Conocophillips Petroleum Company UK Ltd v v Snamprogetti Ltd & Anor

This claim relates to a gas rig in the Hewett field in the North Sea. The claimant, which I shall call Phillips, bore the name Phillips Petroleum Company United Kingdom Limited at the start of these proceedings. On 28^th November 2002 it changed its name to ConocoPhillips Petroleum Company UK Limited. The first defendant, Snamprogetti Limited, which I shall call Snamprogetti, is a company that carries out engineering design. The second defendant, Snamprogetti International S.A., is sued as guarantor of the liabilities of the first defendant.

Snamprogetti undertook with Phillips to provide engineering services, including specification of equipment, for the introduction of offshore compression facilities for gas from wells known as Upper Bunter wells. Those facilities were to be installed on an existing platform in the North Sea known as platform 52/5A. The contract in question is dated 6^th July 1994 and is known as contract UK 0601. The claim is for damages for breach of that contract, plus a small sum said to be due under it. As originally pleaded, the claim for damages was quantified at £5,394,747. By an amendment made shortly before the start of the hearing, the claim was particularized at £3,405,702.58.

Phillips shared the expenses of and revenue from the production and sale of gas from the Hewett field with six other companies owning rights to the gas in the field. Those companies have been called the co-venturers. Phillips had (by a small margin) the largest share in the co-venture. Under an agreement with them entitled the Unit Operating Agreement Phillips was designated as Unit Operator. The Unit Operator was required under that agreement to pay all unit expenses and was entitled to reimbursement from each co-venturer in proportion to its participation in the co-venture. For the incurring of large expenses, Phillips required the specific authority of the co-venturers.

The purpose of the new compression facilities on platform 52/5A was to extract gas from the Upper Bunter wells at a greater rate than would otherwise be possible, in order to meet maximum demands for gas in some of the winter months. Platform 52/5A also received gas from wells known as Zechstein wells. Those wells contained gas at a high pressure, enabling it to flow at sufficient rates without further compression on that platform. The Upper Bunter gas was described as sour gas (in contradistinction to sweet gas) because it contained traces of hydrogen sulphide.

There was a 24-inch diameter undersea pipeline, known as the sea line or the H line, from 52/5A to another platform, 48/29, known as the field terminal platform. Lines from other wells in the Hewett field converged at that platform. A compressor known as the alpha compressor was situated on the field terminal platform. The compression facilities on the field terminal platform had been designed by Snamprogetti. There were two 30-inch lines, the A line and the B line, from the field terminal platform to an onshore terminal at Bacton, in Norfolk. The A line was reserved for sweet gas, i.e gas containing no sour gas. The gas from the A and B lines was commingled at Bacton. At Bacton the gas was compressed to a pressure of some 1000 pounds per square inch (psi) for delivery to British Gas. Snamprogetti was also carrying out work in relation to the facilities at Bacton. There was at one time a proposal, not implemented, for ‘offshore commingling’. That proposal was to allow the A line also to be used for gas containing sour gas.

The following is an outline of the works that the compression project involved. The Upper Bunter gas was compressed by a centrifugal compressor, which I call the main compressor. The main compressor was driven by a gas turbine co-axial with it. The fuel for the gas turbine was Zechstein gas. Before being fed to the gas turbine, that fuel was compressed by a reciprocating compressor. I call that compressor the fuel gas compressor. The main compressor and turbine and associated pipework were assembled on a skid offshore. The fuel gas compressor and its associated pipework were assembled offshore on another skid. The skids were then fitted to the 52/5A platform, the fuel gas compressor skid being borne by the main compressor skid. Other pipework, connections, gauges and so on were fitted on the platform. The offshore work was carried out by a contractor, AMEC Process and Energy Limited (“AMEC”).

During the course of commissioning, in April 1996, the works were stopped by Phillips. At the end of May 1996 a hazard and operability review (“HAZOP”) was carried out. Such a review is a normal part of the procedure to verify the safety and integrity of a process design. Snamprogetti had carried out two such reviews in the course of the detailed design of the 52/5A compression facilities. But this one was ordered specially by Phillips. It was conducted under an independent chairman by Baker Jardine and Associates Limited. Representatives of Phillips and of Snamprogetti attended. It made a number of recommendations concerning the compression facilities. Following the Baker Jardine HAZOP Phillips decided to carry out extensive works on the compression facilities. Phillips required, and obtained, authorization from the co-venturers for the necessary expenditure. Those works were carried out by a contractor known as KYE Limited (“KYE”). With the object of ensuring that those facilities would be in commission by the end of October 1996, Phillips decided to use a support vessel, Seafox by name, to enable the work to proceed as quickly as possible. A substantial part of the damages claim is for the costs of hiring Seafox.

I list below the names of the witnesses before me whom I mention in this judgment. The witnesses of fact for the claimant mentioned below, except Mr. Abernethy, were all employees of Phillips. Mr. Abernethy was employed by another company and seconded to Phillips. I identify their posts as at the material time. They were:

By the contract, Snamprogetti undertook among other things to carry out preliminary engineering, including review of the design basis memorandum. It undertook the preparation of specifications and drawings, and to prepare all detailed design drawings. It undertook to provide procurement services, including invitations to bid and technical and commercial evaluations of bids, and to inspect equipment and materials supplied by vendors. It undertook to review factory witnessed test procedures and to verify the acceptability of those tests. The contract contained a long list of services to be undertaken by Snamprogetti, of which the foregoing is a short extract.

Miss Lindsay Boswell Q.C., for Snamprogetti, submitted that the contractual duties of Snamprogetti were those of an engineer providing professional services. Its actions were to be measured against the standards prevailing at the time of its acts or omissions. She relied on the following passage from the judgment of Bingham L.J. in Eckersley v. Binnie (1988) 18 Con LR 1 at 80,81:

Mr. Peter McMaster, counsel for Phillips, submitted that whilst the contract was for the provision of professional services, it extended to other services. In my judgment, all the services undertaken by Snam, or at any rate those relevant to the issues in this case, can reasonably be described as professional services. That is, I think, recognized by clause 3.6 of the contract. I do not read clause 3.5 as implying that there necessarily exist second class contractors that perform work of a nature similar to that undertaken by Snamprogetti. Rather, it recognizes what may be described as the high class nature of the work undertaken. Nevertheless, whilst not requiring the services of a paragon, combining the qualities of polymath and prophet, the contract did, in my judgment, require something more than ordinary competence on the part of Snamprogetti. It required the competence of the no doubt limited number of firms that undertake work of the relevant kind.

Many of the criticisms of Snamprogetti in this case are criticisms of the equipment supplied pursuant to Snamprogetti’s design. Those criticisms are relevant only in so far as they reflect defects in the work undertaken by Snamprogetti. The work of a first class contractor may not be intended to produce Rolls Royce equipment. The contract provided, in clause 2.1 of the Scope of Work,

Clause 29.8 of the General Conditions of the contract provides:

The Seafox was hired in order to have the compression facilities on 52/5A available by the beginning of November 1996. It is said that the co-venturers required that availability in order to avoid heavy penalties that British Gas was likely to impose on the co-venturers if production of gas in the winter of 1996-97 fell short of the amounts required by their contractual obligations. It is said that the costs of Seafox were reasonably incurred in mitigation of damage.

In my judgment, the damage in question constitutes, or arises out of, loss of production. It also constitutes loss of revenue, since the account between British Gas and the co-venturers was a running account. The penalty is not paid out: it is deducted from payments made by British Gas to the co-venturers. Thus the damage in question is damage for which Snamprogetti is not liable, since it falls within clause 29.8 of the contract. It follows that the costs of mitigating that damage are not recoverable.

In the event of failure on the part of the co-venturers to supply agreed quantities of gas to British Gas, British Gas was entitled to impose a penalty on them. In order to avoid the risk of a shortfall in the supply of gas to British Gas, Mr. King decided that the remedial work should be completed, and the facilities on 52/5A made ready to produce the full capacity, by the end of October 1996. He gave evidence that it was probably between the middle and the end of June 1996 that he made that decision. I accept that evidence. In consequence of that decision, he decided that it was necessary to use Seafox in order to expedite the remedial works. It is in issue whether it was necessary to complete the remedial works by the end of October 1996. I adopt the expression remedial works without assuming liability on the part of Snamprogetti.

Much of the evidence relating to this question is documentary. Mr. McMaster made submissions on the admissibility of documentary evidence. He submitted that the documents were important, but they had limitations: the court must decide the case on all the evidence, including the documents. So far, I accept his submission. The weight to be attached to documentary evidence must depend not only on its inherent probability but also on the other evidence in the case. Mr. McMaster further submitted that the fact that a statement appeared in a document was not, without more, proof of its truth. There was no application to admit hearsay evidence without notice. Many witnesses had been called to speak to the facts in the documents. It was the evidence of those witnesses that the court must assess using the contents of the documents as appropriate. I do not accept the full extent of Mr. McMaster’s submission. It is true that the parties’ agreement of the documents is simply agreement as to their authenticity. But at least since the passing of the Civil Evidence Act 1995 the civil courts have regularly admitted documents in evidence without consideration of the rules about hearsay (see Phipson on Evidence, 15^th edition, 669). No point was taken about this until Mr. McMaster made his submission in his final speech. Indeed, Mr. McMaster, like Miss Boswell, has relied on the documents. In those circumstances, I treat the fact that a statement is made in a document that is before the court as evidence of the truth of that statement. It is not, of course, “proof” of the truth of the statement if by “proof” Mr. McMaster meant conclusive evidence. Nor is the court bound to accept the evidence in the document.

In support of his decision, and in order to justify it to the co-venturers, Mr. King caused a calculation to be made of the net benefit of avoidance of the penalty. That calculation was made on or about 11^th July 1996. On 8^th July 1996 a letter from Mr. D.J.White was sent to the co-venturers’ representatives on a committee known as the Reservoir Engineering Geology and Geophysics Committee (‘REGG’). Mr. D.J.White was head of the reservoir department of Phillips and chairman of the Hewett REGG. Attached to that letter was a histogram. The letter reads as follows (the initials OCM referring to an Operating Committee Meeting):

In spite of the statement in the letter that the histogram showed an estimate for production, it is clear from the other terms of the letter that the histogram was an estimate of production capacity, possibly net of usage as fuel gas by the producer. That conclusion is supported by the wording on the histogram and by the assumptions set out in attachment 2 to the letter. Using the histogram, Mr. King estimated that without the main compressor on 52/5A there would be a total under-delivery of 0.8 billion standard cubic feet of gas in the period November 1996 to April 1997, almost all of which would occur in November and December 1996. I am satisfied that if Mr. King had correctly interpreted the histogram, his estimate was a reasonable one. He calculated that that under-delivery would involve a penalty, the effect of which would be that the co-venturers would suffer a net reduction in revenue of £4 million. That sum he apportioned as to £3 million to 1996 and as to £1 million to 1997. The amount of the putative reduction in revenue and its apportionment are not in issue. For the purpose of calculating the net benefit of avoidance of the penalty by the avoidance of one year’s delay in the completion of the compression facilities on 52/5A, Mr. King used an estimate dated 4^th July 1996 which he had obtained as an estimate of the cost necessary to complete the compression works for 52/5A. That estimate was in the sum of £4.3 million. For the “one year delay” case, Mr. King apportioned that £4.3 million as to £2 million in 1996, and £2.3 million in 1997. (The actual calculations appear to use £1.9 million and £2.4 million, but there are rounding errors, which I shall ignore). For completion of the work by the end of October 1996 (the “base case”), on the other hand, the whole £4.3 would be spent in 1996. On that basis, he instructed a Mr. Heberlet, an economist employed by Phillips, to run a one year delay case in order to show the benefit of avoidance of the penalty by avoidance of the one year’s delay. Mr. Heberlet calculated the post-tax net present value of the one year delay case as £3.8 million, using a discount rate of 10 per cent. a year. The base case had a corresponding net present value of £5.9 million. Thus the effect of Mr. Heberlet’s calculation was to show a net present value of £2.1 million as the benefit of avoiding the delay. The corresponding figure using a zero rate of interest was £2.4 million.

The estimate of costs amounting to £4.3 million included not only remedial works but also outstanding works and an item of £1 million described as original overspend. It was based on the use of Seafox. If and in so far as the use of Seafox increased the cost of carrying out the works, the estimate of £2.1 million was an over-estimate, since Seafox was not needed on the one year delay case. Mr. King expressed the view in evidence that the use of Seafox did not add to the expense. It is said that the use of Seafox increased the costs by about £1 million. I have carried out a rough and ready calculation, which is the best I can do, of the net present value of avoiding the delay on the assumption that the use of Seafox increased the costs by £1 million. As I have said, on the calculation sheet of the one year delay case the cost of £4.3 million had been apportioned as to £1.9 million in 1996 and as to £2.4 million in 1997. I have apportioned £3.3 million, in the place of £4.3 million, in the same proportions, retaining in addition a figure of £1.2 million as capital expenditure for 1996 which is common to both the base case and the one year delay case. The pre-tax net present value of the one year delay case at the discount rate of 10 per cent. I calculate as £9.36 million (compared with Mr. Heberlet’s figure of £8.6 million). It is not possible for me to calculate the tax, but it seems that allowing for the discount the rate must average about 54 per cent., leaving a post-tax net present value of £4.3 million. On that basis, the net present value of the benefit of avoiding the penalty would be £1.6 million. I should add that net “present” value actually means value at July 1994.

But net present value was not the only consideration of the co-venturers. Rates of return have not been explored in this case. I am unable to say whether the co-venturers would necessarily have agreed to completion of the works by the end of October 1996 if the net present value of avoiding the penalty had been £1.6 million. But I am by no means satisfied that if Mr. King had correctly interpreted the histogram his decision to recommend completion of the works by that date or the decision of the co-venturers to accept that recommendation was unreasonable. Whether Mr. King had correctly interpreted the histogram is a question to which I now turn.

Miss Boswell submitted that the 52/5A compression facilities were not intended to add to production or sales of gas until the fourth quarter of the calendar year 1997. That information was to be derived from a document entitled Hewett Gas Sales Forecast and dated 4^th July 1996. Although that information was available to Phillips, it was not disclosed to the co-venturers at the material time. The document showed projected sales of gas with and without the 52/5A compressor for the calendar years 1996 to 2005 and, for each of those years, the difference (described as incremental annual gas sales) between the former and the latter. The incremental sales for 1996 were shown as nil. Miss Boswell submitted that the likely reason for the nil figure was that at low suction pressures the Upper Bunter wells were found to produce sand with the gas. Thus the whole of the theoretical increase in production from drawing down the Upper Bunter wells using the 52/5A compressor could not be obtained. But equivalent production could be obtained by directing the 52/5A gas through the alpha compressor, thereby using the alpha compressor to pull down both the Upper Bunter and Zechstein wells to a suction pressure that did not cause sand production. The document did indeed state that the figures for sales without the 52/5A compressor were based on the assumption that gas from 52/5A would go into the alpha compressor.

Mr. King was cross-examined on this subject. I found Mr. King to be a defensive witness, unwilling to admit even the obvious. I was not impressed by his evidence. The document dated 4^th July 1996 that I have mentioned was the first page of a four-page document produced by the reservoir department of Phillips. When Mr. King was first cross-examined on that first page, the other three pages had not been disclosed. The copy document on which Mr. King was first cross-examined did not show its date, since part of the original had, in the course of business, been covered by a yellow sticker which happened to cover the date. He said that the document was probably provided to him during the first week of July 1996. Mr. King accepted that at the time the document was prepared it showed that there was to be no production using the 52/5A compressor in 1996. He said he did not know when it was prepared, but it certainly did not reflect Phillips’s intent at the time that he provided it to Phillips’s economists. He said that from the time of his involvement in the project [that is, early January 1996] it was expected that the compressor would operate from the start of the 1996 winter. The sticker in question contained a message in Mr. King’s hand to John Gerd, who ran the Phillips economic group at Woking. It said “Herewith latest profiles…” It was put to Mr. King that it would appear that the document must be dated around the period in July when he had prepared the cost estimate [which was dated 4^th July 1996]. He answered that he thought that “latest” meant latest to him, not latest in the context of Phillips. I find that answer inherently implausible, though it is fair to say that Mr. King was speculating. In fact the document was dated 4^th July 1996, as the complete version of four pages disclosed late revealed. (The transcript references to the evidence I have mentioned above are Day 23, p.116; Day 29, pp. 150, 151; Day 24, p.137).

Miss Boswell cross-examined Mr. King about the histogram which was attachment 1 to the letter of 8^th July 1996. The histogram showed the contribution assumed to be provided by the 52/5A compression (some 35 million scuffs of gas per day at the end of October 1996, diminishing to some 25 million scuffs per day at the end of April 1997). She asked him on what pressures that histogram had been based. Mr. King said that he could not answer the question. He thought it would be better answered by the person or group that developed the histogram, namely the Phillips reservoir engineering group. He believed it was David White who had constructed the document, and he, Mr. King, had had the benefit of some discussions with him. (Day 24, pp.119, 112; Day 29, p.75). Neither Mr. White nor anyone else from the group was called as a witness. No documents were disclosed throwing any light on the production of the histogram. No documents were produced contradicting the figures in the Hewett gas sales forecast.

Further documents from the reservoir group, also disclosed only during the course of the hearing, were production profiles for the Hewett Field dated 22nd November 1995. They showed substantial annual sales of Upper Bunter gas without compression facilities starting in the year 1996 and continuing to 1998, tailing off in 1999 and 2000 to zero in 2001 and thereafter. Figures for Upper Bunter with compression facilities were also shown. They showed zero for 1996, and sales building up from a small start in 1997 to a peak in 2003 and falling off thereafter. The sales from Upper Bunter with compression facilities started to exceed the sales from Upper Bunter without compression facilities in 1999. Those figures support the proposition that it was not intended to use the Upper Bunter compression facilities until 1997.

In more detail, the four-page document entitled Hewett Gas Sales Forecast breaks down as follows. The first page in point of time was dated 3^rd July 1996. It is headed “Including Zechstein B11, and Little Dotty 48/30-15 commingled with Della and new 52/5a compressor curve 2 July 96. Case Sens7”. That has been referred to as Case Sensitive 7. The other three pages are each dated 4^th July 1996. One of them is headed “Case Sens8. Including Zechstein B11, and Little Dotty 48/30-15 commingled with Della. New 52/5a compressor curve (2 July 96) is not included. 52/5a gas into Alpha compressor”. The third is headed “Case Sens7-8. Incremental reserves due to new 52/5a compressor”. Each of those three pages shows quarterly gas sales and annual gas sales by calendar year, among other information. The gas sales shown in the document headed Case Sens7-8 are, for each quarter and each calendar year, simply those shown in Case Sensitive 7 minus those shown in Case Sensitive 8 for the corresponding quarter or year, as the case may be. The figures on Case Sensitive 7-8 from the third quarter of 1996 to the third quarter of 1997 inclusive are all zero. That is not so for the preceding three quarters, going back to the fourth quarter of 1995. But neither party relies on the figures for those quarters in Case Sensitive 7 or in Case Sensitive 8. Mr. McMaster said that the first three quarterly figures were an apparent nonsense: there was probably a reason for them, but he could not identify it. The annual gas sales figures on the fourth page, which is the page to a copy of which Mr. King’s note to Mr. Gerd was stuck, were the same as the respective figures on the Case Sensitive 7, Case Sensitive 8 and Case Sensitive 7-8 pages, save that for the calendar year 1996 the figure for Case Sensitive 8 was adopted for both the Case Sensitive 7 and the Case Sensitive 8 columns, so that the incremental sales for 1996 were shown as zero.

The “new 52/5a compressor curve 2 July 1996” is not in evidence. Other compressor curves are in evidence. They show the relationship, for a given compressor, between suction pressure, outlet pressure, speed of compressor, rate of gas flow and power consumption. They also show the lower limits of flow below which the compressor cannot operate (“surge lines”) as a function of power and discharge pressure.

Another document from the reservoir department disclosed during the course of the hearing was dated 20^th June 1996. It was described by a covering letter as showing how the Bacton-Hewett area could be operated, and was headed “Theoretical well flows for given DCQ”. (DCQ stands for daily contract quantity, a term defined by the agreement with British Gas). That document was addressed to, among others, Mr. Halliwell, who was questioned about it. He did not recall receiving the document, but he agreed that it appeared to show that what was being considered was a flow configuration in which the Upper Bunter wells from 52/5A came in only at the highest nominations (periods of highest demand from British Gas), though it might continue to be run in that way if nominations were reduced, and that in so far as there was compression, the compression was through the Alpha compressor. (Day 45, pp.165, 168, 169).

Thus all the documents disclosed from the files of the reservoir department proceed on the basis that there was perceived to be no advantage in using the 52/5A compressor in 1996. Mr. McMaster put forward a speculative argument that the calculation of incremental annual gas sales contained in the Hewett Gas Sales Forecast was based on the assumption that the 52/5A compressor would not be available in 1996. (Indeed, Mr. King speculated that they were prepared for the purpose of the one year delay case). To that end, submitted Mr. McMaster, the figures of gas sales for the contract year 1996/97 were set equal to those in Case Sensitive 8. But that implies that the figures in Case Sensitive 7, which was printed out on 3^rd July, incorporated figures not printed out until 4^th July. That is, of course, not impossible, but it reduces the probability of what in any event I find an unconvincing speculation. All this speculation has been considered necessary because there is no evidence on the point from any witness from the relevant department.

There is evidence from Mr. Halliwell as to the purpose and effect of the Hewett Gas Sales Forecast (Transcript, Day 45, pp.184, 185):

The message which Mr. King wrote on the yellow sticker to Mr. Gerd was

Thus the figures used by Mr. Heberlet show no difference in production between the base case and the one-year delay case. The calculations simply show the effect on present value of incurring a penalty of £4 million, mostly in 1996, with delaying capital expenditure of £2.4 million from 1996 to 1997. The production figures, though rounded to 100,000 scuffs per day, are consistent with those set out in the fourth page of the Hewett Gas Sales Forecast dated 4^th July 1996.

Miss Boswell submitted that either Mr. King and Mr. White had not read and understood the production profiles entitled Hewett Gas Sales Forecast and had misunderstood the meaning of capacity, or they were simply trying to create a figure for loss of production in 1996 which could not be derived from the production profiles. If, she said, the production profiles had been produced merely for “economics” (a reference to Mr. King’s message to Mr. Gerd written on the yellow sticker), then Phillips would have been able to produce to the court another production profile generated at about the same time which showed the intended production for 1996 using the 52/5A compressor. No such document had been produced. If Mr. King had wanted to produce for the co-venturers a proper comparison between the economics of completing the compression works in 1996 and completing them in 1997, he could have obtained production profiles on different bases and assumptions specifically identified. He did not.

I am satisfied that Mr. King had not read the first three pages of the profiles. There is no evidence whether Mr. White had seen them. Mr. King appeared to have had little recollection of the fourth page. I am satisfied that he could have taken little notice of it when he sent it with the sticker to Mr. Gerd, since it showed that the 52/5A compressor, if installed, would make no difference to gas sales in 1996. What Mr. King relied on in instructing Mr. Heberlet that the super shortfall penalty would be incurred and would cost £4 million by way of loss of revenue was the histogram. He did not understand how it had been arrived at, but he understood its effect. It is true that it represented capacity, not sales. But it was highly relevant to Mr. King’s calculation.

The information in the histogram, however, was not the reason, or at any rate the original reason, for Mr. King’s decision that the compression facilities should be installed by the autumn of 1996. He had made that decision before the end of June. The letter of 8^th July mentioned a request made at the OCM meeting on 30^th May for an estimate of how delays in the 52/5A compression project would affect Hewett production capacity. The part of the minutes of that meeting that has been disclosed records no such request. The minutes do, however, state the following:

First, there is another histogram which, though differing from attachment 1, contains the same information as respects 52/5A compression as that contained in attachment 1. It appears with inter-office correspondence from Mr. White dated 14^th August 1996 and relating to a different platform. That makes it improbable that the histogram at attachment 1 was devised for the particular purpose of creating figures to support the AFE supplement.

My second reason is this. Attachment 1 is not necessarily inconsistent with Case Sensitive 7-8, even though the latter is entitled “incremental reserves due to new 52/5a compressor”. That is because the comparison in Case Sensitive 7-8 is between using the 52/5A compressor and sending the gas from 52/5A through the alpha compressor. The estimate of Hewett capacity without the 52/5A compressor said in the letter of 8^th July 1996 to have been provided by the histogram appears to be based on the assumption that the 52/5A gas will not be passed through the alpha compressor. The only reference to the alpha compressor in the assumptions set out in attachment 2 to the letter is assumption 5, viz. “Alpha Compressor aftercooler is assumed to be fully cleaned”.

My third reason is that the 52/5A compressor was indeed used from 18^th October 1996. Mr. Hobson gave the following evidence (Transcript, Day 32, p.20):

I accept that evidence. Further, there are in evidence figures of 52/5A gas flows from 1^st November 1994. The same source (described as a table of daily production flows) shows H line pressures and 52/5A compressor suction pressures. Whilst it is clear that those figures are not entirely reliable, they support that evidence of Mr. Hobson. They also show that the 52/5A compressor was used from 18^th October 1996, the H line not being connected to the alpha compressor.

The suction pressures of the 52/5A compressor in October and November 1996 varied between about 100 and 115 psia; the H line pressures were mostly between 190 and 275 psia. (Psia means pounds per square inch absolute, i.e. measured above the vacuum; psig means pounds per square inch gauge, i.e. measured above atmospheric pressure. The difference is approximately 15 psi). Miss Boswell submitted that whatever might have been the data giving rise to the histogram at attachment 1, it was clear from the Basis of Design that an increment of the order predicted in the histogram could be achieved only if the compressor could suck the Upper Bunter wells down to a suction pressure of 69 psia and discharge at a pressure of 143 psia. I am not persuaded that that is correct. Some figures were in evidence which showed shut-in pressures for 52/5A Upper Bunter and Zechstein wells for years up to 2001. (Shut-in pressures are pressures in the well when the gas is not flowing). They appear to be historic data, in that they are described as shut-in pressure data in the heading of the document in which they appear. Shut-in pressures for the Upper Bunter wells in 1996 were shown as lying between 186 and 189 psi (higher than those shown for 1995). Mr. Sylvester-Evans said that he understood that those figures were in psig. Dr. Robinson said that in the absence of other information he tended to assume it was psig, but he did not know. The figure of 69 psia referred to by Miss Boswell as giving rise to a flow increment of the order of 30 million scuffs a day was based on flowing wellhead pressures which must have been related to the assumed shut-in wellhead pressure of 167 psia stated in the Basis of Design, sheet 5. There is no evidence what the increment would be with a shut-in wellhead pressure of 186 psia or 186 psig and any given H line pressure, but I can see no reason why an increment of 30 mmscf/d should require a suction pressure as low as 69 psia.

Further, Dr. Robinson gave his opinion that the table of daily production flows showed a real contribution [from the 52/5A compressor] of numbers around 30 million scuffs [a day] in the actual pipeline conditions prevailing in 1996 (transcript, Day 66, p.69). I find that figure entirely credible in the light of my own understanding of the evidence concerning the relationship between gas flow between two points in a given system and the pressures at those two points. I accept his evidence on this point.

The evidence concerning the relationship between gas flow and pressure itself comes from Dr. Robinson. His evidence implied (and I take judicial notice of the relevant calculation) that the mass rate of flow of gas in a cylindrical pipe carrying gas at a uniform temperature is proportional to the square root of the difference between the squares of the absolute pressures at the ends of the pipe. The constant of proportionality depends on the dimensions of the pipe, the temperature and the molecular weight of the gas. The situation is not so simple where, as here, the pipework in question is of a complicated shape, or where the temperature is not uniform. The papers show instances of what may be empirical formulae used by Phillips where the gas flow is stated to depend, not precisely on the square root of the difference in question, but on powers differing slightly from 0.5 (a power of 0.5 represents a square root).

I conclude that Mr. King’s decision that it was necessary to have the 52/5A compressor in operation by the end of October if the super-shortfall penalty was to be avoided was reasonable. Whether it was economic to do that, having regard to rates of return, is a separate question.

Phillips complained that the fuel gas compressor selected by Snamprogetti induced vibrations throughout the platform.

The fuel gas compressor was a single-cylinder reciprocating compressor. Its crankshaft was driven by a single-speed electric motor with belt drive. The compressor operated at a frequency of 551 revolutions per minute, or 9.18 Hertz (cycles per second). When running, it caused vibrations of which the claimants complain.

Single-cylinder reciprocating compressors are peculiarly liable to cause vibrations. The vibrations are the result of unbalanced cyclical inertial forces. Those forces are associated with the cyclical accelerations of the massive parts of the compressor, specifically the piston, piston rod, crosshead, connecting rod and cranks. Those forces would be substantially balanced by equal and opposite forces in the case of a horizontally-opposed twin cylinder compressor. That is a compressor with two cylinders on the same axis but situated on opposite sides of a single crankshaft, and driven by diametrically-opposed cranks. In general, any twin- or multiple-cylinder compressor is likely to be better balanced than a single-cylinder compressor. One way of balancing a single-cylinder compressor is to fit it with opposing moving masses analogous to those in a horizontally-opposed twin. The fuel gas compressor on the 52/5A platform had a horizontal cylinder and was fitted with a balance-weight on the crankshaft which reduced the unbalanced horizontal forces but increased unbalanced vertical forces. It was a compromise which Snamprogetti considered acceptable.

Simple harmonic motion is an oscillatory motion where the acceleration is always directed towards a central position and is proportional to the instantaneous displacement from that position. If, for example, a point moves round the circumference of a circle with uniform speed, the projection of that point on any diameter of the circle moves with simple harmonic motion about the centre of the circle. Since the generating point has a constant acceleration, always directed towards the centre of the circle, the projected point has its maximum acceleration, equal to that of the generating point, when it is at the extremities of its motion. The projected point has zero acceleration as it passes through the centre of the circle. At that instant, the acceleration of the generating point is at right angles to the selected diameter, and so has a zero component along the diameter. In general, the component along that diameter of the acceleration of the generating point at any instant is proportional to the projection along the diameter of the radius joining the centre to the generating point. That component represents the acceleration of the projected point, which is thus proportional to its distance from, and directed towards, the centre.

When the compressor is operating, the big end of the connecting rod moves with uniform speed around a circle. The centre of that circle lies on the axis, produced, of the cylinder. Thus the motion of the projection of the big end on that produced axis is simple harmonic motion. But the motion of the piston is not simple harmonic motion. The reason is that the distance of the piston from that projection is not constant. At each end of the stroke, its distance from the projected point is equal to the length of the piston rod plus the length of the connecting rod. But that is not so at any other point in the cycle. At any other point, the distance is equal to the length of the piston rod plus the length of the projection on the produced axis of the length of the connecting rod. The connecting rod is the hypotenuse of a right-angled triangle of which the projection is another side, shorter than the hypotenuse. Thus for example when the crank is at quadrature the piston is closer to the crankshaft than when the piston is at the centre of its stroke.

The oscillatory motion of a point may be plotted as a graph of position against time. Where the point is moving with simple harmonic motion, the curve obtained is a sine curve. That is because the projection, on the diameter, of the radius vector to the generating point is proportional to the cosine of the angle between that radius vector and the diameter, or to the sine of its complement. The angle may be expressed as the product of the angular velocity of the crankshaft and the time interval measured from some datum. The motion of the piston can be exactly expressed as the sum of an infinite series of sine curves. The first, or fundamental, represents the simple harmonic motion of the projected point, it has an amplitude equal to the radius of the circle, i.e. the length of the crank, and its frequency is the frequency of rotation of the crankshaft. The others (‘harmonics’) are of even multiples of the fundamental frequency and of ever-diminishing amplitudes as the multiple increases. It is common ground that the effect of all the harmonics except the first is negligible. Thus, the relevant frequencies are the fundamental frequency and twice that frequency.

When an object is slightly deformed, elastic forces resist its change in shape, and tend to restore it to its natural shape. The magnitude of those forces is proportional to the degree of displacement and depends on the shape of deformation. If the object is released, it will oscillate until the energy used in deforming it is dissipated. The more rapidly the energy is dissipated, the more heavily damped is the object said to be. The nature of the oscillation will depend on the shape to which the object is deformed. The frequency or frequencies of such an oscillation, known as natural frequencies, will depend on the physical characteristics of the object. Each natural frequency corresponds to a particular mode of deformation, or “mode shape”.

The vibration of the fuel gas compressor was capable of causing the members of the structure of the platform to vibrate. The different members of the structure had different natural frequencies. Any member that was excited at a frequency at or near a natural frequency of that member was liable to resonate, i.e. to oscillate with an amplitude greater than would otherwise be the case. The greater the amplitude, the greater the stress in the member. Thus resonance was liable to cause damage to the structural member of the platform, by fatigue or otherwise. What I have said about the structural members of the platform applies also to the pipework, instruments and gauges on the fuel gas compressor skid. Oscillations can also have an adverse effect on persons subjected to them. Whilst the platform was intended to be not normally manned, the presence of persons on the platform had to be allowed for.

The propensity of an unbalanced reciprocating compressor to cause vibrations is, I am satisfied, a fact notorious to the engineering profession and was known to Snamprogetti at the material time. Snamprogetti specified the balance weight to reduce vibration and carried out calculations of the response of the structure of the platform to be expected from the vibration of the fuel gas compressor.

The claimants’ case was this. An unbalanced reciprocating compressor should not have been specified. The module which was to support the new equipment, including the fuel gas compressor, was a structure in the form of a cantilever, projecting from the side of the existing platform. As such, the structure was of a relatively flexible design attached to an edge of the platform that itself had some flexibility. In those circumstances equipment should not have been considered which generated unbalanced forces during its operation. Had it been impossible to specify a compressor that would minimize vibration and operate in a balanced manner, a fuel gas system without a compressor should have been designed. Phillips’s case was that it might or might not have been possible to secure a suitable compressor (i.e., one that did not induce unacceptable vibrations): Phillips made no positive case that any other type of compressor would have been viable.

I should add for the sake of accuracy that the support structure was not strictly a cantilever, since it was knee-braced.

A single-cylinder fuel gas compressor, not fully balanced, was certainly not an ideal compressor for use on the platform. A publication produced from Snamprogetti’s library states the following. Any stationary compressor must be anchored to a solid mounting. Most reciprocating compressors require attention to foundation design. A proper foundation minimizes vibration and prevents its transmission to adjacent building structures. Those remarks are clearly directed to onshore installation of compressors. The extract from that publication that is before the court gives no indication of the title or date of the publication. Dr. Walker gave evidence that usually in land-based applications a single-cylinder reciprocating compressor would be mounted on a large block of concrete that itself is often buried in the ground. The large mass reduces vibrations to acceptably low levels. Mr. Middleton agreed that that was sound engineering advice for reciprocating compressors on the ground.

Mr. Richard Davies, Chief Structural Engineer at Phillips, gave clear evidence that a compressor of the kind specified by Snamprogetti should not have been used, and that he was unaware of the intention to use it. The following are extracts from his cross-examination (Day 10, pp.72, 64):

There is evidence before me of the use offshore throughout the world of reciprocating gas compressors. Hardly any, if any at all, are unbalanced single-cylinder compressors. A single-cylinder compressor was supplied to Phillips for intermittent use on a platform known as the Maureen platform. On the same platform, there is a three-cylinder compressor which, as Mr. Davies told the court, caused vibration by out-of-balance forces. Whilst that compressor was much more massive than that used on platform 52/5A, the Maureen platform is much more massive than the 52/5A platform. Mr. Sylvester-Evans gave evidence that there had been four large compressors and two large unbalanced reciprocating pumps on the Piper Alpha platform in the North Sea. In the early days, he said, there had been a series of piping failures associated with the reciprocating compressors and also with the reciprocating pumps.

Dr. Walker had made enquiries of various suppliers, including Peter Brotherhood Ltd. He wrote in one of his expert’s reports:

It seems to be a reasonable inference from the foregoing passage that the single-cylinder compressor in question was not balanced.

Mr. Gilbert Thomson, who held the position of Principal Engineer with Snamprogetti, was involved in engineering and specifying the requirements for, among other things, the fuel gas compressor. No particular type of compressor was specified. Invitations to tender for the supply of the fuel gas compressor were issued to potential bidders by Snamprogetti on behalf of Phillips. No criticism is made of the list of invitees to tender. Bids were received from three companies, George Meller Ltd., Gruppo Benelli, and Dresser Rand (U.K.) Ltd. George Meller offered two compressors, the single-cylinder reciprocating compressor that was accepted, and a diaphragm compressor. Gruppo Benelli offered a reciprocating compressor. Dresser Rand offered a screw compressor.

The Gruppo Benelli offer was rejected on the grounds that insufficient technical information was provided (even after specific request), that the compressor required cooling water and that the weight of the package at 4.5 tonnes exceeded the target weight of 4 tonnes. I am satisfied that the first-mentioned ground was sufficient for the rejection of the offer.

The screw compressor of Dresser Rand was primarily a refrigeration compressor. In the opinion of Snamprogetti, it was not made of suitable materials. The material failed to comply with the requirements of NACE (National Association of Corrosion Engineers). Compliance with those requirements was a condition laid down in the invitation to tender. Dr. Walker considered that the specification by Snamprogetti of compliance with NACE requirements was unreasonable, since the equipment would not be exposed to the sour Upper Bunter gas, but only to Zechstein gas, which was sweet. However, at a design HAZOP study held in August 1994, conducted by M. George of Snamprogetti and attended by, among others, A. Hess of Phillips Operations Department and M. Webster of Phillips Engineering Department, the possibility of sour gas operation at some future date was mentioned. An observation that the system was designed for operation with sour gas was noted, and it was decided to add a note to the P & I (process and instrumentation) drawings to that effect. Whilst that does not relieve Snamprogetti of responsibility, it does suggest that there are two reasonable views on this subject. I am confirmed in that view by evidence from Mr. Sylvester-Evans, who said (Day 70, p.29) that it was very common to assume that reservoirs would sour. It was a problem that had occurred in the past and therefore it was common to apply the NACE standard especially where you had sour gas on board.

Moreover, there was another objection to the materials of the screw compressor: the screw and its casing were made of cast iron. Mr. Thomson gave evidence that in the event of fire cast iron cracked and promoted the fire (Day 36, p.44). I accept that as a sufficient reason for rejecting the screw compressor.

The screw compressor was rejected also on the grounds that the electric motor rating at 55 kW was too high, the footprint of the equipment was too large and the equipment was too heavy. Dr. Walker expressed the view that the power requirement at 55 kW was only marginally greater than that of the chosen compressor (45 kW). He also said that some redesign of the compressor module could have accommodated the larger footprint area and weight compared to those of the reciprocating compressor. In rejecting the screw machine, Snamprogetti laid the foundation, he said, for all the vibration problems that occurred later in the project.

The figures giving rise to the objections on the ground of weight and footprint were these. The weight was given as between 6.5 and 8 tonnes. The length and width of the footprint were given as 4.8 and 2.5 metres respectively. Those figures were supplied by Dresser Rand on 2^nd June 1994 as “Approximate package dimensions and weights”. On 5^th July 1994 Dresser Rand sent what they described as more accurate weights and dimensions for the package. They were 9300 lb (4.22 tonnes) and 16 ft by 6.2 ft (4.9 m by 1.9 m). The technical evaluation was completed on 13^th September 1994, and it was the former figures that were used. It was put to Mr. Thomson in cross-examination that the objection to the screw compressor on the ground of its excessive weight was not necessarily correct. He said he would have to check, but he would be most surprised if the weight was not between 6.5 and 8 tonnes. It is not necessary for me to make any finding on the question of the weight or the footprint of the screw compressor.

The power requirement as stated by Dresser Rand was 45 kW for the main driver and 1.5 kW for an auxiliary oil pump. It appears that the main driver motor was an American motor which was unacceptable on the ground that it did not comply with British Standards. Mr. Thomson had to provide for a motor which would have consumed 55 kW. It is not in issue that the Dresser Rand screw compressor would have required a 55 kW motor.

I am satisfied that the screw compressor would not have given rise to any significant problem of vibration. It seems that it was rejected with regret. There is a memorandum which appears to be in the hand of Mr. Whittington, Snamprogetti’s project manager from June 1996, saying “We wanted screw compressor, but only C.I. Also took more power (limitations)”. C.I. means cast iron.

The diaphragm compressor offered by George Meller was rejected on three grounds. First, its power requirement was 72 kW. That was considered excessive having regard to the stated limit of 45 kW. Second, the footprint was too large at 6.3 m by 3 m. Third, the weight was 10 tonnes, substantially in excess of the target of 4 tonnes.

I am satisfied that Snamprogetti acted reasonably in taking into account the electrical power consumptions of the offered compressors. I consider the power constraints affecting the platform in paragraphs 416 to 421 below, under the heading Electrical power requirement.

The structural engineering of the compressor module has not been explored in relation to static forces. The effect of Mr. McMaster’s submissions was that it was for Snamprogetti to justify rejection of the diaphragm compressor. It is clear on the evidence that the weight of the compressor was an important consideration. Mr. Thomson considered 10 tonnes excessive. There is no contrary evidence that a satisfactory design of the compressor module could have been made allowing for a weight of 10 tonnes for the fuel gas compressor. I am satisfied, having regard also to what follows in relation to the fuel gas compressor, that Snamprogetti acted reasonably in rejecting the diaphragm compressor.

That left only the reciprocating compressor offered by George Meller. That had an acceptable footprint of 4m by 3m, an acceptable motor rating of 45 kW, and was thought to have had a weight of approximately 4 tonnes, which was acceptable. The order was placed on 12^th October 1994. In early December 1994, George Meller advised an increased weight in excess of 6 tonnes.

Mr. McMaster submitted that the weight limit of 4 tonnes which Snamprogetti adopted was arbitrary and unnecessary. The weight of the compressor substantially exceeded that limit. He pointed to a document of Snamprogetti entitled Conceptual Engineering Weight Report and dated 25^th June 1993 which showed in Appendix A1 a weight of 7000 kg for the fuel gas compressor. When that document was put to Mr. Thomson in cross-examination, he said he had never seen it before. It was part of Snamprogetti’s Offshore Compression Study Report to October 1993. Appendix A1 was an appendix to an attachment to that report entitled Conceptual Engineering Weight Report. In that attachment, it was stated that the weights summarized were for the purpose of preliminary re-analysis of the deck and jacket (meaning, as I understand it, the platform and its loads). The weights were described as discipline weights, and it was stated that ‘Discipline weights are modified as necessary to include engineering design allowance and in other cases to remove cut and waste (cost estimating) allowance’. I do not derive assistance from the figure of 7000 kg.

The reasons why the 4 tonne limit was imposed, though obviously relating to the design of the support structure, have not been explored. The same applies to the dimensions of the footprint. Dr. Robinson and Mr. Sylvester-Evans were agreed that weight, space and power requirements, among other factors, were relevant to the selection of a compressor.

Snamprogetti were responsible for the design of the extension to the 52/5A platform required to accommodate the main compression facilities. That extension comprised a module deck, roof and bracing. Snamprogetti advised Phillips to commit themselves to the purchase of the reciprocating fuel gas compressor from George Meller before they, Snamprogetti, had calculated the effect of the unbalanced oscillatory forces on the module. Mr. McMaster submitted that Snamprogetti were in breach of their duty in so advising Phillips. Snamprogetti were confident that they could design the structure so as not to vibrate unacceptably. Mr. McMaster submitted that the state of mind of Snamprogetti was irrelevant. It was their duty to ascertain that the compressor would not cause unacceptable vibrations. He also submitted that the calculations were in any case flawed.

Mr. McMaster’s criticisms of the calculations relating to the vibration of the structure were based on the evidence of Dr. Walker. I accept that the calculations did not lead to an accurate prediction of the response of the structure to the vibrations caused by the fuel gas compressor. But I do not accept that as a criticism of the choice of fuel gas compressor. It is not claimed that Snamprogetti’s design of the module itself constituted a breach of any duty owed by Snamprogetti to Phillips.

On 11^th February 1996, during commissioning of the fuel gas compressor, with the compressor running under no load, it was considered that the vibration levels on the structure and on some of the pipework and fittings associated with the compressor might be excessive. A contemporary note of Mr. M. G. Critten, of George Meller Ltd., describes the situation. It says

Phillips commissioned Acoustic Technology Limited (‘ATL’) to conduct an initial survey of vibration. That survey was carried out by Mr. John Richardson on 24^th February 1996. He measured vibration velocity levels at 27 locations on the fuel gas compressor, fuel gas compressor skid, pipework, ancillaries and the main structure (i.e., the compressor module) while the fuel gas compressor was running on no load. In addition, with the compressor not running he carried out impact response tests to determine the natural frequencies of the fuel gas compressor, the south diagonal brace and the fuel gas filter. ATL issued their report on 27^th February. It stated that the natural frequencies of the south diagonal brace and the fuel gas filter were very close to the running speed or twice the running speed of the compressor. As those items were lightly damped, they would be easily driven by excitation from the compressor. The natural frequency of the knock out drum drain valve/pipe was not tested, but in view of the very high level of vibration measured at 18.75 Hz, it was likely that its natural frequency coincided with the second harmonic of compressor speed. (The expression second harmonic clearly means what I have called the first harmonic in this judgment).

Mr. Davies first heard about the vibration problem from Mr. King, by telephone in February 1996. I accept that from the evidence of Mr. Davies. Mr. Davies’s evidence was that Mr. King had mentioned to him that there were major vibrations to the diagonal support beams of the cantilever structure, and that two of the braces in particular exhibited significant vibration. I do not accept that in fact two of the braces exhibited significant vibration. The report shows significant vibration of the south diagonal brace in the north-south direction, but insignificant vibration of the north diagonal brace. Either Mr. Davies’s memory of the details is inaccurate or Mr. King was mistaken when he made his telephone call to Mr. Davies.

Mr. Davies heard of a suggestion from Snamprogetti that the end connections of the braces should be changed to stiffen them. That would increase their resonant frequency. Mr. Davies considered that to be an impractical suggestion since the necessary welding would not be easy to carry out. He devised the idea that the braces should be filled with liquid. By increasing their mass, that would reduce their resonant frequency. The parties met on 1^st March 1996 to discuss the problem.

In spite of Mr. Davies’s view, expressed in his evidence, about the unsuitability of using a single-cylinder unbalanced compressor on the platform, at no time did he enquire of Snamprogetti why they had advised the use of such a compressor or offer any criticism to them about it.

Following the meeting of 1^st March, Mr. Natarajan sent a note dated 4^th March 1996 to Mr. Davies. Mr. Natarajan was the Snamprogetti engineer who had calculated the effect on the structure of the unbalanced forces from the fuel gas compressor. The essence of what he said was this. He understood that the main vibration problems were on a diagonal member, number 153 (the ‘south diagonal’). It was described as 273 X 8 CHS, which means a circular hollow section of 273 mm diameter and of 8 mm wall thickness. The south diagonal had an unacceptable level of mid-span vibration in the north-south direction. The local vibration mode of that member was very close to twice the running speed of the reciprocating compressor. Although the north diagonal had identical geometric characteristics, its vibration was slight. Therefore, only the mid-span vibration of the south diagonal was influenced by the running speed of the fuel gas compressor. In Snamprogetti’s dynamical analysis calculations, the sectional properties of the diagonals had been taken as 219 X 10 CHS, contrary to the fabrication drawings. The detailed drawings were developed subsequently to the dynamic analysis. The end constraints of the two diagonal members had been correctly applied, with moment releases at supports in the north-south direction. The natural frequency of vibration of the diagonal using 219 X 10 CHS was 14.1 Hz, which was between the operational frequencies of the compressor. He had attached to the memorandum a summary of the natural frequencies of vibration of the diagonals with three possible end conditions. That table confirmed that by changing one of the end conditions to fixity the local vibration problem could be minimized if not eliminated. The additional restraints against bending could be achieved by stiffening the gusset plates either by welding or by using mechanical connections. Increasing the mass of the diagonal by filling it with a suitable material would also eliminate the present mid-span vibration problems. The density of the filling material was crucial. For example, using a non-structural filling material with a density of 2000 kg per cubic metre could easily give rise to a natural frequency close to the running speed and introduce another set of vibration problems. On that basis, he recommended that a suitable modification be made to the south diagonal. He indicated his willingness to be of further assistance, if required.

Mr. Davies’s solution was adopted. The south diagonal was filled with water containing, at Snamprogetti’s suggestion, anti-corrosion material. On 6^th March 1996 AMEC sent to Snamprogetti AMEC’s proposal to fill the brace with water, inviting Snamprogetti to advise their acceptance or an alternative arrangement. On 12^th March Snamprogetti sent to Phillips their comments on AMEC’s proposal. They stated that they did not recommend filling the diagonal with a water solution. They commented as follows. They had calculated that filling the diagonal with water would introduce a mid-span vibration frequency of 12.6 Hz, which lay between the running frequencies of the fuel gas compressor and was close to the running speed. The main disadvantages of water filling were leakage, ice formation and thermal expansion, corrosion, inspection and maintenance requirements, and introduction of stress concentration in the structural member at its ends. Changing both the end restraints against bending in the north-south direction would bring the frequency of mid-span vibration to 40 Hz, away from the running frequencies of the fuel gas compressor. By changing one of the end conditions to fixity, the corresponding frequency of vibration would be 28 Hz, again separated from twice the running frequency of the compressor. It was recommended that achieving end fixity to the south diagonal brace was by far the suitable [sic] and permanent way of minimizing the mid-span vibration of the member. They recommended that in the first instance a gusset restraint should be added to one end of the diagonal brace at 90 [degrees] to the existing gusset, shaped and welded to the outside of the tubular [member] and supporting deck beam. The new gusset would have to be installed in two pieces and if possible welded to the existing gusset. If the vibration was still unacceptable a second gusset should be installed at the other end of the member.

I note in passing that a resonant frequency of 12.6 Hz was regarded by Snamprogetti (in the person of Mr. Natarajan) as undesirable (though not, apparently, one of the main disadvantages of water filling), whereas a frequency of 14.1 Hz was apparently regarded by Mr. Natarajan as acceptable in his initial dynamic analysis carried out before the detailed drawings of the structure had been developed. 12.6 Hz is proportionately further from the fundamental frequency of the fuel gas compressor than the frequency of the first harmonic is from 14.1 Hz. However, the force amplitude of the first harmonic of the fuel gas compressor is only about 30 per cent. of the force amplitude of the fundamental. This point has not been explored.

Phillips instructed ATL to carry out a further vibration survey. That survey was carried out from 16^th to 18^th March 1996. ATL reported on 19^th March 1996. The survey was conducted when the fuel gas compressor was running on load and after the south diagonal brace had been filled with water. Also, as the report states, the knock out drum drain line support had been improved and the thermo-syphon vent tank pipework had been braced to the compressor suction line. The report included the following statements. The impact response test on the south diagonal brace showed that filling the tubular brace with water had lowered the fundamental natural frequency from 19 Hz (empty) to 13.5 Hz (full of water). That had reduced the vibration level in the north-south direction from 19.5 to 3.5 mm/second. The bracket added to support the knock out drum drain had reduced the vibration level in the vertical direction from 134 to 2 mm/sec. Apart from those items, the vibration levels were generally similar to, or higher than, those measured previously with the fuel gas compressor working off load and prior to the modifications. Of the items tested, the report mentioned some which were subject to relatively high vibration levels. They were the LSLL-903 level switch, the fuel gas filter, the [fuel gas] compressor suction line, the thermo-syphon, the LSHH-915 level switch and, on the gas turbine compressor set, the fuel gas solenoid valve, the turbine lube oil vent and the turbine air inlet filter framework. Though not impact tested, those items probably had natural frequencies close to the excitation frequencies associated with the fuel gas compressor.

None of the items stated as having been subject to relatively high vibration levels were part of the structure of the platform. It is clear from an examination of the detailed results set out in the report that in the eleven locations on the structure of the platform (other than the south diagonal brace) tested, the only significant changes were to the vibrations of the north diagonal brace and of the tubular north-south diagonal brace. Those changes represented in both cases increases in the overall levels of vibration, but not to an extent which gave rise to any complaint. Dr. Walker did, however, mention (Day 51, p.81) the increase in the vibration level of the north diagonal brace as an illustration of the complex inter-relationship between the members of the structure. He attributed that increase to the filling of the south brace with water. I accept that that must have been at least a major cause of the change. The load on the compressor may have contributed.

Mr. Mobbs was on the platform when the test was carried out. He gave this evidence about it (Day 19, p.97):

I am satisfied that after the south diagonal brace was filled with water the structure was adequately designed and built to accommodate the dynamic loads imposed on it by the fuel gas compressor so as not to vibrate unduly. However, the pipework, vessels and instruments on the fuel gas skid and the turbine set in some instances vibrated to an unacceptable extent.

The minutes of a meeting between the parties held on 2^nd May 1996 record that there were still problems with vibration on the platform but the main problem related to instruments. Phillips undertook to provide Snamprogetti with a copy of the ATL report of 19^th March. It was not until 15^th May 1996 that Phillips sent to Snamprogetti a copy of that report. In his covering letter, Mr. Rayner asked Snamprogetti to review the data contained in the ATL report against their own vibration analysis, including all associated equipment and instrumentation, and to advise what actions were necessary to reduce the levels of vibration to within acceptable limits.

On 25^th May Mr. Natarajan sent his report to Snamprogetti in London. He recommended changing the end condition of the north diagonal member of the structure. As to the equipment and instruments mounted on the fuel gas skid, he said they were not included in Snamprogetti’s vibration analysis model. He suggested that a comprehensive review of the equipment mounted on the fuel gas skid should be carried out in consultation with the FG skid manufacturer [a reference to George Meller Limited] to arrive at permanent measures to minimize the present high vibration levels. He added that adequate clamps/support schemes must be developed to minimize vibration levels on the turbine enclosure and on the turbine air inlet filter element mesh.

The Baker Jardine HAZOP took place from 20^th to 24^th May 1996. The report of that HAZOP was dated 28^th May. Among the matters stated in the report were the following. There had been a number of operating problems with the equipment due both to changes in the operating conditions from the original design data and also, possibly, due to deficiencies in the design of the package. The timing of the HAZOP had been intended to allow the action recommendations to be considered and acted upon during the summer, when no compression was required, so that the platform would be fully operational in time for the start of the winter operation mode. There were sixty-nine recommendations. The number of recommendations was higher than would normally be expected for a plant that had been installed and commissioned. The majority of the issues concerned minor changes to the plant which were unlikely significantly to affect the design or cause major modifications offshore. There were, however, several significant issues concerned both with the safety and with the philosophy of the design and construction. The overall safety and operability of the design would depend on their satisfactory resolution.

In relation to the vibrations caused by the fuel gas compressor, recommendations were contained in a HAZOP action response sheet. The meeting giving rise to it was held on 22^nd May, and the response sheet was dated 23^rd May. The response sheet stated that remedial work had reduced vibration, but some instruments and nozzles were still vulnerable. It stated that two reports had been written by ATL. The recommendations were as follows (action required of those whose abbreviated names appeared in brackets):

There is a further internal memorandum of Snamprogetti dated 13^th June 1996 which covers four HAZOP action numbers, including number 2.3-11, which relates to the vibrations. It was written by Mr. A. Fletcher. The relevant part of that memorandum repeats the substance of what Mr. Natarajan had said in his memorandum of 25^th May. The memorandum of 13^th June was apparently attached to a copy of the HAZOP action response sheet mentioned above. On that copy, under the heading ‘Action resolution’, the following is written in manuscript:

There is no evidence whether Mr. Natarajan’s memorandum of 25^th May or Mr. Fletcher’s memorandum of 13^th June was communicated to Phillips. However, as Mr. McMaster indicated in the course of cross-examining Mr. Middleton (Day 56, p.153), it can be assumed that the memorandum of 13^th June was so communicated since it was attached to the HAZOP action sheet.

The following exchange took place later in the same cross-examination (Day 56, page 173):

I reject Mr. McMaster’s analysis of the situation. The HAZOP action response sheet recommended three courses of action. First, item 1), Snamprogetti had to review the second ATL report and advise on remedial measures. The only remedial measures which they could specifically advise related to the north and south diagonals, where they recommended changing the end condition, as they had before. As to the equipment and instruments, they recommended that a comprehensive review should be carried out in consultation with the manufacturers (George Meller). They did not advise Phillips against going to ATL, or discourage them from doing so. Indeed, consulting ATL was an obvious way of conducting a comprehensive review. There was nothing to prevent Phillips from asking Snamprogetti to take part in such review. Recommendation 2) of the response sheet clearly left it to Phillips to decide whether to obtain specialist advice.

A meeting was held between Phillips and Snamprogetti on 20^th June 1996. The minutes record that it was decided that the fuel gas compressor should be treated as lower priority work. However, a summary of problems was needed so that George Meller could be contacted for a tripartite meeting later. That was to be actioned by Snamprogetti. Thus it appears that Phillips did indeed intend Snamprogetti to be involved in the comprehensive review.

Dr. Walker calculated the expected fatigue life of certain welds on the discharge bottle (also called the fuel gas receiver), an item of equipment on the fuel gas compressor skid. He applied finite element analysis, using the vibration levels reported by ATL on 19^th March 1996, to calculate the cyclical stresses. The bottle stood on a cylindrical skirt which was welded to a horizontal flange at the bottom. The welds which he considered were that weld, the weld joining the bottle to the skirt, and a weld joining to the bottle a nozzle attached to piping supporting the LSHH-915 level switch. In the case of the first weld, he calculated that a fatigue crack would begin to develop after four days of operation (and would rapidly develop). In the case of the weld attaching the bottle to the skirt, he calculated a fatigue life of 8 months, which would be reduced if the bottle had been inaccurately positioned. In the case of the nozzle, the fatigue life he calculated was 30 hours.

Mr. Middleton considered that the finite element analysis could not properly be used because it assumed that the foundation on which the bottle was mounted was not itself vibrating. I accept that as a theoretical objection. But in my judgment Dr. Walker’s assumption was a reasonable one for the purpose of carrying out an approximate calculation. In relation to the nozzle, Mr. Middleton also made the point that no account had been taken of possible movement of the bottle at the root of the nozzle. Moreover, one had to allow for an unknown phase difference between the oscillations in the nozzle and the oscillations in the bottle. Those points cast doubt on the accuracy of the calculation, but in my judgment the principal objection to Dr. Walker’s analysis is that he assumed that the support of the level switch was perfectly rigid. I cannot accept his calculation of the fatigue life of the weld at the nozzle as being necessarily even approximately correct. Nevertheless, the pipework and instruments on the fuel gas compressor skid mentioned in the ATL report of 19^th March as showing relatively high vibration certainly required treatment so as to reduce their vibration to acceptable levels. So did the two items on the gas turbine skid mentioned in the report.

The ATL report also showed that there was significant vibration of the casing of the fuel gas compressor and of the south, north and middle feet of the fuel gas compressor. The root-mean-square [(rms)] velocities recorded by ATL at those points were up to, but did not exceed, 9.0 mm/second. The level of those vibrations, though outside the limits applied by Snamprogetti in the design of the module, fell well within those stated in a standard, VDI (Verein Deutches Ingenieur) 2056. That standard, which is dated 1964, relates to the vibrations of machines. It divides machines into groups. Group D is machines and motion parts having mass effects which cannot be balanced, and installed on high-tuned mountings (rigidly mounted). The standard says (p.11) that for Group D, root-mean-square velocities of 20 to 30 mm/second, and in certain cases higher than that, may arise without any complaint necessarily resulting. Group S includes machines and motion plants [sic] having mass effects which cannot be balanced, and installed on low-tuned mountings (flexibly mounted). Closely similar standards apply in British Standard 4675 : Part 1 of 1976. The comment in relation to Group S (Class VI in the British Standard) is clearer in the language of the British Standard. It reads, so far as material, ‘The resiliently mounted machines in Class VI permit a greater tolerance [sc., than Class V, which corresponds to Group D] in this respect. There is an isolation effect and the forces transmitted by the mounting into the surroundings are small. Under these circumstances vibration levels measured on the machine side of the mounting system are greater than those measured when the machine is fastened to a large relatively rigid support’. BS 4675 : Part 1 was superseded by BS 7854 : Part 1 in 1996.

Dr. Walker considered that Group D was not applicable because the fuel gas compressor was not mounted on a rigid foundation. The lightweight steel frame cantilevered off the main platform provided a flexible support to the single-cylinder reciprocating compressor. I reject that evidence. Though the fuel gas compressor was not on a block of concrete, it was rigidly mounted. It is true that it was attached to a module that had a degree of flexibility. In any case, the less rigid the mounting, the greater the speed of vibration allowed by the standard. Dr. Walker said that Group M applied. But Group M applies to machines ‘having only rotating parts’, and thus excludes reciprocating machinery.

Dr. Walker considered that BS 7854 was more relevant than VDI 2056 to the vibration on the fuel gas skid. BS 7854 has an apparently more rigorous standard limiting the permitted root-mean-square vibration velocity to 2.8 mm/second. However, that British Standard does not apply specifically to reciprocating machinery. Class II, on which Dr. Walker relied, is similar to Group M in VDI 2056, though the express restriction to machines having only rotating parts is absent. Electric motors of 15kW to 75kW output are said to be typical. It is obvious that such machines are likely to produce less vibration than reciprocating machines, as appears to be recognized by VDI 2056. BS 7854 states that it is a basic document which sets out general guidelines for the evaluation of mechanical vibration of machines; that it is intended that evaluation criteria for specific machine types will be provided for different machine types; and that the limited criteria provided are a short-term expedient only.

Dr. Walker considered that the electric motor fell within Class II of BS 7854. The root-mean-square speed of the vibrations on the south foot of that motor was 4.25 mm/second as revealed by the ATL report of 19^th March 1996, and he concluded that they thus exceeded the permitted level of 2.8 mm/second. However, it follows from the vibration spectra and figures reported by ATL that the greater part of the vibrations of the electric motor was caused by the reciprocating compressor, and not by the motor itself. BS 7854 states at the beginning, under the heading Scope, that the evaluation criteria relate only to the vibration produced by the machine itself and not to vibration transmitted to it from outside. Thus I reject Dr. Walker’s evidence on this point. The vibrations of the fuel gas compressor and of the electric motor were not excessive.

Mr. Thomson, while agreeing that Phillips were entitled to expect that at the time of commissioning there would not be a serious vibration problem to resolve, gave evidence that it is usually a simple matter to prevent the instruments vibrating. He mentioned bracing (Day 37, pp.51, 47).

Mr. Middleton considered that a proper program of modifications to pipework and instruments on the fuel gas compressor skid would have been likely to be successful. He mentioned three methods of reducing the vibrations to a harmless level: bracing, stiffening and application of inertia dampers. He gave an illustration of bracing together attachments to the fuel gas compressor receiver to neutralize the vibrations of level switch LSHH-915. He envisaged a stiff connection between two nozzles attached to the receiver a quadrant apart. If properly designed, it would take the resonant frequency away from the excitation frequencies (Day 54, p.93 et seq.; Day 56, p.110). Dr. Walker said (Day 51, p.7) that an object should not be braced to something that is vibrating at an unacceptable level. I do not accept that that is an absolute prohibition, since the bracing can affect the resonant frequencies of the system.

Mr. Middleton described the inertia damper as an additional mass/spring system attached to the vibrating component. By choosing the right characteristics very large reductions of vibration of the basic system could be achieved.

Dr. Walker gave evidence that the elimination of vibrations in the structure and equipment would have required a lengthy and carefully planned program of technical assessment before any solution at all could be designed. He estimated that to carry out the necessary survey, numerical modelling, design and modifications was likely to take six to eight weeks. Dr. Walker included the structure of the module in that estimate. He considered that even the pipework and instruments on the fuel gas compressor skid should not be rectified on a piecemeal basis, since a change to one part of the system would be likely to affect other parts.

Mr. Middleton gave the following evidence (Day 54, p.100 et seq.):

Although Mr. Middleton considered that ATL had probably identified the worst cases of vibration, he thought that Phillips should have instructed a firm like ATL or Mitsui Babcock to do a much more comprehensive survey than had been carried out. It was put to him in cross-examination that it might have been impossible to find a firm that could carry out the work without delay. He said that he knew that ATL were geared up to moving into action pretty quickly. From what he knew of the way ATL operate, they would probably be out on the platform measuring within two weeks of being instructed:

The pipework on the fuel gas compressor skid was designed and supplied by George Meller. Snamprogetti had laid down a specification for the fuel gas compressor. That specification included a seventeen-page document entitled ‘General specification for piping design in mechanical packages for offshore facilities’. It is not clear that the precise document is before the court, since the first page (sheet 1 of 17) has the reference Specification No. GSP/ENDP/5962, whereas the pages that follow (sheets 2 to 17 of 17) have the reference Specification No. E-99-S-P-0014. No point has been taken about that. I shall assume that the document, if not the actual specification, is substantially the same as the actual specification in all relevant respects. The specification provided for the use of gusset plates where fatigue stress due to vibration could cause failure of a pipe branch at the point of attachment. It provided that piping components and pipework should be designed and installed so that their natural vibration frequencies in operation were not excited by a regular forcing frequency from other system equipment such as a pump or fan. In case of dispute, the principal natural frequency of a suspect piping installation as determined by calculation or measurement and vibration frequency analysis on the fabricated system should be at least 20 per cent. removed from any significant coupled exciting frequency.

The following evidence was given by Mr. Sylvester-Evans (Day 70, p.38):

As regards the effect of vibrations on human beings, Mr. Morrell gave evidence about his experience when he was on the platform for the first time in February 1996. At that stage, the gas turbine, the main compressor, the after-cooler and the fuel gas compressor skid had all been installed. Some connections had been made, but not all the interfaces had been completed. When the fuel gas compressor was started, he was alarmed at the extent of what he called the vibration problem. He could feel the vibrations. They affected the skid itself, walkways, handrails and other things. At one point he stood on a walkway approximately 80 feet from the fuel gas compressor, off the deck itself, and still felt the vibrations. His entire body was moving, though it was not bad enough to give blurred vision. His major concern was with the pipes and instrument panels. He said that the south-west diagonal brace on the main compressor skid was resonating very badly. You could physically see it moving about. I accept that evidence of Mr. Morrell. But it is worth mentioning that the maximum vibration double amplitude (i.e., peak to trough) of any member of the structure whose vibration was measured by ATL on their first visit, namely that of the south diagonal brace, was less than 0.6 mm.

Vibrations measured by ATL on the main original structure, i.e. off the compressor module, were negligible.

The Department of Energy publishes a document entitled Offshore Installations: Guidance on design, construction and certification. Appendix A52 to the fourth edition of that document deals with noise and vibration. Vibrations are divided into five categories. They are as follows:

Dr. Walker considered the vertical components of the vibrations. From the accelerations shown, the corresponding velocities can readily be calculated. They were independent of frequency for frequencies of 8 Hz and above. For lower frequencies, higher velocities were permissible. Dr. Walker ignored that latter fact, though some of ATL’s spectra showed significant contributions from lower frequencies. Nevertheless, I accept Dr. Walker’s approach as a reasonable approximation, which is all that he claimed for it. In particular, I accept that he was right to calculate the rms velocity over the whole range of frequencies studied by ATL (stated in the text of the report of 19^th March 1996 to have been 0 to 100 Hz, but shown in the graphs of the spectra from 3 to 50 Hz only). Dr. Walker identified 13 locations tested by ATL which were in areas which he considered it was reasonable to expect that operatives might be likely to work. On the basis of the velocities reported by ATL on 19^th March 1996, two of those locations fell within category I. Eight fell within category II. Two fell within category III. One fell within category IV.

The Department of Energy guidelines state in paragraph 52.4.3:

The largest horizontal components of the vibrations at any of the 13 locations selected by Dr. Walker fell well short of the upper limit to category II for horizontal vibrations.

Since the compression facilities on platform 52/5A were not intended to be used continuously throughout the year, much of their maintenance could be carried out when the facilities were not working and when there would be no significant vibration. It is by no means clear that in relation to their effect on human beings the vibrations measured by ATL were unacceptably high in the circumstances. The material before me is insufficient for me to make a finding on the point.

On the totality of the evidence, I find that excessive vibrations could have been rectified within six to eight weeks of the taking of a decision to instruct a firm such as ATL or Mitsui Babcock to carry out the necessary survey and design work. That includes the time taken to find a firm willing to undertake the work. The work in question includes not only the work necessary to rectify the vibrations caused by the fuel gas compressor to the pipework and instruments and to things on the main compressor skid but also any work found necessary to rectify the vibration of any walkways, gratings or other places where operatives or maintenance engineers would have to work. At the material time, however, Phillips could not have been certain that such vibrations could be rectified. It would have taken them at most six to eight weeks to be certain.

I consider now the question whether compression of the fuel gas was necessary. The gas turbine driving the main compressor required fuel gas at a pressure between 188 or 190 psig and 250 psig. The fuel gas was Zechstein gas taken from a point a short distance above the confluence of the Upper Bunter and the Zechstein gas, and thus at a pressure only a few pounds per square inch above that at the main compressor discharge. The main compressor had to be capable of providing a discharge pressure of 143 psia in 1995, falling to 100 psia in 2003 to 2005. It is manifest that on the basis of that design, a fuel gas compressor was necessary.

Mr. McMaster submitted that an alternative design could have been effected. The shut-in wellhead pressure of the Zechstein gas was over 1000 psig in 1994 and 1995. Fuel gas could have been tapped off at a suitable pressure near one of the Zechstein wellheads, upstream of the confluence of the gas flowing from that wellhead with the gas flowing from the other Zechstein wellheads. A pressure reduction valve downstream of the tap-off point but upstream of that confluence point would prevent the high-pressure gas from reducing the flow from the other Zechstein wells. It would, however, reduce the flow of gas from the wellhead in question. Such a system was described by Dr. Robinson. I am satisfied that it would have worked in principle. It would have required the use of a fuel gas heater, since the driving turbine of the main compressor required superheated gas. No such heater was required with a fuel gas compressor, since the compression superheated the gas.

I am satisfied that the option of using a fuel gas heater was discussed by representatives of Snamprogetti and Phillips. Mr. Tomlinson gave evidence to that effect, though he did not start work on the project until later (Day 40, pp.116, 117, 121). In an early version of the Hewett/Bacton offshore compression study report, dated 9^th July 1993 and issued for the approval of Phillips, in sub-section 2.2.1, and repeated in the final issue of the report dated 1^st October 1993, in sub-section 2.3.1, there appears this passage:

Thus the system discussed was not the same as that described by Dr. Robinson. Dr. Robinson’s system would not work if the fuel gas were tapped off below the confluence point of the gas from the individual Zechstein wells. That would inevitably be so if the gas were tapped off below the Zechstein gas metering skid. But the requirement or putative requirement to take the fuel gas from downstream of the metering skid was changed in 1994. The gas for the fuel gas compressor was in fact tapped off upstream of that skid.

In the final issue of the Hewett/Bacton Offshore Compression Study Report, dated 1^st October 1993, it is stated in sub-section 1.3.3 :

In his original expert’s report, Dr. Robinson said :

In oral evidence, he corrected the last sentence of the first of the two paragraphs from which I have quoted above, so as to read:

Mr. Green, in a draft witness statement which was in evidence, prepared on 12^th March 2002, the day before he died, said this:

It is implicit in that evidence of Mr. Green that Phillips had, at any rate by implication, required Snamprogetti to take the supply of the fuel gas from a point downstream of the metering skid. The existence of that requirement on the part of Phillips was challenged in cross-examination of Mr. Tomlinson. The following exchange between Mr. McMaster and Mr. Tomlinson took place (Day 40, pp.109 to 113):

I am not satisfied that Phillips specifically imposed a requirement on Snamprogetti that the fuel gas had to be taken off downstream of the metering skid (though it is clear that Phillips were aware of Snamprogetti’s intention to do so). Subject to that, I accept all of the evidence of Mr. Green that I have quoted above. In particular, I accept the figures of reduction in production that the use of two dedicated Zechstein wells for fuel gas would have entailed. Dr. Robinson was not in a position to give any such figures.

It was put to Mr. Tomlinson in cross-examination that a fuel gas system without compression, for example a system using a heater, would have “won easily” against a compressor on points of low weight, small footprint and low power consumption (Day 40, p.115). Mr. Tomlinson agreed. I accept that such a system would have had those considerable advantages.

I am satisfied that a fuel gas system of the kind described by Dr. Robinson was not overlooked by Snamprogetti. In or about 1991 Mr. Green himself was the lead process engineer in relation to compression facilities on platform 48/29A. He said in his first witness statement (signed on 7^th December 2001) with reference to that platform that as there was a source of high pressure fuel gas, there was no need to instal a fuel gas compressor. In his submissions, Mr. McMaster said that the fuel gas system designed by Snamprogetti for the 48/29A platform used two wells configured in accordance with the layout shown in the design that Dr. Robinson drew in the course of his oral evidence. Moreover, the design that Snamprogetti produced in 1996 for the fuel gas system on platform 52/5A involved the dedication of two wells.

I find that Snamprogetti reasonably rejected the use of a fuel gas heater in place of a fuel gas compressor. The matters mentioned in paragraphs (a) and (b) of Mr. Green’s evidence quoted above were sufficient reasons for such rejection. That is so notwithstanding the advantages of such a system.

Given that the avoidance of fuel gas compression was not an option, in my judgment Snamprogetti had no alternative to the single-cylinder fuel gas compressor that they recommended, in the sense that they reasonably rejected all the others. Even if one of the others might reasonably have been chosen, I am satisfied that Snamprogetti were not in breach of the duty they owed to Phillips in recommending the single-cylinder compressor. It is not suggested that they should themselves have designed a compressor.

In their particulars of claim, Phillips pleaded that by reason of the vibration problem, the fuel gas compressor was ‘perforce’ abandoned. It was not perforce abandoned. It was abandoned because in the event, unforeseen by either party at the material time, it was never needed. (Phillips’s pleading was indeed amended. By re-amendment, it was pleaded that the claimant decided to instal a fuel gas heater because it was unlikely that works to remedy vibration problems or other remedial works for the fuel gas compressor system could be completed before the end of October 1996, if at all. The fuel gas compressor was eventually abandoned and the heater used in its place. The claimant did not see fit to incur further cost and effort in attempting to resolve the vibration problems. In that the compressor would never be used for the purpose for which it was designed it was perforce abandoned).

The Basis of Design was prepared by Snamprogetti from information provided by Phillips. That included the suction and discharge pressures required of the main compressor over the stated years. Following review by Phillips, the Basis of Design was revised and issued on 3^rd June 1994. Snamprogetti are not criticized in relation to the preparation or content of the document, but there is a dispute as to its meaning. Mr. McMaster submitted that it showed a requirement that the main compressor should be capable of producing an outlet pressure of about 225 psi (whether 225 psia or 225 psig does not appear, and is not material for present purposes). Miss Boswell submitted that it showed a requirement that the main compressor should be capable of producing an outlet pressure of 143 psia, but no greater requirement.

The Basis of Design says in section 2.2:

There is only one table on sheet 5. It shows compressor outlet pressures for the 3000 HP case (the case adopted) varying between 143 psia in October 1995 to 100 psia from 2003 to 2005, as the Upper Bunter shut-in wellhead pressure is predicted to diminish over the period.

Dr. Robinson said (in his first report) that those figures did not represent limits, but were simply typical. The table did not define the operating envelope. That could be seen by reference to some figures given on sheet 11 in section 4 of the document, headed “Operating conditions”. Those figures showed a flowing wellhead pressure for Zechstein gas of 200 psig forecast for 1995. The main compressor must give a discharge pressure to blend Upper Bunter gas into the Zechstein gas. The pressure drop between the wellhead and the point of confluence was likely to be less than 20 psi. Thus the compressor must be capable of discharging at about 180 psig.

In his oral evidence (Day 63, p.48) Dr. Robinson expressed the opinion that the flowing wellhead pressures of Zechstein gas shown in section 4 corresponded to rates of flow set out in section 2.3. On that basis, the flowing wellhead pressure of 200 psig corresponded to a rate of flow of 32 mmscf/d. Some figures in a table on sheet 8 of the basis of design were put to Dr. Robinson in cross-examination. Those figures, contained in a table said to show capacities by reservoir, showed a capacity of the Zechstein wells on 52/5A predicted to be about 55 mmscf/d in 1995. Dr. Robinson accepted that that could indicate a flowing wellhead pressure lower than 200 psig. But it could be that the wells had been re-calibrated since the production of the figures in section 2.3. It was put to him that what was contemplated when Zechstein gas was required for maximum nominations was that there would be a very high flow of Zechstein gas, which meant inevitably that that Zechstein gas would be flowing at a much lower pressure than the pressures set out on sheet 11. He accepted that that appeared to be a reasonable interpretation (Day 63, p.52).

The relevance of the figures on sheet 11 to the compression facilities is not apparent if Snamprogetti’s case is right. The front page of the Basis of Design states that the document relates to compression facilities at Hewett 52/5A. But the introduction states that it defines the basis for the offshore compression facilities to be installed on platform 52/5A and for modifications to the Bacton compression, amine and sulphur recovery facilities. Thus the figures may be relevant to the situation where the main compressor on platform 52/5A is not working. Nevertheless, I regard the figures on sheet 11 as an item of evidence in support of Mr. McMaster’s contention. I return to this point below.

Two graphs were attached as Appendix II to the Basis of Design. Mr. McMaster relied on a curve drawn in manuscript on one of those graphs as showing that the H line would be subjected to pressures of up to 225 psi and that accordingly the main compressor would have to be capable of discharging Upper Bunter gas at that pressure. His argument depended on an analysis of another document appended to the Basis of Design, called the flow strategy. That document appears as Appendix III and is headed Flow Strategy for Winter 1993/94. It is introduced in section 2 of the Basis of Design, which says that it is the flow strategy used in the design.

One of the graphs in Appendix II relates to the 30-inch A line, the other to the 30-inch B line. They are both graphs of FTP pressure in psi (the ordinate) against flow rate in mmscf/d (the abscissa). FTP refers to the field terminal platform, also known as the 48/29A platform. Each graph shows four members of a family of curves. Each such curve is a plot of entry pressure of gas into the 30-inch line at the FTP against the flow rate of that gas. For any given curve, the lower the flow rate, the less the pressure. The pressure at zero flow rate appears as the intercept of the curve with the ordinate. Since with zero flow rate the pressure along the line would be uniform, the intercept represents the pressure on arrival at Bacton of the gas to which the curve applies. Thus the curve can be used to determine what pressure is required at the FTP for any given flow rate of gas to arrive at Bacton at the pressure represented by the intercept. The four curves are almost straight lines and almost parallel, especially at flow rates above about 150 mmscf/d. Especially below that pressure, they diverge. Because of that departure from linearity, it makes a difference whether the pressures shown are in psia or in psig. It is not stated which they are.

Drawn in manuscript on each of the two graphs that I have described are two curves. They show increasing pressures for reducing flows. The curves are roughly parallel. There exists a range of flow rates common to both curves. Within that range, at any given flow rate one of the curves, which I shall call the upper curve, shows a higher pressure than the other. Confusingly, the upper curve on the graph relating to the A line is marked “Curve 2”, and the lower is marked “Curve 1”, whereas on the graph relating to the B line it is the other way round. Mr. McMaster relied on the lower curve, Curve 2, on the graph relating to the B line.

Appendix II is introduced in section 4.2 of the Basis of Design in the following terms:

There was a compressor on the FTP known as the alpha compressor. That was the compressor giving the “49/29A” (i.e., 48/29A) compression mentioned in section 4.2 of the Basis of Design quoted above. At the material time, the alpha compressor fed gas only into the A line, though the system was later altered so as to be capable of feeding gas into either the A line or the B line. Curve 1 on the graph for the A line shows a pressure at the FTP (i.e. at the discharge of the alpha compressor) of approximately 190 psi at a flow rate of 200 mmscf/d. I call that the starting-point of Curve 1. Curve 2 on the A line graph starts at a pressure of 250 psi and a flow rate of about 240 mmscf/d. Thus the text explains that Curve 1 shows, as a function of flow rate, the pressure to which the alpha compressor is to compress the gas from 75 psig. And it explains that Curve 2 shows, as a function of flow rate, the pressure to which the alpha compressor has to compress the gas from 95 psig. It is also relevant to the present issue that it was contemplated that where 52/5A gas was to be fed to the alpha compressor, the suction pressure would be 95 psig (110 psia). That would be a few pounds per square inch less than the discharge pressure required of the main compressor on the 52/5A platform in those circumstances.

Manuscript Curve 1 on the graph for the B line (this time, be it remembered, Curve 1 is the upper curve) starts at approximately 265 psi. and 200 mmscf/d. (It thus appears from the text that the FTP pressure figures shown on the ordinate were being interpreted as representing pressures in psig). Curve 2 starts at about 136 psi and a flow rate of about 155 mmscf/d. The pressure remains constant down to a flow rate of about 130 mmscf/d, and at that point starts to rise to 250 psi at about 40 mmscf/d. There being no compression at the FTP for gas flowing into the B line, those figures indicate an H line pressure at the 52/5A platform of a few pounds per square inch over 136 psi at a B line flow rate of 155 mmscf/d rising to a few pounds per square inch over 250 psi at a B line flow rate of 40 mmscf/d. Where Upper Bunter gas was to be used, the discharge pressure from the main compressor on the 52/5A platform would have to be slightly above the H line pressure. It was upon those figures that Mr. McMaster’s argument depended. The starting-point of the lower manuscript curve on each graph was marked with the date 1995 and shown as moving with time towards lower flow rates in the general direction of the family curves.

Further light is thrown on the graphs by a document prepared by Snamprogetti entitled Operating Philosophy, which relates to the compression facilities at the Bacton terminal. The same graphs appear as attachment 1 to that document. The flow strategy appears as attachment 2. The version of the document prepared for review by Phillips was dated 27^th May 1994; revision 1, made apparently in the light of Phillips’s comments, is dated 17^th January 1995. They do not differ in any respect relevant to the present issue.

The following is stated by way of explanation in section 3.1 of the Operating Philosophy, at sheet 8. Section 3.1 relates to offshore commingled operation:

In section 3.2, relating to onshore commingled operation, it is said at sheet 10:

It will be noted that the pressures on the graphs are here said to be psia, not psig as in the Basis of Design.

The Operating Philosophy summarizes the position in relation to the onshore commingled operation in section 4.2 at sheet 15:

Thus both the Basis of Design and the Operating Philosophy contemplate that the lower manuscript curve on the B line graph reflects the shutting down of the lower-pressure wells as the flow falls. For reasons which appear from what follows, I reject an argument of Mr. McMaster to the contrary.

When there is a pressure gradient of gas in a pipeline, the gas flows from the point of higher pressure to the point of lower pressure. The higher-pressure gas drives the lower-pressure gas in front of it, thereby losing some of its potential energy. I am satisfied on the evidence of Dr. Robinson that for any given rate of flow of gas in a given pipeline, the higher the average pressure of the gas, the less is the loss of power in that way. Thus in the interests of economy it is desirable to have as high a pressure as possible for gas entering the A line or the B line at the FTP.

Moreover, the lower the pressure of the gas on arrival at Bacton, the greater the amount of energy required to compress it to the pressure (about 1000 psi) required for delivery of the gas to British Gas. Thus it is desirable to minimize loss of the pressure at which the gas emerges from the wellhead.

There is provision for control of the flow of gas from each well by means of a control valve near the well. If the valve is adjusted so as to reduce the flow of gas, pressure will build up upstream of it until the rate of flow of gas from the well is reduced to match the flow through the valve. In view of the large volume of gas downstream of the valve, the adjustment of the valve will have no immediate discernible effect on the pressure downstream of it. Thus the adjustment of the valve in those circumstances will cause a loss of potential energy as the pressure of the gas passing through the valve falls. That loss is irrecoverable. If, instead of using the valve, the pressure downstream of it were increased to a level slightly below the upstream pressure, such as to allow the same flow, that energy would not be lost. That conclusion was reached by Dr. Robinson by a slightly different process of reasoning. Thus it is more economic to control the flow of gas by means of the pressure of gas in the sea line than by using flow control valves. Dr. Robinson said that the valve should be viewed as a device for minor adjustment of flow, to balance one well against another. It is not intended as the primary means of controlling pipeline flows. The production objective is to maximize the flow of gas from each well in use, in such a way that the pipeline pressure is maximized. In that way the need for gas at any time can be met, whilst minimizing the use of energy at Bacton. (Dr. Robinson’s second supplementary report, 28^th August 2002, paragraph 2.22). I read his reference to maximizing the flow of gas as meaning minimizing the reduction of flow caused by use of flow control valves. On that reading, I accept that evidence as representing the ideal state of affairs from the point of view of economy. But it does not represent the state of affairs envisaged by the figures in the flow strategy, on which Mr. McMaster’s argument depends.

Dr. Robinson explained the association of high pressure with low mass flow shown by the lower manuscript curve on the graph for line B as follows. One starts with a steady state, i.e. a situation where the mass flow is constant and the pressure at any given point in the system is constant. The mass flow of gas into the system from the wells equals the mass flow out of the system at Bacton. If the take of gas at Bacton is reduced, the mass of gas in the pipelines upstream of it will increase, since the rate of influx is initially unchanged. That causes an increase of pressure everywhere in the system, including the back pressure at the wellheads. That in turn causes a reduction in flow from the wells. In time, a new state of equilibrium will be reached at a lower mass flow and higher pressures than before. The same applies, mutatis mutandis, if the take at Bacton is increased. It is not desirable, said Dr. Robinson, for reasons of economy, to make the changes at the wellhead end of the system by means of flow control valves (save for minor adjustments). Again, I accept that evidence as representing the ideal state of affairs.

Whilst Dr. Robinson’s explanation of the shape of the lower manuscript curve in the graph of line B is not the same as that given in section 3.2 of the Operating Philosophy quoted above, it is not inconsistent with it.

The flow strategy shows Upper Bunter production at 13 mmscf/d, to be fed into the B line, when the B line is passing a total of 67 mmscf/d. Therefore, said Mr. McMaster, reading off from the lower manuscript curve on the graph for the B line, the pressure in the B line at the FTP has to be 200 psi. Indeed, the Upper Bunter gas should start to flow when the flow in the B line is 54 mmscf/d. That implies a pressure of about 217 psi in the B line according to the curve on the graph. The main compressor on the 52/5A platform would have to discharge gas at a few pounds per square inch above those pressures.

The flow strategy shows a constant feed into the B line of 39 mmscf/d from wells 48/30-9 over a range of flows in the B line from 54 to 157 mmscf/d. Those flows in the B line correspond to pressures ranging from 217 psi down to about 136 or 138 psi according to the curve. It is manifest that a constant flow over such a range of pressures requires more than tweaking of the control valves for those wells. On Dr. Robinson’s scenario, the pressure would have to be substantially constant over the whole range of flows in the B line in order to ensure a constant flow from wells 48/30-9. That constant pressure would also have to sustain variable flows from two of the Zechstein wells.

Moreover, the flow strategy, at any rate in so far as it relates to the B line, does not contemplate turning off the lower-pressure wells first as the flow diminishes. (The flow strategy being applicable to winter 1993/94 clearly is not based on use of the 52/5A compressor).

Thus the explanations of the lower manuscript curve in the graph for line B given in the Basis of Design and in the Operating Philosophy and by Dr. Robinson appear not to accord either with figures in the flow strategy or with the order of shutting-down the wells shown in that strategy. That curve may be an ideal curve for economical use of the B line, but in my judgment it is unjustifiable to draw from it the conclusion that the main compressor on the 52/5A platform was required to discharge Upper Bunter gas at pressures higher than those stated in the table on sheet 5. I am fortified in my conclusion that the lower manuscript curve may be an ideal curve by the terms of the note accompanying the marked-up graph when, on 21^st March 1994, it was sent by Phillips to Snamprogetti. That note reads, so far as material, “Please find curves showing ideal max. pressure at FTP as a function of rate”.

The normal method of operation described in the Operating Philosophy is that the pressure rises as the flow requirement falls, but it assumes that the low pressure wells are shut down as the flow requirement falls. The low pressure wells must by definition surely include the Upper Bunter wells, notwithstanding the proposed 52/5A main compressor. Indeed, it was contemplated that the 52/5A compressor would not be used in the summer, when demand is less than in the winter. If the lower pressure wells could not be cut out, then the H line pressure would be 138 psia, consistent with a discharge pressure of 143 psia for that compressor. The pressure of 138 psia is also consistent with a flow in the B line of 150 mmscf/d, the figure mentioned in the Operating Philosophy as the normal rate.

There was put before me an internal document of Phillips dated 28^th April 1994. It showed results of a simulation model in relation to the reservoir serving platform 52/5A, as confirmed by Mr. Halliwell in evidence (Day 44, p.131). It was sub-headed “Assumption of compression on 52/5A at 1^st October 1994”. For the 3000, 4000, 6000 and 7500 HP cases it showed flow rates of individual Upper Bunter wells, increasing with each increment of power. It showed the corresponding flowing wellhead pressures, decreasing with each increment in flow rate. It showed pressures at inlet to the compressor, decreasing with each increment in power. It showed sea line pressures: in all those cases, 143 psia. The sea line pressure with no compression was shown as 141 psia.

The Phillips document of 28^th April 1994 was put to Mr. Halliwell in cross-examination (Day 44, p.132):

The reader familiar with the general characteristics of compressors concerning the relationship between power, speed, throughput of gas, pressure and pressure head would well understand that a compressor specified in accordance with sheet 5 of the Basis of Design would be likely to be capable of operating to some extent, unspecified, outside the envelope of the specification. Thus I accept the foregoing evidence of Mr. Halliwell except to the extent, if any, that he was implying that the specification on sheet 5 indicated that any particular performance, outside the envelope, was specified.

Dr. Robinson, in his second supplementary report, dated 28^th August 2002, said that the data on sheet 5 of the Basis of Design represented the maximum Upper Bunter flow for each year, at the predicted minimum pressure conditions within the sealine. Mr. Tomlinson, he said, appeared to agree with that interpretation. The relevant passage of the evidence of Mr. Tomlinson is this (Day 39, p.90, 92, 93 to 95):

I am satisfied that the data on sheet 5 of the Basis of Design represented conditions of maximum flow for each situation described. Lower flow would entail higher wellhead pressures. The flow might be adjusted by flow control valve (not an ideal method). It might be adjusted by adjusting the speed of the compressor. In that sense, I accept that the figures in the table on sheet 5 are typical. It is clear from the compressor curves that are in evidence that a reduction in flow does not necessarily involve a change in the discharge pressure.

Dr. Robinson gave another reason for saying that the pressures in sheet 5 were typical: that only one figure (for each month mentioned in the table) was given as the shut-in wellhead pressure for Upper Bunter, whereas in practice different wells were likely to have different shut-in wellhead pressures. Those pressures would be similar, but by no means identical. I accept that evidence.

The reference on sheet 11 of the Basis of Design to the flowing Zechstein wellhead pressures strongly suggests, in my judgment, that the main compressor must be capable of operating at discharge pressures slightly above the H line pressures at the 52/5A platform that are implied by the flowing Zechstein wellhead pressures applicable to 1995 and thereafter. I conclude that the Basis of Design required the main compressor to be capable of operating at a discharge pressure of at least 215 psia, allowing for uncertainty in the forecast. There was no specification of the flow rate that it had to be capable of sustaining at that discharge pressure.

I am satisfied that Snamprogetti must have appreciated that the main compressor must be capable of working at discharge pressures in excess of 200 psia. Although the pressure of the fuel gas for the main compressor had to be at least 188 psig (203 psia) according to the Basis of Design, Snamprogetti had stated in their Offshore Compression Study Report of 1^st October 1993 that the installation of the fuel gas compressor could be delayed, as I have mentioned in paragraph 123 above under the heading Vibration. Moreover, Snamprogetti’s statement that the installation of the fuel gas compressor could be delayed by several years implies that the main compressor was not expected to be working at its 1995 discharge pressure specified in sheet 5 of the Basis of Design, 142 psia, at all.

For the fuel gas supply to the main compressor, the pressure required at the edge of the gas turbine skid lay between 190 psig (according to the manufacturer’s drawings; 188 psig in the Basis of Design) and 250 psig. The fuel gas regulator (supplied by EGT as part of the main compressor package) controlled the pressure to the gas turbine at about 165 psig.

The fuel gas was Zechstein gas taken from a point downstream of the confluence of the supplies from the various Zechstein wellheads. Originally, the fuel gas supply was to be taken downstream of the Zechstein metering skid. Pursuant to a letter of 10^th February 1995 from Phillips to Snamprogetti, the supply was in fact taken from upstream of the metering skid.

The fuel gas compressor was designed to compress, and was capable of compressing, gas at inlet pressures corresponding to the range of main compressor outlet pressures specified on sheet 5 of the Basis of Design. That range was 143 psia for 1995 to 100 psia for 2003 to 2005. The corresponding inlet pressures for the fuel gas compressor were 142 psia and 99 psia respectively. The specification (Reciprocating Compressor Data Sheet) provided that where the inlet pressure was 142 psia the fuel gas compressor would compress the gas to a pressure of 233 psia at a flow rate of 59,373 scf/hr. Where the inlet pressure was 99 psia the fuel gas compressor would compress the gas to a pressure of 218.5 psia at a flow rate of 25,998 scf/hr. In fact, the latter flow rate would be substantially higher. Mr. Tomlinson (Day 38, p.127) thought that the latter figure was the usage of the turbine “at the end of life” (sc., when the main compressor outlet pressure was 100 psia).

In the Basis of Design, the flowing Zechstein wellhead pressure stated for the year 1995 was 200 psig. That implies an H line pressure of at least 180 psig, probably more. At any given time when Zechstein gas was flowing, the pressure upstream of the metering skid would be a few pounds per square inch higher than the pressure downstream of it. In their Offshore Compression Study Report of 1^st October 1993, section 1.3.3, Snamprogetti stated that a fuel gas heater would suffice in place of a fuel gas compressor for the early years. (The precise words are quoted in paragraph 123 above). It follows that Snamprogetti were well aware, or should have been, that the main compressor was required to operate with an H line pressure of at least 180 psig. Thus they must have been aware that the fuel gas compressor had to be capable of operating at an inlet pressure of at least 180 psig.

The fuel gas compressor was not designed to operate with an inlet pressure of 180 psig, nor was it capable of doing so. Dr. Robinson gave evidence in his first report (paragraph 11.18), and I accept, that there was a high pressure trip on the fuel gas receiver. He said that it was set at 230 psig. The actual figure was 240 psig (255 psia). That trip would stop all production on the platform. At an inlet pressure of 200 psig, he said, the discharge pressure from the fuel gas compressor would be about 290 psig, well above the trip setting. I am satisfied that he was wrong about that. As Mr. Sylvester-Evans explained, there was a recycle pressure control valve (PV 923, sometimes called PCV 923) which was set to control the discharge pressure of the fuel gas compressor at about 218 psig at the start of the project. Some variation in the discharge pressure could readily be catered for by the fuel gas regulator on the gas turbine skid. The problem lay in the fact that when the fuel gas compressor had an inlet pressure of approximately 200 psig, it was scarcely compressing the gas, so that the gas was not superheated. In consequence, liquids condensed out of the gas in the fuel gas supply line. When commissioning was attempted, that problem was experienced. It was not acceptable for the operation of the main compressor. I am satisfied that the same problem would have existed with an inlet pressure of 180 psig. The requirement to take the supply of fuel gas from upstream of the metering skid made little difference, and was not a cause of the problem. In that connection I accept the evidence of Mr. Tomlinson (Day 39, p.115) that the upper limit of inlet pressure for the fuel gas compressor to provide sufficient superheat was somewhere between 143 and 180 psia (128 and 165 psig).

That was an obvious error in design which in my judgment constituted a breach on the part of Snamprogetti of their contract with Phillips. What was needed to ensure sufficient superheat in the absence of a fuel gas heater was a suction pressure control valve for the fuel gas compressor. That would have reduced the inlet pressure to a sufficiently low value to give sufficient superheat on compression. Mr. Sylvester-Evans had carried out what he described as a back-of-an-envelope isentropic calculation which, allowing some margin, showed that there would be sufficient superheat with an inlet pressure as high as 175 psia (Day 68, pp.10, 11). I accept that evidence. Thus the inclusion of a suction pressure control valve limiting the suction pressure to 175 psia would have avoided this problem.

In his second supplementary expert’s report, of 28^th August 2002, Dr. Robinson commented on some oral evidence that had been given by Mr. Tomlinson. The fuel gas compressor was designed to supply 59,373 scf/hr of gas, compressed from an inlet pressure of 142 psia to a discharge pressure of 233 psia. Mr. McMaster cross-examined Mr. Tomlinson about the capacity of the fuel gas compressor to supply enough compressed fuel gas to the turbine of the main compressor. The following exchange took place (Day 40, p.34) (the references to E4, 161 are references to sheet 5 of the Basis of Design):

Mr. Tomlinson’s statement that the two compressors effectively would be in step is correct. But his statement (which was made with reference to the main compressor) that if the discharge pressure reduces the pressure ratio reduces is not necessarily correct. In particular, sheet 5 shows that as the discharge pressure falls from 143 psia the pressure ratio increases from 2.07 to a maximum of 3.36 when the discharge pressure is 121 psia, then reducing to 2.86 as the discharge pressure falls to 100 psia. And his statement that the power is the square of the pressure ratio (meaning, as was obvious, that the power is proportional to the square of the pressure ratio) was not correct.

Dr. Robinson made play with those errors. In his report of 28^th August 2002, he produced a table (Table 3) in which he set out his calculations of the power output of the main compressor in terms of its power output so calculated for the year 1995, using the figures in sheet 5. He carried out those calculations on the basis, as stated by Mr. Tomlinson (Day 39, pp.51, 52), that the power of the main compressor was proportional to the rate of flow and to the square of the pressure ratio. Those calculations showed that if the fuel flow supplied by the fuel gas compressor in 1995 was just sufficient to provide the calculated power in 1995, the fuel flow in each subsequent year until 2001 would be insufficient to provide the power calculated for the year in question.

Dr. Robinson was simply scoring points. He himself stated in his report that he did not accept as valid what he described as Mr. Tomlinson’s rule of thumb. The points that he made were that the (wrong) results in the previous paragraph should have led Mr. Tomlinson to investigate the problem more seriously. His actions were illogical and far below those that would have been expected of a first class contractor. The truth appeared to be that the effect of suction pressure on the capacity of the fuel gas compressor was largely or completely overlooked by Snamprogetti.

I do not accept that the effect of fuel gas pressure on the capacity of the fuel gas compressor was overlooked by Snamprogetti. But I am more concerned with whether the fuel gas compressor was actually defective in the respects alleged by the claimants. On the question of the adequacy of the fuel gas compressor to carry out its functions I derived considerable assistance from Mr. Sylvester-Evans. Mr. Sylvester-Evans carried out calculations of the amounts of fuel gas available under three different conditions against the corresponding requirements of the main compressor for fuel gas. The three conditions were fuel gas compressor inlet pressures of 142, 125 and 175 psia. He used as his basis for those calculations the assumption (which was not in issue) that at an inlet pressure of 142 psia the fuel gas compressor could compress 59,373 scf/hr, as stated in the Reciprocating Compressor Data Sheet. He assumed that the discharge pressure of the main compressor exceeded the inlet pressure of the fuel gas compressor in each case by 15 psi. He also carried out calculations of the power required by the main compressor to compress Upper Bunter gas to those discharge pressures. He then calculated what flow of fuel gas was needed to meet those power requirements. In each case there was more than enough gas available to meet the power requirement. I accept those calculations as substantially accurate. They showed that the availability exceeded the requirement by proportions ranging from 9 per cent (at a fuel gas compressor inlet pressure of 125 psia) to 30 per cent (at a fuel gas compressor inlet pressure of 175 psia).

Mr. Sylvester-Evans explained his assumption that the pressure drop between the discharge of the main compressor and the inlet of the fuel gas compressor was 15 psi. He took that as the pressure drop across the after-cooler of the main compressor and along the pipework from there to the sealine. He treated the pressure drop across the Zechstein metering skid as negligible. Slightly varying estimates were given by the witnesses for the latter figure. I am satisfied that it did not exceed approximately 5 psi. As to the 15psi pressure drop between the discharge of the main compressor and the sealine, the figures emanating from Phillips and incorporated in the Basis of Design included no allowance for that pressure drop. Mr. Sylvester-Evans made no allowance for a pressure drop across any suction control valve for the fuel gas compressor. He considered (Day 73, p.78) that the pressure drop across it would be small, not much more than “perhaps a couple of psi”.

In order to carry out his calculations, Mr. Sylvester-Evans used compressor curves, i.e., graphs showing the characteristics of the main compressor. Those curves applied to a suction pressure of 65.5 psia. He also used a graph showing the thermal efficiency of the main compressor as a function of compressor speed and power output. From one graph of compressor curves, he read off the speed of the main compressor required to discharge a flow of gas of 155,035 lb/hr at the discharge pressure in question. From the other graph of compressor curves, he read off the power required to drive that flow of gas through the compressor at that speed. He added 200 horsepower to allow for power losses in driving the compressor. The efficiency of the turbine in turning the heat energy in the fuel into mechanical energy he determined by reference to the thermal efficiency graph, which showed specific heat input in British thermal units per horsepower hour. Mr. Sylvester-Evans said (Day 68, p.30) that the calculations were rough and ready calculations.

Mr. Sylvester-Evans evidently selected the flow of 155,035 lb/hr because it was specifically marked on the graphs and facilitated reading them. I have calculated that that flow represents approximately 81 mmscf/d, which is substantially greater than the flows required by sheet 5 of the Basis of Design. Those flows vary from 74.1 mmscf/d for an inlet pressure of 69 psia and a discharge pressure of 143 psia through 39.3 mmscf/d for inlet and discharge pressures of 36 and 121 psia to 18.5 mmscf/d for inlet and discharge pressures of 35 and 100 psia. On that account Mr. Sylvester-Evans’s calculated requirements of power and fuel will be over-estimates. On the other hand, the use of the graph applying to a suction (i.e. inlet) pressure of 65.5 psia will tend to give rise to an under-estimate in the case of 125 psia sea line pressure, for in that case sheet 5 shows an inlet pressure of only 43 psia.

I have carried out a calculation, exactly on the lines explained by Mr. Sylvester-Evans, for the 1999 case in sheet 5. That has a flow of 39.3 mmscf/d (75,200 lb/hr), an inlet pressure of 36 psia and an outlet pressure of 121 psia. Using Mr. Sylvester-Evans’s assumption, I have taken the inlet pressure of the fuel gas compressor as 106 psia. I have used the compressor curves for a suction pressure of 33.80 psia, the nearest available to 36 psia. The required speed is 94 per cent, the horsepower is 3300, to which I have added 200, and the thermal efficiency factor is 10,750. That leads to a fuel requirement of 40,000 scf/hr. The available fuel flow is 44,320 scf/hr, an excess of 11 per cent. The use of a lower suction pressure than that contemplated in sheet 5 leads to a slight over-estimate of the fuel requirement.

To test the bottom end of the range, I have taken the 2003 case in sheet 5, viz. flow rate 18.5 mmscf/d, suction pressure 35 psia, and discharge pressure 100 psia. I have taken the inlet pressure of the fuel gas compressor as 85 psia. Again I have used the compressor curves appropriate to a suction pressure of 33.80 psia. Here, the main compressor has to run on recycle. That creates inefficiency, but does not affect the method of the calculation. I have chosen a flow through the main compressor (including recycled gas) of 50,000 lb/hr (26 mmscf/d). The speed is 84 per cent and the power requirement is 2000 horsepower. Adding 200 horsepower yields a thermal efficiency factor of 12,800 and a fuel requirement of 29,900 scf/hr. The fuel available is 35,540 scf/hr, an excess of 19 per cent. (For any given compressor, there exists a minimum flow rate, dependent on other factors such as the pressure, below which the phenomenon of surge will occur. Surge is an unacceptable fluctuation in pressure. To avoid surge, flow through the compressor is maintained by recycling gas that has passed through it).

On the evidence of Mr. Sylvester-Evans, plus my calculations based on it, I am satisfied that the fuel gas compressor would have met all the requirements of sheet 5 with a margin to spare, provided that a suction pressure control valve were fitted giving an upper limit of 175 psia for the inlet pressure in order to maintain sufficient superheat.

The claimants contend that there should have been a different pressure control valve. They say that the inlet pressure of the fuel gas compressor should have been limited to 95 psia. They rely on the offshore compression study of 1^st October 1993, which states at paragraph 2.3.3:

The point was mentioned again by Dr. Robinson in his report of 28^th August 2002 (paragraph 6.4):

The claimants have interpreted the expression “the highest fuel gas rate…..to meet the requirements of the gas turbine” to mean the highest fuel gas rate that the turbine could at any time require. That is a possible reading of the expression. But in its context I take it to mean that even at the lowest inlet pressures, the fuel gas compressor must be capable of supplying, at the constant pressure, a flow of gas sufficient for the contemporaneous requirements of the gas turbine. As I have found in paragraph 185 above, it was capable of doing that. (In fact, at the lowest inlet pressure, 99 psia, the specification did provide for a somewhat lower outlet pressure, viz. 218.5 psia compared with 233 psia). It is true that if such a suction pressure controller had been used, the fuel gas compressor would not have been capable of supplying sufficient fuel to the turbine to allow the main compressor to fulfil the requirements of sheet 5 for the earlier years, i.e. for higher discharge pressures of the main compressor. I understand Mr. Tomlinson’s agreement that the sentence in question was the correct way to size a compressor to be based on a similar understanding of its meaning to mine, particularly having regard to the passage I have emphasized.

At the HAZOP meeting held on 21^st May 1996 it was recommended that a suction pressure control valve controlling the inlet pressure at 80 psig be added to the fuel gas compressor. The cause given was high Zechstein operating pressure. The relevant worksheet stated under the heading “Consequences”:

Mr. Tomlinson attended the HAZOP meetings. He was cross-examined about the HAZOP recommendation (Day 39, pp.47, 51, 52):

Dr. Robinson said that the use of a fixed set point at around 80 psig, as recommended at the Baker Jardine HAZOP, was the logical way to proceed (his report of 22^nd August 2002, paragraph 6.19). He reached that conclusion because he considered that variations in the inlet pressure to the fuel gas compressor would render control of the outlet pressure impossible or inefficient. In order to consider that evidence of Dr. Robinson, I must explain the system in more detail.

The fuel gas was supplied to the turbine at a constant pressure, 165 psig. That pressure was controlled by the fuel gas regulator, a valve situated on the main compressor turbine skid. Dr. Robinson and Mr. Sylvester-Evans differed over the operation of the fuel gas regulator. Both witnesses agreed that the regulator valve could not properly handle rapid variations in the pressure of the gas upstream of it. Dr. Robinson said (Day 65, p.26) that it was a device to produce a constant pressure drop. Thus if the pressure upstream of it varied [sc., slowly], the pressure downstream of it would likewise vary. Mr. Sylvester-Evans, on the other hand, said (Day 67, p.138) that the valve was designed to establish a constant pressure downstream of it. Dr. Robinson accepted that it was intended to produce a constant pressure downstream of it; but that it was intended for use where there was a constant pressure upstream of it. I am satisfied that it is Mr. Sylvester-Evans who is correct in this respect. He drew my attention to a technical description of the valve, I assume emanating from the manufacturer. It is clear from that description that the outlet pressure was regulated, and that that regulated pressure plus the pressure applied by a spring balanced the fixed pressure of a trapped mass of gas. Provision was made for correcting the pressure of the trapped mass of gas if it varied. The spring controlled the opening of the valve. If the downstream pressure departed from the required value, the resulting imbalance would cause the spring to move, restoring the downstream pressure to its required value.

The pressure upstream of the regulator was the outlet pressure of the fuel gas compressor. It was specified in the Reciprocating Compressor Data Sheet as 233 psia for a flow of 59,373 scf/hr and 218.5 psia for a flow of 25,998 scf/hr. It was nevertheless intended to be held constant over periods of time. Rapid fluctuations in it were not acceptable. It was controlled in the following way. Gas flow excess to the requirements of the turbine was taken off at the outlet of the fuel gas compressor and returned via a spillback line, through a control valve PV 923, to the inlet of the fuel gas compressor. The pressure at the outlet of the fuel gas compressor was controlled by controlling the flow though the spillback line. The degree to which the control valve was open was controlled by a device sensitive to the pressure upstream of the valve (i.e., to the outlet pressure of the fuel gas compressor). The flow through the control valve depended also on the pressure downstream of it (i.e., the inlet pressure for the fuel gas compressor). By controlling the flow through the spillback line so as to maintain a constant pressure at the outlet of the fuel gas compressor, the device ensured that the flow of gas supplied by the fuel gas compressor to the regulator matched the flow required by the turbine.

In paragraph 11.18 of his first report, dated 31^st January 2002, Dr. Robinson gave the following as one of his reasons for having a suction pressure control valve:

I have already indicated that Dr. Robinson was quite wrong in stating that the discharge pressure of the fuel gas compressor could be expected to be around 290 psig. The following exchange took place during his cross-examination (Day 65, p.31):

It is clear that the point made in justification of a suction pressure controller in paragraph 11.18 of Dr. Robinson’s first report is a bad one. I cannot accept that it can fairly be described as an approximation. No witness, let alone an expert witness who, like Dr. Robinson, has previous experience of giving expert evidence, can reasonably think that a wrong explanation can help a judge to understand a point. This matter detracts, to some extent, from Dr. Robinson’s credit as a witness in this case.

Dr. Robinson did give another reason in his first report for saying that a suction pressure controller was required. It is that other reason which he sought to support in a later report and in his oral evidence. The “detailed calculation” was adduced in support of that other reason.

In his first report, dated 31^st January 2002, Dr. Robinson said in paragraph 11.21 that upstream pressure control was essential for efficient operation of the pressure control loop using PV 923 in the spillback line. A control loop of that type was efficient only when the discharge side of the control valve was at a steady pressure. The main purpose of PV 923 was to control the pressure of fuel gas at entry to the gas turbine, for which the gas take-off was upstream of that control valve. PV 923 must respond to demands from the turbine for changes in the gas flow and ensure that the pressure of 200 psig [sic] was always maintained for the fuel gas going to the turbine. It could not achieve that efficiently if the pressure on its discharge side was also changing in an uncontrolled manner. Without upstream pressure control to the fuel gas compressor, the pressure on the discharge side of PV 923 would “float” on the sealine pressure. Valve PV 923 would thus be required to achieve control in response to changing pressures on both its upstream and downstream sides. Such a requirement was grossly inefficient. It represented a design which would never have been provided by a competent engineering contractor.

I note in passing that it is not accurate to say that the main purpose of the recycle valve PV 923 was to control the pressure of fuel gas at entry to the gas turbine. Its main purpose was to control the pressure of fuel gas at entry to the gas turbine skid, upstream of the regulator valve. The regulator valve controlled the pressure of gas at entry to the turbine. The regulator valve could do that in spite of any slow variations in the pressure upstream of it, within the stated limits (190 to 250 psig).

Pursuant to the Baker Jardine HAZOP, Mr. Tomlinson prepared a redesign of the fuel gas compressor introducing a suction pressure control valve capable of limiting the inlet pressure to 99 psia. In paragraph 6.6 of his report dated 28^th August 2002, Dr. Robinson said that the problem with the simple addition of the suction pressure controller was that the introduction of the controller, and consequent reduction of suction pressure [sc. to about 95 psia], would severely limit the capacity of the fuel gas compressor. That reduction would be so severe that the turbine would be almost useless. Mr. Tomlinson was cross-examined about that (Day 39, pp.33, 34, 37):

Dr. Robinson criticized the idea of having different set points for the suction pressure control valve. He set out his argument in paragraph 6.9 and the succeeding paragraphs of his report of 28^th August 2002. It was this. The fuel gas compressor must deliver gas to the gas turbine at a set pressure. The flow rate of gas to the gas turbine can vary from its maximum value, when the turbine power is a maximum, down to zero, when the fuel gas compressor is working at total recycle. When the turbine is operating at full power a small flow must be maintained through the recycle, or pressure cannot be controlled. The recycle (spillback) control valve must be capable of controlling the pressure, with a flow from 110 per cent. of the maximum flow required by the turbine, down to a flow of only 10 per cent. of that maximum. Because of the wide control range, valve PV 923 should have a nearly linear flow characteristic. That means that equal increments in valve opening should give nearly equal increments in flow, for any given pressure drop across the valve. A control valve can normally operate between 10 per cent. and 90 per cent. of its full range. (It is common ground that the recycle valve as supplied was in any event unsuitable. That is a separate question, but the properties that it should have had are relevant to the present question).

The argument continued: if the setting of the suction pressure controller were changed from 95 psia to 142 psia, the flow rate through the fuel gas compressor would increase in the ratio 142:95=1.49. Thus in full recycle, the recycle control valve would need to handle 1.49 times the amount of gas that it would need to handle at suction conditions of 95 psia. The pressure drop across the valve would decrease from 233-95=138 psi to 233-142=91 psi. The capacity of a valve varies as the square root of the pressure drop across it. The effect of a reduction in pressure drop from 138 to 91 psi on a linear valve would be to require it to open in proportion to v(138/91)=1.23. The combined effect would be to require the valve to be 1.49 x 1.23=1.83 more open at full recycle at suction conditions of 142 psia than at suction conditions of 95 psia. Given that the system was designed to suction conditions of 95 psia the valve would simply be wide open on full recycle at 142 psia, and no control would be achieved. And the valve could not pass the required flow of gas. For at 95 psia the valve would be 90 per cent. open at the full flow rate appropriate to that suction condition. Manifestly, it could not open further by a factor of 1.83. If the valve were linear above 90 per cent. open, it could pass only 1.11 times that flow rate. Yet the increase required of it would be by a factor of 1.49 [I have slightly amended Dr. Robinson’s wording without affecting the argument].

If, on the other hand, the recycle valve were designed to cover suction pressure controller set points from 95 to 142 psia, the position would be this. At 10 per cent. open, the valve would pass one-eleventh of the full recycle rate at 95 psia suction pressure. At 90 per cent. open, it would pass the full recycle rate at 142 psia suction pressure, i.e. 1.49 times the maximum recycle rate at 95 psia suction pressure. Not only that, but at 90 per cent. open and with a suction pressure of 95 psia, the recycle valve would pass 1.83 times the maximum recycle rate applicable to that suction pressure. Thus for maximum recycle rate at 95 psia suction pressure, the valve would be open only to the extent of 52 per cent. That would be grossly inefficient, because necessary adjustments to the valve in response to changes in pressure upstream of it would have to be unnecessarily fine. That would make control difficult. [The percentage figure of 52 is the value of x in the equation

Those arguments were originally directed against the idea of having variable set points for the suction pressure control valve. In cross-examination, Dr. Robinson relied (Day 65, p.34) on those arguments generally in support of his view that there should be a suction pressure controller limiting the inlet pressure of the fuel gas compressor to about 95 psia. He accepted that the changes in sealine pressure [and hence suction pressure] would be slow-moving (though large). Nevertheless, he considered that that did not affect the foregoing arguments. He said in evidence (Day 60, p.48):

Thus Dr. Robinson was making these points. 1. The recycle valve cannot effectively control its upstream pressure if its downstream pressure is varying in an uncontrolled manner. 2. If a suction pressure controller is used and set at 142 psia, the recycle valve would have an insufficent range of opening to handle all contingencies. 3. If a suction pressure controller is used and set at 95 psia, the whole range of the valve is not used, and that is inefficient.

Mr. Sylvester-Evans agreed (Day 72, p.113) that when the turbine imposed its maximum requirements for fuel gas there should still be some flow of gas along the spillback line, which might be 10 per cent. of the maximum requirement of the turbine. He said that the margin could be refined somewhat, depending on the particular control valve used.

Mr. Sylvester-Evans said that given that changes in the sealine pressure were slow, the system could respond to it smoothly. He gave these answers in cross-examination (Day 72, p.119):

In my judgment, Dr. Robinson’s view that having a suction pressure control valve set at 95 psia was the logical way to proceed is unjustified. According to his own criteria, the recycle valve in that case would be required to pass, when 90 per cent. open, 110 per cent. of the maximum flow required by the turbine even in the early years when the highest powers would be required of it. And it would be required to pass, when 10 per cent. open, 10 per cent. of the maximum flow required by the turbine in the later years. Those constraints would require of the recycle valve PV 923 a sensitivity in terms of incremental flow per incremental opening not far divergent from the implicit figure which he described as giving rise to gross inefficiency.

I have carried out the following calculations of fuel gas flow required: 1. On the basis of the 1995 figures in sheet 5 of the Basis of Design, namely flow through main compressor of 74.1 mmscf/d, inlet pressure 69 psia (I have had to use the curves for 65.50 psia), main compressor outlet pressure 143 psia. 2. The same as 1, except that the outlet pressure is taken as 157 psia, following Mr. Sylvester-Evans. 3. On the basis of the 2003 figures, assuming a flow through the main compressor, including recycle, of 50,000 lb/hr (26 mmscf/d), inlet pressure 35 psia (I have had to take the 33.80 psia curves), outlet pressure 100 psia. 4. The same as 3, except that the outlet pressure is taken as 115 psia, following Mr. Sylvester-Evans, and the flow, including recycle, has been taken as 60,000 lb/hr. (31 mmscf/d). 5. For the early years and a sealine pressure of 175 psia, the same as calculation 1, save that the outlet pressure is 190 psia.

Calculation 1 has a main compressor speed of 80 per cent., power of 3700+200 horsepower, efficiency factor 10850 and fuel requirement 45000 scf/hr. For calculation 2, the corresponding figures are respectively 81½ per cent., 4000+200 horsepower, 10700 and 47700. For calculation 3, the figures are 84 per cent., 2000+200 horsepower, 12800 and 29900 scf/hr. For calculation 4, the figures are 90 per cent., 2600+200 horsepower, 11750 and 35000 scf/hr. For calculation 5, the figures are 86 per cent., 4800+200 horsepower, 10150 and 53900. Calculation 3 is the calculation mentioned in paragraph 185 above.

Taking the first four calculations in pairs, the difference between 110 per cent. of 45000 scf/hr (calculation 1) and 10 per cent. of 29900 scf/hr (calculation 3) is 46500 scf/hr. The difference between 110 per cent. of 47700 (calculation 2) and 10 per cent. of 35000 (calculation 4) is 49000. As between calculations 5 and 4, the corresponding difference is 55800. For a linear recycle valve complying with Dr. Robinson’s requirements, those differences will have to represent a range of 80 percentage points of the opening of the fuel gas recycle valve PV 923. They amount respectively to 580, 612 and 697 scf/hr per percentage point.

Dr. Robinson considered that it would be grossly inefficient to have the required opening of the recycle valve limited to 52 per cent. when operating at 95 psia suction conditions. The maximum flow available from the fuel gas compressor at that inlet pressure was 39720 scf/hr. According to Dr. Robinson, that would have to be at least 110 per cent. of the relevant maximum requirement of the turbine. Thus the range of flows that valve PV 923 would have to handle over a range of 42 percentage points of opening would be 10/11 of 39720, or 36110 scf/hr. That amounts to 860 scf/hr per percentage point. That in my judgment is not so different from the amounts in the previous paragraph as to justify describing the latter as grossly inefficient whilst the former are implicit in Dr. Robinson’s recommendation. I find Dr. Robinson’s evidence on this point unconvincing.

Dr. Robinson appreciated that the use of a suction pressure control valve with a set point of 95 psia would render the fuel gas compressor incapable of providing sufficient fuel to the turbine in the early years. In his view, that went to show that Snamprogetti had not thought through the design properly.

A parameter with the symbol Cv is used in relation to valves. It has been called the capacity of the valve. (Dr. Robinson used the expression capacity in a different sense in paragraph 203 above). For any given setting of a given valve, it is a constant proportional to the ratio of the mass flow of gas through the valve divided by the square root of the product of the pressure difference across the valve and the density of the gas upstream of the valve. The wider open the recycle valve is, the greater the mass flow for any given pressure difference and upstream density. Thus Cv increases as the valve is opened. For a linear valve, Cv is proportional to the amount of opening. Those points are common ground. (In a document put to Mr. Sylvester-Evans in cross-examination the expression Cv, called the valve coefficient, was used as a constant for a valve. Its product with the flow characteristic of the valve, which was a function of the valve lift, was equivalent to the “capacity of the valve” Cv as used before me. I adhere to the latter usage).

Mr. Sylvester-Evans carried out comparative calculations of the Cv values required of the recycle valve PV 923 in various circumstances. Those circumstances were the three conditions mentioned above in relation to which he calculated the flow of fuel required by the turbine and the flow available from the fuel gas compressor. In each case, he calculated the value of Cv for the minimum spillback, representing the case of maximum flow required by the turbine. Those values of Cv ranged between 435 for a fuel gas compressor inlet pressure of 125 psia to 2,225 for a fuel gas inlet pressure of 175 psia. In all cases, he assumed that the pressure upstream of the recycle valve was 233 psia. That implies that the gas upstream of the recycle valve was at the same density in all cases, and so the density could be ignored (as it was) for comparative calculations. I am satisfied that those figures are correct; nor were they controversial. The range represents a factor of about 5.

I have calculated Cv values also for the 1999 case in sheet 5 of the Basis of Design (106 psia fuel gas compressor suction pressure) and for 85 psia (2003 case). The corresponding flow figures for the maximum requirement of the turbine and for the fuel gas compressor are mentioned in paragraphs 184 and 185 above. The respective differences represent the minimum recycle flows. The values of Cv are respectively 383 and 464. (If in the latter case one took the fuel gas compressor discharge pressure as 218.5 psia, as in the Reciprocating Compressor Data Sheet, the comparative value of Cv would be 504). The minimum Cv value that emerges is thus 383. The range 383 to 2225 represents a factor of 5.8.

The maximum value of Cv occurs on full recycle at 175 psia. The flow is 73170 scf/hr, and the value of Cv is 9608. The range of Cv from 383 to 9608 represents a ratio of 25. The maximum value of Cv for the cases in the Reciprocating Compressor Data Sheet is for the flow of 59373 scf/hr. It is 6224. That is the case where the inlet pressure of the fuel gas compressor is 142 psia. The range of Cv from 383 to 6224 represents a ratio of 16.

Mr. Sylvester-Evans gave evidence about a suitable recycle valve (Day 68, pp.41 to 48). In his view, an equal percentage valve would be a reasonable approach for this application. An equal percentage valve is a valve where, over its range of operation, e.g. from 10 per cent. fully open to 80 per cent. fully open, the gas flow is an exponential function of the degree of opening (i.e. the degree of opening is a logarithmic function of the gas flow). That is, any given increment in opening will produce the same fractional increment in flow. For example, if the range of Cv of the valve is represents a factor of 50 from 10 per cent. open to 80 per cent. open, then for each increment of opening representing 10 per cent. of fully open, the value of Cv will increase by a factor of about 1.75. Mr. Sylvester-Evans made the position clear from a diagram that he drew. I am satisfied that my understanding of the equal percentage valve is at least substantially correct, notwithstanding some confusion in the oral evidence, possibly caused by a question of mine.

Mr. Sylvester-Evans said that a valve might well offer a Cv range of 30 or 50 or even more. The rated case for the equal percentage valve might be 59373 scf/hr at 70 per cent. open. But such a valve, if properly sized, would easily take 73170 scf/hr (the maximum flow of the fuel gas compressor at an inlet pressure of 175 psia) at 80 per cent. open. When the turbine was working, the flow through the valve would be at the low end of the range. In that case, for a given required change in Cv, the equal percentage valve would give a greater distance of valve travel (i.e. opening) than a linear valve working over the same range of Cv. I accept that evidence. A calculation of mine illustrates the position. Consider two valves, one of each type, both operating over a range of Cv representing a factor of 20. The linear valve operates from 10 per cent. open to 90 per cent. open. The equal percentage valve operates from 10 per cent. open to 70 per cent. open. To increase the Cv by a factor of 5 from the value of Cv when the valve is 10 per cent. open would require an increment of 16.8 per cent. in opening for the linear valve, and of 32.2 per cent. in opening for the equal percentage valve.

Mr. Sylvester-Evans said, and I accept, that the higher increment in valve travel for a given increment in Cv in theory gave more controllability. To that extent he was in agreement with Dr. Robinson. Mr. Sylvester-Evans said (Day 73, p.33)

The position would be the other way round at the top end of the range. But that situation, full recycle, would be used only for short times. Mr. Sylvester-Evans referred to start-up. He acknowledged that if the valve did not work at 175 psia then the only way you could start up the turbine at such a sealine pressure would be to lower the pressure in the sealine. Mr. Sylvester-Evans thought that there was “every likelihood” that a recycle valve would have worked with the inlet pressure of the fuel gas compressor even as high as 175 psia. And he expressed the view (Day 73, p.54) that there should be a wide suite of valves available.

As to the possibility of finding an alternative to the unsuitable valve PV 923 that was supplied, Mr. Sylvester-Evans said (Day 73, p.46):

I am satisfied that a suitable valve or possibly pair of valves could have been found to replace the existing valve PV 923. And I conclude, contrary to the opinion of Dr. Robinson and to the recommendation of the HAZOP, that in any event the fuel gas compressor did not need a suction pressure control valve limiting the inlet pressure to 95 psia or thereabouts.

It is evident from the pleadings (see paragraph 133 above) that the fuel gas compressor was never needed, and that the fuel gas heater served in its place. There is no direct evidence why that was so. It is not for the court to speculate. Possibly the diminution in flow of Zechstein gas caused by the dedication of Zechstein wells to the fuel gas heater was permanently acceptable. But it is worth mentioning some evidence that may be relevant if that is not so. That evidence, which I consider in the following paragraphs, concerns the following propositions. 1. Shut-in wellhead pressures of Upper Bunter gas were substantially higher than those forecast by Phillips. 2. Average H line pressures may have increased so that lowest pressures did not fall below the minimum pressure (190 psig) required of the fuel gas. 3. If, contrary to 2, there continued to be fluctuations in H line pressure below 190 psig, the alpha compressor was available, if necessary, to draw and compress the Upper Bunter gas.

The forecast Upper Bunter shut-in wellhead pressures stated in sheet 5 of the Basis of Design were 168 psia in May 1995 falling to 146 psia in October 1996 and falling further thereafter. The Phillips document mentioned in paragraph 40 above showing shut-in pressure data has not been spoken to by any witness of fact. It does not state whether the pressures shown are expressed in terms of psia or in terms of psig. But Dr. Robinson and Dr. Sylvester-Evans were of one mind in thinking that the figures were probably expressed in terms of psig. The shut-in wellhead pressure for Upper Bunter well A1 in 1995 was shown as 174, which probably means 189 psia. For 1996, four of the Upper Bunter wellheads were shown as having shut-in wellhead pressures lying between 186 and 189 psi, which probably means between 201 and 204 psia.

The historic figures for the shut-in wellhead pressures of the Upper Bunter wells in 1996 represent a substantial increase over the figure forecast and appearing in the Basis of Design. I am satisfied that up to 1996 at any rate the shut-in wellhead pressures of Upper Bunter wells were substantially higher than those forecast by Phillips and used by Snamprogetti in the Basis of Design.

When the Basis of Design was prepared, some figures of historic H line pressures were available. They were daily figures for the period from 25^th August 1993 to 26^th September 1993. They range between 168 psig and 301 psig (183 psia and 316 psia). No flow of gas is shown from the 52/5A platform. Subsequently, figures of H line pressure from 1^st October 1994 were available. They vary between 128 psig and 295 psig (143 psia and 310 psia). Dr. Robinson, in his first expert’s report, dated 31^st January 2002, said that over short periods of days or weeks the pressure variations were considerable. That was entirely to be expected. They were typical of pipelines of that sort. Those variations in pressure would be understood by a competent contractor and would be anticipated. If Snamprogetti interpreted the data given to them as meaning that the sea line would operate at a fixed pressure all the time, it was not applying the standards of a first class contractor.

I accept Dr. Robinson’s evidence that Snamprogetti ought to have appreciated that fluctuations in H line pressure were to be expected. It does not follow, however, that they should have expected the range of pressures reported. In particular, where no flow of gas from 52/5A was shown, the figures would afford little, if any, guidance.

There is before me a memorandum dated 8^th February 1996 sent by Mr. S. Woodington (Snamprogetti project engineer) to Mr. J Gibbons of Phillips. It says, among other things:

On 11^th September 1995 Mr. D. White of the Phillips reservoir engineering group sent a fax message to Mr. Halliwell and Mr. Hobson. Some diagrams, described as nodal diagrams, were sent with the fax. One of those diagrams was headed “B line configuration: Gas capacity (mmscf/d) at 1/1/96”. There was another diagram showing the corresponding pressures. For the main compressor on platform 52/5A it showed suction pressure of 123 psia and discharge pressure of 244 psia. The H line was shown as entering the B line downstream of the alpha compressor, where the pressure was 235 psia. In his fax message, Mr. White asked the addressees to review the nodal diagrams and to confirm that they agreed with the flow configuration of the wells and that the pressures given at the various nodes were reasonable. He expressed concern with the compression ratio [2.0] that the model assumed for the 52/5A compressor.

Those documents were put to Mr. Halliwell in cross-examination (Day 45, p.20). It is implicit in his answers that those pressures were indeed expected by Phillips. Clearly, no fuel gas compressor would be necessary with a pressure of 244 psia in the H line.

When the Basis of Design was prepared the alpha compressor discharged only into the A line. That fact was implicit in the following paragraph on sheet 4 of the Operating Philosophy:

The HIPS (High Integrity Protection System) is a system designed to protect pipework, vessels, machinery and other items to which gas has access from overpressure, that is pressure from gas at a pressure higher than that for which the item has been designed. This claim is concerned with the HIPS in the main compressor system on platform 52/5A. The system consists of a device to measure pressure of the gas and firmly to cut off supply of the gas to the vulnerable parts of the equipment if the measured pressure exceeds a set limit. To increase reliability, the system operates in triplicate. That is, there are three separate devices to measure the pressure, and two of them registering excess pressure are sufficient to cut off the supply of gas. Thus the system works even if one of the devices fails. Moreover, one of the devices at a time can be removed for maintenance and testing to ensure reliability. The system is designed to have a probability of failure in any year of not more than one in a million.

The shut-in wellhead pressure of the Zechstein gas was over 1000 psi. The HIPS was designed to guard the main compressor and the pipework and equipment upstream of it, in particular the suction knock-out drum, from overpressure arising from unintentional (“spurious”) closing of a valve in the sealine. If that were to happen, gas could back up from the Zechstein gas line to the main compressor and the pressure could build up fairly rapidly to the shut-in pressure of the Zechstein wells. That would greatly exceed the design pressure of the relevant pipework and equipment. The suction knock-out drum had its own pressure relief valve, but it was common ground that that had insufficient capacity to cope with a surge in gas of the magnitude against which the HIPS was intended to guard.

Gas passed from the Upper Bunter wells through the knock-out drum, then through the compressor and after-cooler and along pipework to join the sealine. An anti-surge line diverged from the compressor discharge line just downstream of the after-cooler and ran back to a point upstream of the knock-out drum. Compressed gas could flow into the anti-surge line, lose its excess pressure as it passed through a valve FV 820, and be re-circulated through the knock-out drum and the compressor. The purpose of the anti-surge line was to allow the compressor to operate when the gas flow required from Upper Bunter was below the lower limit at which the compressor could operate.

The HIPS as designed by Snamprogetti was positioned immediately downstream of the point where the anti-surge line diverged from the discharge line. It was set at 435 psig. There was a non-return valve situated between the main compressor and the after-cooler. The design pressure of the main compressor and of the pipework downstream of it (including the pipework upstream of FV 820) was 435 psig; the design pressure of the pipework upstream of the main compressor (including the knock-out drum and the pipework in the anti-surge line downstream of FV 820) was 250 psig.

If the pressure as detected by the HIPS exceeded 435 psig several things would happen. The Zechstein gas line would be isolated. Certain other valves would be operated; in particular, FV 820 would be opened. The main compressor would be tripped, i.e. stopped. And a valve upstream of the knock-out drum would be opened, allowing gas to vent to the atmosphere.

In addition to all that I have mentioned, there was a compressor by-pass line. That line diverged from the line from the Upper Bunter wells upstream of the point of confluence of the compressor anti-surge line. Between those points of divergence and of confluence, there was a valve XV 803. The compressor by-pass line joined the compressor discharge line downstream of a valve XV 841 which was itself downstream of the HIPS. The compressor by-pass line contained a non-return valve. It was common ground that non-return valves were not a sufficient protection of the items upstream of them.

Phillips claimed that the HIPS as designed was inadequate for its purpose. If the by-pass were in use, then valve XV 841 would be closed. The HIPS would thus not be subjected to any surge in pressure from the Zechstein gas line, and would in consequence not operate. The compressor by-pass line would be subjected to overpressure. Moreover, valve XV 803 was not automatic. Although the procedure was for an operator to close that valve when the by-pass line was in use, the reliability of that procedure was insufficient for the exacting standards of safety required. Thus the whole of the compressor system, including the knock-out drum and the after-cooler, was insufficiently protected. (In fact, I am satisfied on the evidence of Mr. Sylvester-Evans that valve XV 803 was automatic).

Snamprogetti submitted that the compressor by-pass line would be used only when the compressor was shut down. In that case, the compressor would be vented to the atmosphere. Thus the protection as designed was adequate. It was true that the protection had to be redesigned, but that was because of a change in operating philosophy adopted by Phillips. Phillips had decided that they did not want the compressor vented to atmosphere every time it was shut down.

Whatever the force of that argument, it is clear that if XV 803 were closed as intended, the compressor by-pass line would not be protected by the opening of the vent.

It is clear that Snamprogetti omitted to allow for the compressor by-pass line in their design of the HIPS. At a design HAZOP conducted by Snamprogetti on 22^nd August 1994 it was reported that the by-pass line had no apparent purpose. The need for it should be reviewed, and it should be deleted if possible. That report was sent to Phillips on 26^th September 1994. There is a letter of 5^th October 1994 from Snamprogetti to Phillips stating, with reference to the HAZOP, that Snamprogetti were continuing to include the by-pass line in the design as Phillips had confirmed that it was required. On a copy of the HAZOP report there is a note dated 12^th December 1994 stating that Phillips did see a purpose for the by-pass and it should therefore be left in. Nevertheless, as appears from internal memoranda of Snamprogetti dated 7^th June 1996 and 12^th August 1996 it is apparent that the HIPS was designed, or its design was left unchanged, on the basis of the proposed change involving the removal of the by-pass line, and that the letter of 5^th October 1994 appeared to be the justification for taking no further action on the by-pass.

In cross-examination, Mr. Tomlinson accepted that the failure to design the HIPS to protect the by-pass line was a serious omission and constituted an error on the part of Snamprogetti. He was right to do so. The error constitutes a breach of the duty that Snamprogetti owed to Phillips.

A number of other points relating to this issue were explored in the evidence and argument, but in the event it is unnecessary for me to deal with them.

The turbine driving the main compressor burnt Zechstein gas, which was largely methane. The claimants alleged that methane was unsuitable for starting the turbine, because it would not ignite reliably. The difficulties starting the turbine led the claimants to instal an alternative start-up system using bottled propane. This was originally a complaint that Snamprogetti should have enquired of the vendor of the turbine (European Gas Turbines: EGT) whether or not methane was a suitable start-up gas.

The claim was amended. As amended, it stated that the turbine was tested at the factory using propane. The turbine started satisfactorily on propane. No tests were performed using methane with the composition of Zechstein gas. (Zechstein gas contains some ethane and small quantities of higher hydrocarbons and other gases). The only start-up gas tested satisfactorily was propane. Notwithstanding that there had been no tests with methane, still less methane with the composition of Zechstein gas, Snamprogetti specified Zechstein gas as the start-up gas. Start-up was unsatisfactory. The problem was cured by using propane but at greater cost because the modifications did not take place until after the compressor had been installed on the platform.

By the end of the hearing, the argument was this. The contract required Snamprogetti “to review factory witnessed test procedure and specifications as prepared by respective Vendors” and to “verify the acceptability of the factory tests”. In performing those obligations Snamprogetti was “to observe and exercise the standards of skill, care and diligence adhered to by recognized first-class contractors performing work of a similar nature”. It is not in dispute that those were terms of the contract. It was argued that Snamprogetti were in breach of contract in not ensuring that the start-up of the machine was factory tested on methane, and that that breach of contract led to the propane start system being installed.

EGT issued a document describing customer-witnessed performance demonstration tests for the turbine. Those were factory tests. They included the provision “Start the turbine and observe that a satisfactory start is accomplished with ‘Turbine Running’ annunciated”. Snamprogetti produced a document entitled “Construction and pre-commissioning requirements at fabrication yard for the installation and test running of the gas compressor package…..” The fabrication yard was the fabrication yard of AMEC at Sunderland. No criticisms were made of either of those documents.

The turbine had four separate combustion chambers, equally spaced around the turbine case. Pressurized air and fuel gas mixed and ignited within each combustion chamber. The hot gases expanded and drove the turbine. Before the main fuel supply to each combustion chamber could open, there had first to be established, and proven, a pilot flame in each can. That was a safety feature designed to prevent the accumulation of an inflammable mixture within the casing and exhaust of the gas turbine. The starter gas was used to establish the pilot flame. Once the control system was satisfied that the pilot flame was proven the main fuel gas supplies were opened. Methane ignites readily, but propane more so since it has a lower ignition temperature. The above facts stated in this paragraph I have obtained from Mr. Sylvester-Evans’s consolidated report.

There were two relevant kinds of detector capable of proving the existence of the pilot flame. One was an ultra-violet system, which detected the flame by the radiation emanating from it. The other, known as the Rustronic Mark II control system, used a thermo-couple to sense the temperature caused by the flame. On 8^th March 1994, during the tender process, Snamprogetti asked EGT whether a propane pilot was necessary. EGT replied that they had found in certain installations that the natural gas [sc., methane] ignited perfectly well, but the flame emitted a low level of ultra-violet [radiation], making flame recognition difficult. They could not guarantee that the use of propane would be unnecessary if a flame detection system were used. They strongly recommended using the “inference flame detection” system, i.e. the thermo-couple system. It was fitted as standard to the EGT Rustronic panel offered as option 1. Snamprogetti’s specification for the compression facilities dated 19^th September 1994 incorporated the Rustronic Mark II control system.

The factory tests of the turbine were carried out at EGT’s factory at Lincoln in January 1995. UV sensors were used in conjunction with propane gas for start-up. Those tests were attended by Mr. Thomson. He gave evidence that there was a problem with start-up, which was resolved. That problem is irrelevant for present purposes. The fact is that the factory tests did not test the use of the thermo-couple system or the start-up of the turbine with methane gas.

Mr. Thomson gave this evidence in cross-examination (Day 37, p.148):

At a meeting held on 4^th May 1995 between representatives of Phillips, Snamprogetti and EGT, future testing of the turbine and compressor at AMEC’s fabrication yard was discussed. It is clear from the minutes of that meeting that at that stage all parties expected those tests to be carried out at the fabrication yard.

Mr. Thomson’s specification for testing at the fabrication yard was the document entitled “Construction and pre-commissioning requirements at fabrication yard” that I have mentioned above. It was dated 5^th May 1995. That specification made no explicit mention of the start-up system of the turbine. But it provided for running the turbine to achieve satisfactory operation of the turbine uncoupled to the compressor, including all trips. That manifestly could not be done without starting it. The specification also provided for achieving satisfactory operation of the coupled turbine/compressor train under roll conditions, which I take to mean slow roll.

Mr. Mobbs, who did not attend the meeting of 4^th May, considered that the slow roll and vacuum tests of the compressor train should be carried out offshore. He sent a memorandum dated 15^th May 1995 to that effect to Mr. Rayner with a copy to Mr. Thurgood. (Mr. Rayner, a contractor and not an employee, acted on behalf of Phillips as project engineer on the 52/5A compression project. He was responsible for co-ordination between Phillips and Snamprogetti in respect of design, procurement and commissioning activities. Mr. Thurgood, an employee of Phillips, assisted with that work). A copy of the memorandum may have been sent on to Snamprogetti. There is a note to that effect on a copy of the memorandum, though there is no oral evidence about it. In fact, the whole test appears to have been cancelled by Phillips. There is no evidence that Snamprogetti were aware of the fact. Mr. Mobbs gave the following evidence in cross-examination (Day 19, p.58):

Mr. Mobbs gave this further evidence in cross-examination (Day 19, p.66):

From the evidence about the actual commissioning of the turbine offshore, I am satisfied that the cause of the difficulty in starting the turbine was a defective combustion chamber: combustion chamber number 2. There were other teething troubles, not relevant to that defect. The nature of the defect does not appear.

My conclusions are these. The factory test of the start-up system incorporated in the design was not carried out because that system was not available at the material time. Both parties expected that further testing of the turbine would take place at AMEC’s fabrication yard. Mr. Thomson wrote a commissioning procedure which would implicitly, but inevitably, involve a test of the start-up system. Phillips decided to abandon such testing onshore, and to carry it out offshore. There is no evidence whether Phillips were aware of the fact that the factory testing had not included the start-up system as designed. There is no evidence whether Snamprogetti were aware of Phillips’s decision to abandon the onshore testing. Evidence on those two matters might have been available from Mr. Rayner, if he had been called as a witness. I am not satisfied that any breach of contract on the part of Snamprogetti in relation to this matter has been established. In particular, Snamprogetti were not under a duty to ensure that the machine was factory tested on methane. Their obligations to review the test procedure and to verify the acceptability of the tests did not include such a duty.

This claim as originally formulated was as follows:

Snamprogetti did not propose to use Upper Bunter gas for summertime storage. Snamprogetti proposed that during the summer shutdown the compression circuits be drained and purged with nitrogen and then kept with nitrogen in them at a pressure slightly above atmospheric. Phillips amended their claim accordingly, and alleged that the use of nitrogen in that way was not practical having regard to the excessive consumption of nitrogen and the need to replenish the system with nitrogen frequently. The amended claim went on to say that nitrogen was replaced both as a purge and as a storage gas by Zechstein gas. That necessitated pipework modifications offshore.

The claim arises out of a recommendation in the Baker Jardine HAZOP. The relevant HAZOP action response sheet (HAZOP action number 3.2.13) proposed manually purging the compressor circuit with Zechstein gas at approximately 5 psig, and then pressuring the circuit with nitrogen up to approximately 5 psig and maintaining that pressure of nitrogen throughout the downtime. It contained a recommendation that Snamprogetti should confirm how Zechstein gas could safely be used as a purge gas without the risk of overpressuring the machine.

The reason why the nitrogen would have been consumed is that it would have leaked out of the compressor through the seals. The leakage would, however, have had the intended advantage of keeping the seal cavities free from dust and debris.

Mr. Tomlinson gave the following evidence in cross-examination (Day 42, p.78):

A meeting was held on 1^st June 1995 between representatives of Snamprogetti, including Mr. Thomson and Mr. Tomlinson, of Delaval Stork, the suppliers of the main compressor, and of Phillips (Mr. Stoker). At that meeting, Snamprogetti expressed their concern regarding the amount of nitrogen usage during the summer period, and stated that a control system would be required with a nitrogen bottle system. On 15^th June 1995 Delaval Stork wrote to Mr. Thomson suggesting an alternative way of pressurizing the compressor with nitrogen where the leakage would be “extremely low”, but no rate of leakage was given. Mr. Thomson replied the following day asking for the magnitude of the required top up “i.e. bottles per [day?] bearing in mind that the platform is unmanned. Please revert at the earliest”. There is no evidence before me that any answer to that letter was received.

On 7^th November 1995 Mr. S. Woodington, project engineer of Snamprogetti, wrote to Delaval Stork. According to the letter heading, copies of it were sent to Mr. Thurgood and to Mr. Stoker and Mr. Jones of Phillips Basingstoke. The letter contained the following passage referring to long term shutdown of the main compressor for the standby summer period:

The letter also mentioned that following the long-term summer shutdown period the main compressor was purged, in accordance with logic requirements specified by Delaval Stork, to eliminate nitrogen from the circuit.

A meeting attended by representatives of Phillips, Snamprogetti, Delaval Stork and EGT was held on 14^th November 1995 to review the compressor logic, and to ensure that there were no problems associated with its interpretation. The minutes of that meeting show that no comments relevant to the present issue were made in relation to start-up, stand by or non-operating modes. According to the minutes, those attending included Mr. Jones, Mr. Stoker and Mr. Hemming of Phillips and Mr. Woodington, Mr. Tomlinson and Mr. Thomson of Snamprogetti.

I conclude that before the Baker Jardine HAZOP was carried out Phillips were aware, without apparent disapproval on their part, of the intention, for the purposes of summer storage, to purge the main compressor with nitrogen and to leave it filled with nitrogen at a pressure slightly above atmospheric, and that a supply of nitrogen bottles would be shipped out specifically for the purpose. Snamprogetti had not, however, ascertained how much nitrogen would be required. It was not until that HAZOP was carried out that the proposal was made to purge the compressor with Zechstein gas. Even then, the proposal was to leave the compressor pressurized with nitrogen.

Mr. Tomlinson gave this evidence (Day 42, p.72):

It is not apparent why the HAZOP members recommended purging the compressor with Zechstein gas. It may have been to save nitrogen, coupled with the view that Zechstein gas was an unsuitable gas for storage. How much nitrogen it would save does not appear. Mr. Sylvester-Evans considered that Zechstein gas was unsuitable for storage since it contained water. Evidently the HAZOP members did not consider that the use of nitrogen for storage would use too much nitrogen.

Zechstein gas had been used for purging during the course of commissioning. A note dated 29^th March 1996 stated that “sweet” gas (i.e. Zechstein gas) had been introduced into the compressor unit for gas purge runs. That had been done by manual operation of valve XV 802 to a pressure of 160 psi. A note of 1^st April 1996 stated that Zechstein gas had been introduced into the compressor and was currently at about 35 psi. It is clear from the evidence of Mr. Sylvester-Evans, including a process flow diagram which he included with his report, that the gas had been introduced from the product gas header via a bypass line back to the main compressor suction header, and hence to the suction of the main gas compressor, at sealine pressure under manual control.

I have mentioned the recommendation of the HAZOP members that Snamprogetti should confirm how Zechstein gas could safely be used as a purge gas without the risk of overpressuring the compressor. They probably had in mind that the process used during commissioning did not avoid that risk. The new design involved taking the supply of Zechstein gas from the fuel gas compressor suction knock-out drum, and later from the fuel gas heater. That ensured that the pressure would not exceed the maximum permissible (250 psig on the inlet side of the compressor).

Mr. McMaster’s argument was that Snamprogetti were in breach of contract in failing to resolve summer storage procedures before the facilities were built. Mr. Tomlinson had, he submitted, accepted that as a result of that failure it had been necessary to come up with procedures in the Baker Jardine HAZOP that had resulted in additional work that should have been done as part of the original construction and installation.

I reject Mr. McMaster’s argument. What Mr. Tomlinson actually accepted was that it was necessary to do further work in order to implement the recommendation of the HAZOP. I am not satisfied that that recommendation was made because it was thought that Snamprogetti’s design would require the use of excessive quantities of nitrogen. It may, for example, have been made because it was thought more convenient to purge the compressor with a direct supply of gas from a well rather than using bottled gas. The only thing shown to have been lacking from Snamprogetti’s design was a statement of the amount of nitrogen required.

Mr. Sylvester-Evans considered that the decision of Phillips to use wet Zechstein gas, with its attendant risk of damaging the seals, rather than nitrogen was a matter of preferential engineering. I accept that evidence.

Dr. Robinson reached the following conclusion in paragraph 16.10 of his report dated 31^st January 2002:

I accept that Snamprogetti’s design was incomplete in that the required quantities of nitrogen had not been specified. In my judgment, it is not correct to describe the consideration given to the matter as scant. Dr. Robinson appears not to have considered the documents by any means as fully as Mr. Sylvester-Evans. That is not a criticism, since I do not know what access he had to them. But in relation to this and other items of complaint I have derived assistance from Mr. Sylvester-Evans’s thorough comments on the documents. I am not satisfied that Snamprogetti’s design was impractical. I am not satisfied that the extra work was necessary.

The claim is in these terms:

The incident in question occurred on 18^th April 1996, during commissioning of the plant. Possibly as a result of erratic operation of the fuel gas recycle valve, the outlet pressure of the fuel gas compressor went outside the permitted limits. That caused the fuel gas compressor and the main compressor to trip. Upon tripping of the main compressor, the compressor circuit was automatically isolated from the wells and the blowdown valve XV 806 of the compressor circuit opened, as it should have done. That was to allow the gas in the compressor circuit to vent to atmosphere. The gas vented through valve XV 806. Valve XV 805 also opened. That valve was not isolated, and gas escaped directly from the Upper Bunter wells through that valve to atmosphere. The release of gas continued for about five minutes until the valve was closed manually.

The error lay in the logic of the system. Valve XV 805 ought not to have opened on such a trip, at any rate without the flow control valves upstream of it being closed.

Snamprogetti admitted that error. It constituted a defect in the design and a breach of contract on the part of Snamprogetti. That is sufficient to decide this issue. However, the questions whether the incident was dangerous and whether it gave rise to a loss of confidence on the part of Phillips in Snamprogetti’s work were strongly contested. I shall give my findings relating to those points briefly.

I am satisfied on the evidence of Mr. Sylvester-Evans, who carefully considered the documents, that members of the Phillips project team approved the various issues prepared by Snamprogetti for the process trip schedule and the cause and effect diagrams. Mr. Sylvester-Evans concluded in paragraph 4.5.24 of his consolidated report:

I accept that evidence, which in my judgment gives a fair assessment.

There was evidence from witnesses called by Phillips that the incident caused danger in two ways. One was danger to helicopters in case the methane entered their engines. The other was the danger to persons on the platform of poisoning by hydrogen sulphide, which is heavier than air and could, it was said, have descended from the stack. I am satisfied that there was no danger. Mr. Sylvester-Evans gave detailed and convincing reasons why the escaping gas would rapidly be diluted by air and that the risk to persons would have been negligible. And the vent stack was designed to take excess gas blowing off directly from the wells. It can be seen from the drawings that near the wellheads there are relief valves. Those are valves that will blow off under excess pressure.

The incident was evidently not thought serious enough by Phillips to merit reporting it in writing to the Health and Safety Executive. Mr. King believed that an oral notification was made (Day 24, p.6). Mr. Halliwell said this (Day 44, p.9):

It appears that the factors that Mr. Halliwell had in mind were to do with the fact that the incident occurred during commissioning, rather than during ordinary operation. I can only conclude that the matter may have been reported orally to the HSE, but that neither Mr. King nor Mr. Halliwell knew whether that was so.

I find that the incident justifiably damaged the confidence that Phillips reposed in Snamprogetti’s work, but that Phillips have blown up the danger aspect out of all proportion.

It is said that two pipelines were incorrectly designed in that the drawings ought to have shown them as sloping, whereas the drawings gave no such indication. The lines were sub-headers for the vent systems. They were line number 1255 for the high pressure vent system and line number 1270 for the low pressure vent system. There is a general requirement to provide slopes for vent pipework to prevent the accumulation in the pipework of liquids, water and liquid hydrocarbons, which natural gas contains. Mr. Tomlinson accepted that there was an error in the design, and that it was not unreasonable to correct it. He was right to do so.

This claim is pleaded as follows. Underlinings represent amendments:

Before the main compressor was introduced, water and liquid hydrocarbons from the Upper Bunter wells were collected in the contactors and removed from the gas stream. The water was sent to the sump tank and the liquid hydrocarbons were re-inserted into the gas stream at the ejector. A quantity of liquid hydrocarbons accompanied the water. Those hydrocarbons were separated out from the water in the sump tank and pumped back into the product gas header by means of the two sump pumps, described in the drawings as gasoline transfer pumps. (The product gas header was a 20-inch line into which Zechstein and Upper Bunter gas was discharged before passing into the H line).

When the compressor was introduced, it was no longer possible to re-insert the liquid hydrocarbons into the gas stream in the former manner, in view of the increase in pressure in the product gas header. Snamprogetti designed the pipework on the basis that all the liquid hydrocarbons would be removed from the gas stream at the contactors or the compressor suction knock-out drum and pass from there into the sump. The existing pumps would pump them into the product gas header.

During the course of design, Snamprogetti carried out a procedure described as a flash simulation, from which they concluded that no liquid hydrocarbons would be present at the operating conditions as provided by Phillips and stated in the Basis of Design. That simulation has not been explored as such. But its conclusions were challenged in the cross-examination of Mr. Tomlinson (Day 42, pp.135 to 143). The relevant passage is this:

In the above passage, Mr. Tomlinson accepted that it was apparently not correct that there was no condensate in the Upper Bunter gas. I take it that he was relying on statements emanating from Phillips that the Upper Bunter wells produced 1.1 barrels of liquid hydrocarbons per 1.1 mmscf. I shall return to that point.

A distinction between the liquids predicted by the flash simulation and entrained liquids was implicit in much of the above questioning. Entrained liquids are liquids in the form of droplets that are carried along the pipework by the flowing gas. Where the speed of flow is low, such droplets tend to collect at the bottom of the vessel. Hence the use of knock-out drums to collect entrained liquids. The large cross-sectional area of the drum compared with that of the pipework reduces the speed of flow. Baffle plates can assist the process.

The method of the flash simulation was not explored. But as I understand the evidence of Mr. Green and Mr. Tomlinson and the documents evidencing the simulation, it is a process of determining whether and if so to what extent water and identified hydrocarbons can exist in the liquid (as opposed to the gaseous) phase under certain thermodynamic conditions (temperature, pressure, etc.) If, as Mr. Tomlinson’s evidence indicates is so in this case, the simulation shows that the hydrocarbons do not exist in the liquid state, that must surely imply that any such liquids formerly entrained as droplets must evaporate or have evaporated. Thus in that case, the simulation does tell you whether any liquid is entrained. The answer is none. If, on the other hand, the simulation shows that hydrocarbons exist in the liquid phase, it does not tell you whether they are entrained or collected at the bottom of a vessel, or partly one and partly the other. I accept Mr. Tomlinson’s evidence quoted above that if you cannot simulate that you have liquids present, then you would assume they could not be entrained.

Mr. Tomlinson gave evidence (Day 42, p.151) that Snamprogetti had been told quite specifically that the contactors had had their internals removed. In the context it is clear that he was saying that it was Phillips who had told Snamprogetti that. That evidence was not challenged or contradicted, and I accept it. That information was incorrect. Removal of the internals from the contactors would prevent the extraction of the liquid hydrocarbons to the ejector and cause them to collect at the bottom of the contactors and pass to the sump. Thus Snamprogetti was unaware that the bulk of the liquid hydrocarbons had by-passed the sump and hence the pumps. Mr. Tomlinson thought, and had reason to think, that the demand on the pumps would not be increased by the introduction of the main compressor.

Nevertheless, Mr. Tomlinson accepted that his misapprehension as to the state of the contactors had been compounded by Snamprogetti’s own error in interpreting the relevant drawings.

Mr. Hobson said in paragraph 52 of his witness statement that the standard figure which Phillips used for entrained liquids coming from the Hewett gas wells was 1.1 bbls/mmscf. (Bbls means barrels). That figure, he said, would have been known to Snamprogetti from their work on the 48/29A compressor project. (1 bbl equals 42 U.S. gallons). Mr. Hobson (Day 32, p.77) accepted that one could not interpret that the figure of 1.1 bbls/mmscf would be the same figure as the figure in relation to any particular well. But he said that it would provide direction and guidance that generally Hewett gas had a composition of 1 to 2 barrels per million.

Mr. Sylvester-Evans said this in his consolidated report:

There was a diary entry of Phillips of 20^th June 1996 which said that one of the pumps was in poor condition, and that the motor was removed for rebuilding. A Phillips handover note of November 1996 referred to problems relating to both gasoline transfer pumps which could not at that time be started or stopped from the control room. There was a handover note of about January 1997 which stated “The production of gasoline is not that much of a problem but should the water dump control fail (due to chamber waxing, sand or outlet to the sea blocking) the level in the water end will rise quickly, thus causing [high] level trip…..All of these problems should be resolved when the new sump tank is fitted during summer shut down”. (Emphasis added).

The capacity of pump P-346 was 36 U.S. gallons per hour; the capacity of pump 67-1 was 180 U.S. gallons per hour. The following is clear from drawing FDD-84-MF-514. The pumps ran in parallel. They were operated by switches which were governed by the level of condensate in the sump tank. The smaller pump was switched on when the level of liquid reached 889 mm above datum. The larger pump was switched on when the level reached 1041 mm. Both pumps were switched off when the level fell to 381 mm. It is clear from the questions put to Mr. Tomlinson and his answers (Day 42, pp.116 to 119) that before the installation of the main compressor the smaller pump had been on more or less continuous duty, and the larger pump was there as an auxiliary for cases where the small pump had insufficient capacity or stopped working.

The highest flow rate of Upper Bunter gas when using the main compressor is shown in the Basis of Design as 74.1 mmscf/d. Using the figure of 1.1 bbls/mmscf gives a flow of liquid hydrocarbons of 143 U.S. gallons per hour. Thus, with the pump logic described above, at that flow rate the smaller pump would be operating continuously and the larger pump would be running for more than half the time. However, even at that flow rate the system could cope with failure of the smaller pump, as it had been able to before the introduction of the main compressor. And at highest flow rates before then, the system could not cope with failure of the larger pump. The fact is that pump 67-1 was not a spare and never had been.

The position is this. Phillips used a standard figure of 1.1 bbls/mmscf for the concentration of liquid hydrocarbons in the gas from the Hewett field. Calculations showed an absence of liquid hydrocarbons under the suction conditions of the new main compressor. How reliable those calculations were does not appear. Mr. Tomlinson’s evidence that he did not know how much reliance Snamprogetti placed on the calculations suggests that not much reliance was placed on them. No doubt Snamprogetti relied on them to show that there was no reason to expect that the new main compressor would give rise to an increase in the concentration of liquid hydrocarbons in the flow of gas. Mr. Tomlinson referred to a reduction in gas flow (arising out of the depletion of the wells) but the compressor gave rise to an increase in flow over the immediately preceding values. Snamprogetti concluded that there was no reason to expect an increase in flow of liquid hydrocarbons. That conclusion itself has not been criticized.

Snamprogetti made an error in failing to appreciate that not all of the liquid hydrocarbons that had previously been collected in the contactors had passed into the sump. In my judgment, Mr. Tomlinson was right to admit that error. It arose from a failure to read the relevant drawing with sufficient care. No doubt it was induced, at any rate in part, by incorrect information from Phillips. But in my judgment it amounted to a failure on the part of Snamprogetti to fulfil the duties it owed to Phillips under the contract.

In my judgment, no consequences giving rise to liability on the part of Snamprogetti arose from that breach of duty. Following the Baker Jardine Hazop, Phillips decided to replace pump P-346 with a larger pump. If their standard figure of 1.1 bbls/mmscf, or anything like it, were applicable to the suction side of the main compressor, the demand on pump 67-1 would be increased by the introduction of that compressor. Thus the decision to replace pump P-346 with a larger pump which could take the whole flow if pump 67-1 failed would not have been unreasonable. But I am by no means persuaded that all engineers would, even in those circumstances, consider it necessary. I find that Snamprogetti’s failure to design for the replacement pump did not itself constitute a breach of duty. Moreover, I am by no means satisfied that the actual flow of liquid hydrocarbons experienced by Phillips after the introduction of the main compressor justified the installation of the new pump.

The most that can be said is that if Snamprogetti had told Phillips that their standard of 1.1 bbls/mmscf implied a flow of 143 gallons of liquid hydrocarbons per hour, Phillips might have decided to replace pump P-346 before they actually so decided. If Snamprogetti had not made the mistake they did, it is possible that they would have told Phillips that. It is, indeed, a matter of elementary arithmetic following from gas flow figures of which Phillips were aware. In my judgment, Snamprogetti were under no duty to give the figure of 143 gallons an hour to Phillips.

Having regard to the information contained in the Phillips documents mentioned above, there is reason to think that Phillips might well have replaced pump P-346 in any event. Phillips have failed to satisfy me that they would not have done so. The burden lies on them to prove their loss, and in respect of this item they have failed to do so.

This is another claim that has been heavily amended. It relates to the drainage of the water used for washing the blades of the gas turbine that drove the main compressor. It is common ground that it was important that that water should drain away without risk of flowing back into the turbine casing. Thus it must drain at no more than atmospheric pressure.

The original complaint was that Snamprogetti designed the drain line to connect to a drain system that was above atmospheric pressure. It is true that Snamprogetti designed the drain line to connect to the sump tank, and that the sump tank was at a pressure above atmospheric. It should not have been at a pressure above atmospheric pressure. It had a six-inch pipe leading to a 12-inch low pressure vent header. The fact that it was above atmospheric pressure was not the fault of Snamprogetti.

The claim was amended so that the material part read as follows:

The description of the drainage system designed by Snamprogetti as described in the claim, though in accordance with drawings, was not in accordance with the system as built. Although the relevant drawing showed a two-inch header running into a ¾ inch line, that was not the situation on site. It ran into a two-inch line which then ran into a one-inch line. That was the only bottleneck.

On the evidence I am satisfied that the back pressure did not arise from the bottleneck. It arose from the sump tank. Mr. Sylvester-Evans, after a careful discussion of the documentary evidence, concluded (paragraph 4.8.35 of his consolidated report):

Phillips solved the problem by re-routing the turbine water wash drain directly overboard. They used 8 metres of ¾ inch diameter piping for the purpose. It was unnecessary for Phillips to alter the drainage piping designed by Snamprogetti, and their doing so was no solution to the problem.

Dr. Robinson’s report of 31^st January 2002 reflected the confused history of the pleading. He said that Snamprogetti correctly piped the water to flow to the sump tank. The piping created a bottleneck, which was a design error and had to be corrected. There was back pressure on the line. In order to save time Phillips re-routed the wash line directly to sea. That allowed commissioning work on the turbine to continue. A Snamprogetti suggestion that the cause of the back pressure be identified and rectified rather than re-routing the drains was valid only when the project had no constraints of time. The investigation would stop all commissioning work on the gas turbine. The project did not allow time to investigate, when a simple solution was possible. Poor piping design was probably the sole cause of the back pressure. The routing of the piping from the turbine water wash was a very clear design error. The error had to be corrected. The water wash was re-routed to sea, as a matter of expedience. That was entirely reasonable in the circumstances.

In that report, Dr. Robinson appears to have contradicted himself on whether the routing of the piping to the sump tank was a design error. (His reference to the routing of the piping must have been a reference to the directing of it to the sump tank rather than to sea or some other destination, not a reference to the route by which the line led to the sump tank, since no question of an alternative route from the turbine to the sump tank was raised). In his report of 28^th August 2002 he quoted an expression showing that for a given rate of flow of a fluid in a pipeline the pressure drop per unit length of pipe was inversely proportional to the fifth power of the diameter of the pipe. Thus one metre of one-inch pipe would create approximately as much pressure drop as 32 metres of two-inch pipe. Bottlenecks would add a significant further pressure drop. He did not back up that evidence with any figures to show the practical effect in this case.

Dr. Robinson clarified his position in cross-examination (Day 62, pp.171 to 173). He said that there were here two problems which might or might not be connected. One was that the water wash would not correctly drain to the sump tank. As a matter of expediency and to resolve it quickly, Phillips put the line overboard. The other was that the bottleneck could not have been allowed to remain. It might have been the reason for the water wash problem, or it might not. He did not know. He did not know what was the cause of back pressure on the line. In a way, the water wash was a red herring which alerted people to the fact that there was a problem, namely the bottleneck, which he did not think anybody would have allowed to exist.

There was a pre-existing drain line on platform 52/5A which involved a 1½ inch drain line running into a ¾ inch drain line. Miss Boswell Q.C. put the relevant drawing to Dr. Robinson in cross-examination (Day 65, p.110):

Mr. Abernethy gave this evidence in the course of his cross-examination (Day 26, p.132):

It was put to Mr. Tomlinson in cross-examination (Day 42, p.159) that it was bad engineering practice to put the flows from a two-inch line into a three-quarter-inch line. He said that it was poor engineering practice at best. Doubtless he would have said the same or something similar if the two-inch to one-inch bottleneck has been put to him. He accepted (Day 42, p.161) that if Phillips had carried out work after the installation of the compression facilities to ensure that the first two-inch line ran all the way to the second two-inch line, that would not be an unreasonable thing for them to do.

I am satisfied that it was poor engineering on the part of Snamprogetti to run the two-inch pipe into the one-inch pipe. They were thereby in breach of contract. I am satisfied that that breach did not adversely affect the drainage of the turbine wash water. I cannot accept that nobody would have allowed the bottleneck to exist. There was a pre-existing bottleneck which Phillips had allowed to exist. Assuming that the pleading represents Phillips’s state of mind, Phillips appear to have removed the bottleneck under the misapprehension that it caused the problem with the drainage. The bottleneck was a red herring, to adopt the language of Dr. Robinson. Given that Phillips were evidently content to have another bottleneck in a drainage system, I am not satisfied that they would have wished to remove this one if they had appreciated the true position. No other reason has been put forward why Phillips should pick on this bottleneck for removal. It would be an arbitrary choice, and if so then in my judgment baseless and unreasonable. If Phillips had not in fact removed the bottleneck, they could not have recovered more than nominal damages for the breach plus possibly any fee paid for the relevant engineering. Phillips have failed to prove the case that they pleaded, and in my judgment are entitled to no more than such damages for this item.

There were seals at each end of the main compressor, between the rotating shaft and the casing. Gas was used to provide a dry seal which prevented the loss of process gas along the shaft of the machine and out to the atmosphere. The gas used for sealing was itself process gas which was bled off from the discharge from the compressor. At that point it was warm and at high pressure. It was piped through heat-insulated pipework to a coalescer and then upwards through more insulated pipework to the seals. The coalescer contained filters designed to remove any solid particles and droplets of liquids. Any water entrained with the sour Upper Bunter seal gas would have been particularly corrosive and possibly damaging to the seals, though they were made of stainless steel. The presence of solid particles and water would have reduced the operational life of the seals. The seals were expensive items. They cost about £80,000.

Phillips claim that the pipework from the coalescer to the seals should have been heat traced. That is to say, the insulation should have contained wiring so that an electric current could be passed along it to generate heat and ensure that the gas in the pipework remained warm enough to prevent condensation of any water vapour in it. The pipework in question was not heat traced. Phillips claim that Snamprogetti were thereby in breach of their contractual duty.

There was an internal memorandum of Snamprogetti, apparently dating from about June 1993, stating “All seals to be traced electrically for dry gas seals”. It is evident that that was not pursued. It was put to Mr. Tomlinson in cross-examination that it was overlooked. Mr. Tomlinson said (Day 43, p.9):

Dr. Robinson’s opinion was this. The decision on whether tracing should be used depends on costs. The cost of heat tracing, when carried out at the correct time, is small, probably of the order of tens of pounds. The cost of replacing the compressor seals is several tens of thousands of pounds. Loss of production also needs to be considered. The provision of heat tracing would have given a greater certainty that the seal gas was dry and hence less corrosive. It was a balance of expenditure of a few tens of pounds against the possibility of a loss of £80,000. The failure to provide heat tracing was a design error. The work, in relation to this item, was not that of a first-class engineering contractor. (Dr. Robinson’s first report, paragraphs 21.2 to 21.4; Day 61, p.97).

The following exchange took place during the course of Dr. Robinson’s cross-examination (Day 65, pp.126 to 129):

Mr. Sylvester-Evans considered that not specifying heat tracing for the seal gas lines was not a Snamprogetti design error. Snamprogetti were entitled to rely on the specialist vendor’s experience. He was of the opinion that the introduction of the heat tracing following the Baker Jardine Hazop was a matter of preferential engineering.

I accept Dr. Robinson’s opinion that the decision whether heat tracing should be used depends on costs, and that the decision involves balancing a small expenditure against a small risk of a large cost. In assessing that balance, a first-class engineer is entitled, in my judgment, to take into account the relevant experience of the specialist supplier. Mr. Thomson’s recall of his thought processes was incomplete (see Day 37, p.129), and it may be that he simply relied on the specialist supplier without himself assessing the balance. However that may be, I am satisfied that the design was such as could have been produced by a first-class engineer. I accept Mr. Sylvester-Evans’s opinion that this was a matter of preferential engineering. This item of claim fails.

I should add this. The amount of expenditure apportioned to this work by Mr. Farrow is £28,547. The amount apportioned by Mr. Charters is £11,912. It has not been suggested that in the absence of the heat sealing the risk of having to replace a seal (let alone of losing production of gas in consequence) is a substantial risk, nor, I think, could it have been. The claimants have not suggested that such expenditure could have been justified at the time of design. It follows that it could not be justified later. Thus even if I am wrong on the question of breach of duty, the damages in respect of this item would not be measured by those sums.

The start-up sequence of the main compressor was said to be incorrect in the following way. On the suction side of the main compressor, there were two valves in parallel, a large valve XV 803 and a small valve XV 802. When the compressor was to be started up, the compressor system was first pressurized. That was done by opening valve XV 802, whilst XV 803 remained closed. When the compressor started up, it was necessary to open valve XV 803. Owing to a mistake in the logic, it was not possible to open valve XV 803. To correct the logic, it was necessary to insert upstream of valve XV 803 a pressure transmitter to provide automatic control of three flow control valves FV 944, 945 and 946 on the contactors, upstream.

The evidence from witnesses for both parties was all one way: this constituted an error on the part of Snamprogetti. I find that Snamprogetti were thereby in breach of duty.

This matter was raised by way of amendment. It concerns the anti-surge controller. (For an explanation of surge, see paragraph 185 above). It arose out of a remark made by Dr. Robinson in his first report, dated 31^st January 2002. He said this:

The flow meter measured the pressure drop across an orifice plate. In order to know the flow rate through an orifice plate, whether the volume flow rate or the mass flow rate, the pressure drop across the plate and the density of the gas must be determined. I am satisfied of those facts. There is apt to be confusion between the pressure difference across the orifice plate, necessary to measure the flow accurately, and the pressure difference across the compressor.

The density of the gas depends on its temperature and pressure. Dr. Robinson gave the following evidence in his report of 28^th August 2002, paragraphs 20.4 and 20.5:

Dr. Robinson and Mr. McMaster relied on evidence of Mr. Tomlinson to demonstrate that Snamprogetti were at fault. His relevant evidence is as follows (Day 43, pp.13 to 19):

I am satisfied that the expression put by Mr. McMaster to Mr. Tomlinson was correct as applied to orifice plates. It appears in International Standard ISO 5167-1980 to which Dr. Robinson referred me. But I cannot accept even so that Mr. Tomlinson was right to accept that if the expression were right, it was necessary to instal the pressure transmitters.

I accept Mr.Tomlinson’s evidence about the refusal of Delaval Stork to use the pressure as an input to the algorithm, and as to his discussion and agreement with Mr. Wells.

Mr. Sylvester-Evans was cross-examined at length about this. He did not pretend to know in detail how the anti-surge device worked. But I am satisfied that he had a sufficient understanding of the relevant thermodynamical considerations. He explained that the original design of Snamprogetti had provided three inputs of information for the device: the pressure drop across the flow meter, or flow transmitter FT 820; the pressure difference across the main compressor, given by the pressure difference transmitter PDT 853; and the setting of the speed controller of the main compressor. I would add that since the device governed the operation of the main compressor recycle valve FV 820, the setting of that valve must have constituted information available to the device. Whether it used that information does not appear. Mr. Sylvester-Evans explained that the device used the ratio between the pressure drop across FT 820 and the pressure difference across the main compressor to calculate how close the compressor was to the surge condition. If the compressor was too close to the surge condition, the device caused the flow through the recycle line to be increased. The device contained some algorithm to make that happen. It is true that its calculation was not precise, but the device would have been adjusted during commissioning to give satisfactory results.

I quote only a small part of the lengthy cross-examination of Mr. Sylvester-Evans on this point (Day 76, pp.31 to 35):

Pursuant to item 3.2.3 of the Baker Jardine Hazop, a proposed change was recorded by Snamprogetti in a P & ID change record form dated 29^th May 1996. It was to replace PDT 853 with two pressure transmitters, one connected to the compressor suction and one connected to the compressor discharge, with outputs connected to the anti-surge controller. A Hazop action response sheet dated 23^rd (?) September 1996 referred to that form and stated:

Mr. Sylvester-Evans gave this evidence (Day 75, p.158):

I do not think that the uncompensated reading of FT 820 can give the actual volume flow. Nevertheless on the totality of the evidence I am satisfied, as Phillips appear to have been at the time, that Delaval Stork knew what they were doing in designing the anti-surge flow controller, and that Snamprogetti properly designed the systems to operate it. There is no evidence that it does not work.

This claim fails.

This claim arises out of the Baker Jardine Hazop. Recommendation 2.2.3 referred to valve XV-895. That is a blowdown valve upstream of the fuel gas compressor suction knock-out drum. The recommendation was that the downstream valve V-832 should be locked open and “FO-896” should be shown as “RO”. Then in heavy type there followed the recommendation “Both these changes to be general for all valves downstream of blowdown valves and restriction orifices”. The effect of the recommendation was twofold: first, that all valves downstream of blowdown valves should be locked open; and second, that all restriction orifices should be labelled “RO”.

This claim relates to the second of those recommendations. The purpose of installing restriction orifices downstream of the blowdown valves was to restrict the flow rate of gas entering the vent header so that the vent system would not be over-pressurized by an excessive flow. The restriction orifice is simply an orifice plate installed in the pipework. The orifice plate is a metal disc with a central hole, the flow requirements determining the diameter of the central hole. A flow orifice is essentially an identical device. It consists of an orifice plate with small tappings upstream and downstream of it to allow the differential pressure across the orifice to be measured. The orifice plate for the flow orifice is normally of a thinner construction than that for the restriction orifice.

Snamprogetti provided a general instrumentation specification for orifice plates and restriction orifices. It was re-issued for engineering on 11^th January 1995. In section 7 it specified that orifice plates should be provided with stainless steel labels engraved with the Phillips tag number, the supplier’s name, the supplier’s serial/equipment number and the orifice diameter. It appears that Snamprogetti specified that they should all be tagged as “FO”, meaning flow orifice.

Mr. Green, in paragraph 76 of his first witness statement, said this:

Dr. Robinson said this in his report of 31^st January 2002:

Mr. Tomlinson gave this evidence during his cross-examination (Day 43, p.20):

I accept that evidence of Mr. Tomlinson. Dr. Robinson said in his report dated 28^th August 2002, with reference to that evidence:

I see the force of Dr. Robinson’s opinion, but it is apparent that there are two opposed views about the labelling of orifices. I am not satisfied that there was any fault in Snamprogetti’s specification in this respect. Mr. Sylvester-Evans considered that the change recommended by the Baker Jardine Hazop was a matter of preferential engineering. In my judgment, he was clearly right.

This is the first of the two items mentioned in paragraph 344 above as having been recommended by recommendation 2.2.3 of the Baker Jardine Hazop. The recommendation refers to these valves as being downstream of the corresponding blowdown valves. According to Dr. Robinson and to Mr. Tomlinson, they are upstream of the restriction orifices. Mr. Sylvester-Evans said that they were downstream of the orifices. I prefer the evidence of Dr. Robinson and Mr. Tomlinson in this respect, though the question is immaterial.

Dr. Robinson wrote in his first report:

I do not accept that the vent system would be rendered inoperative. There remain the pressure relief valves, and I accept evidence of Mr. Sylvester-Evans that there would be no risk of over-pressurizing the piping. However, as he said, in the event of an emergency, if one of the valves were inadvertently left closed, it would prevent the blowdown of a section of plant, and that could be hazardous.

Mr. Tomlinson agreed that the valves upstream of the restriction orifices should be locked open. He said that Snamprogetti had found one case where it had not been shown (on the drawings) that the valve was to be locked open.

Mr. Sylvester-Evans wrote in his consolidated report of 18^th September 2002 that he had checked all the automatic blowdown connections and, with one exception that was noted by the Hazop team, all valves downstream [sic] of the orifices were shown on the process and instrumentation diagrams to be locked open.

I find that Snamprogetti were in breach of contract in relation to that one valve in failing to show it on the drawings as to be locked open.

Before the main compressor was installed there was provision for blowing down the Zechstein gas metering skid for maintenance purposes by means of a connection from a point immediately upstream of the ejector to the vent header. The ejector was later isolated. That isolated the connection to the vent header. Thus that connection could no longer be used for blowing down the metering skid. Phillips claim that Snamprogetti should have devised another blowdown system, but that they overlooked the need to do so. It had to be installed later.

I am satisfied on the evidence of Mr. Tomlinson (Day 43, p.31) that the ejector was isolated by means of blinding, that is, the insertion of semi-permanent spades into the line so as to block it each side of the ejector. Thus it would not have been possible, as suggested by Mr. Sylvester-Evans, to have gained access to the blowdown line by opening a valve upstream of it.

Mr. Tomlinson said that a simple and economical solution would have been to remove a pressure gauge PI 0526 (actually PI 6526) and run a hose from that connection to a suitable vent line. There were probably several instances where the P & IDs provided for such an arrangement. He was not able to say that it was a convenient route.

Dr. Robinson considered that such an arrangement was unrealistic. He considered (Day 61, p.125) that the volumes of pipework that needed to be evacuated were sufficiently large to warrant a hard piping connection. He said:

In his consolidated report, paragraph 4.13.11, Mr. Sylvester-Evans noted that OIM (offshore installation managers’) handover notes of 24^th April 1996 recorded that following an emergency shutdown there was a problem with valve XV 940. Valve XV 940 was located downstream of the metering skid and the ejector. The note stated “Had to vent from fuel gas take off (cellar deck) to compressor”. Mr. Sylvester-Evans concluded (ib., paragraphs 4.13.15, 4.13.16):

I reject Mr. Sylvester-Evans’s view that this was a matter of preferential engineering, particularly since the very facilities required had existed before the ejector was isolated.

Mr. Green stated in his witness statement that the metering skid and ejector system did not form part of Snamprogetti’s modification works in the contract for detailed design engineering works. Miss Boswell submitted that Phillips appeared to have been making their own decisions about the ejector and its blinding without full reference back to Snamprogetti. I nevertheless find that Snamprogetti had a sufficient involvement in the isolation of the ejector (and hence of the blowdown line) for them to be regarded as having undertaken duties in relation to the design. I find that this was a defective design and that Snamprogetti were in breach of duty.

After the ejector was isolated, Phillips caused the removal of a spool piece (i.e. pipework) connecting the ejector suction header for Upper Bunter gas to the ejector. That left a dead end of the header closed by a blind flange. Upstream of that blind flange there was a valve, which was to be kept closed. If that valve should leak, gas pressure would build up behind the blind flange. At the Baker Jardine Hazop it was recommended (node 2.1.2) that a pressure gauge be fitted to the blind flange. The object was to enable workers who might have to remove the blind flange, e.g. if the spool piece were to be reinstated, to know in advance whether there was gas pressure behind the flange.

Mr. Green, Mr. Tomlinson and Mr. Sylvester-Evans all considered that the fitting of the gauge was a matter of preferential engineering. Dr. Robinson did not share that view. He regarded the addition of the gauge, at a relatively low cost, as an important safety aid. However, he freely acknowledged that this item, alone, might be viewed by some competent engineers as preferential engineering. On that evidence, I am satisfied that this claim must fail.

Miss Boswell did not accept that this item fell within Snamprogetti’s design. She submitted that the decision to remove the spool piece without fitting a pressure gauge was taken by Phillips without reference to Snamprogetti, and cited a document in support. That point was not answered. However, it is unnecessary for me to decide it.

It is unnecessary for me to explain this item, since there is no evidence of any breach of contract and the damages claimed are nil.

These two claims have tacitly been abandoned.

Dr. Robinson said that Snamprogetti were in error in failing to amend Phillips drawing P & ID 504 by the deletion of a reference to PSD4 and the substitution of a reference to PSD3. That would have caused those responsible for the electronics to modify the shutdown logic so as to cause the main compressor to shut down when the Upper Bunter contactor valves were closed. The main compressor needed to be protected directly the valves were closed, not as a consequence of closing those valves. The revision to the shutdown logic was essential to the safe and reliable working of the process.

This matter has not been explained in evidence in any sufficient detail to satisfy me that Snamprogetti are at fault in this respect.

The Baker Jardine Hazop at node 2.2.2 recommended that the fuel gas compressor start sequence should automatically drain the suction piping prior to starting the machine. The Hazop worksheet stated that for operational reasons, start-ups should be as similar as possible to machines on other platforms. The recommendation was made in order to make the purge and start operations the same for all compressors in general. I take it that “all compressors in general” refers to Phillips compressors.

Dr. Robinson made little of this. He said (Day 60, p.62):

Mr. Sylvester-Evans said that it was unlikely that liquids would be formed in the suction piping. Liquid found downstream of the aftercooler during commissioning was a consequence of the lack of superheat. He said in paragraph 4.1.7.43 of his consolidated report:

Whether or not Snamprogetti relied on Meller’s advice, in my judgment Mr. Sylvester-Evans was wrong to say that they were entitled to do so, in the sense that the receipt of such advice absolved Snamprogetti from responsibility to Phillips on the point.

On 22^nd July 1996, Snamprogetti wrote to George Meller expressing their surprise that there were no convenient drain ports on the fuel gas compressor for the purpose of fitting an automatic drain valve. That supports Dr. Robinson’s evidence that it is a fairly standard procedure to have such valves. George Meller’s reply dated 6^th August 1996 stated that unless specifically requested it was not their practice to provide automatic drain ports. It was their understanding that the process operating conditions and inlet knock-out pot would not necessitate additional drains. However they had no objections to those modifications being carried out. That supports the case that such valves were not necessary on this system.

I am not satisfied that Snamprogetti fell short of the standards required of a first class contractor in this respect.

There is no doubt that George Meller supplied an unsuitable valve for the fuel gas compressor recycle valve PV 923. The allegation against Snamprogetti is that the valve was wrongly specified by Snamprogetti. The valve as supplied turned out to be grossly oversized. I accept evidence from Mr. Tomlinson (Day 41, pp.9, 21) that Snamprogetti left the sizing of the valve to George Meller. Snamprogetti expected to have George Meller’s calculation for checking. But it appears that it was not received or checked. In my judgment, Snamprogetti’s failure to follow the matter up and check the calculation constituted a breach on the part of Snamprogetti of the duty they owed to Phillips.

The fuel gas compressor was protected against pressures above 240 psig by a HIPS. In his first report, dated 31^st January 2002, paragraph 13.3, Dr. Robinson expressed his view about a suction pressure control valve. He said that a suction pressure control system should have been installed downstream of valve XV-892. That accorded with the recommendation contained at node 2.2.5 of the Baker Jardine Hazop. The relevant part of the recommendation stated:

With reference to the suction pressure control valve, Dr. Robinson continued:

At node 2.2.4, the Hazop worksheet reported that the demand on the HIP system would be reduced by the recommendation 2.2.5 adding an upstream pressure control valve. It might not be necessary to initiate an ESD-2 (a shutdown procedure) from the HIP. One of the recommendations was to recalculate the protection value of the HIP given the deletion of the ESD-2 action, and to determine if the reliability was within acceptable limits (to be done by Snamprogetti).

Dr. Robinson gave a further explanation of his view in the course of his examination in chief (Day 60, p.77):

Mr. Sylvester-Evans agreed that it might be necessary to change the set point of the HIPS. It was not a costly exercise. It would be part and parcel of a review of how the system was set up for use.

The foregoing evidence of Dr. Robinson and Mr. Sylvester-Evans was given in relation to the suction pressure controller actually introduced. That limited the suction pressure to 84 psig. I have found such a controller to be unnecessary, and indeed undesirable. A controller limiting the pressure to 175 psia (160 psig) would have been sufficient (see paragraph 224 above). It does not appear whether any adjustment of the fuel gas compressor HIPS would have been necessary if such a controller had been installed. Thus I am not satisfied that any breach of duty on the part of Snamprogetti in this respect has been shown.

A further point was raised under this heading. The Baker Jardine Hazop node 2.2.4 included a recommendation to add “LO” (meaning “lock open”) to all suction valves on the impulse lines for the HIP system. Dr. Robinson explained (Day 60, p.78 and first report, paragraph 13.11) that the pressure detectors had valves in the lines to them which needed little locks to keep them locked open, because you must not be able to shut off the pressure to those HIPS detectors, otherwise they become inoperable. That was another design error on the part of Snamprogetti.

Mr. Sylvester-Evans said (Day 69, p.18) that he thought that if one of the valves in question were left closed a malfunction should show up in the distributive control system. If there are too many locks people do not replace them. It was the subject of debate in the industry, and a matter of preference. I accept that this is a matter of preference, and not something for which Snamprogetti are liable.

This item was the subject of a recommendation in Baker Jardine Hazop node 2.3.9. The recommendation was to change the logic to ensure that a trip of the turbine and main compressor led to a trip of the fuel gas compressor package. It was stated that the existing logic might lead to an abort of the fuel gas compressor start-up (sc., after a trip of the turbine and main compressor).

There is insufficient evidence to support this claim.

In addition to the question of vibration, the claimants relied on other alleged defects in the fuel gas system as increasing the risk, as of June 1996, that the fuel gas compressor could not be put into operation by the end of October 1996. It was said in the particulars of claim that on their own those defects could have been remedied individually, and in some cases were. Some gave rise to monetary claims, and some did not. Those that gave rise to monetary claims are mentioned above. The claimants relied on most of those in this connection also.

The items which do not give rise to monetary claims but upon which Phillips rely as adding to the risk that the fuel gas compression system could not have been made to work by the end of October 1996 are the following items.

A Baker Jardine Hazop action worksheet for node 2.3.2 contains a recommendation that the setting of a temperature trip and alarm should be updated to reflect the new operating range arising from the introduction of the fuel gas compressor suction control valve. The specification of the work, as explained by Mr. Sylvester-Evans by reference to a Hazop action response sheet, was to reset the trip from 212F to 280F. It is not clear whether the work was actually done; probably not, since it was relied on in support of the above-mentioned risk. Snamprogetti did, however, revise the instrument process sheets accordingly. Mr. Sylvester-Evans said that the adjustment would be readily achieved by changing the instrument set points on the control panel. That was a simple routine operation readily achieved by operators. I accept that evidence.

It does not appear whether this adjustment would have been necessary with a suction pressure control valve set at 175 psig. I doubt it. But the work is trivial in any event.

During commissioning, it was found that when the turbine was started the outlet pressure of the fuel gas compressor fell by about 50 psi, causing the system to trip. Dr. Robinson attributed that, correctly in my judgment, to an excessive ratio between the volume of gas downstream of valve XV 924 and the volume of gas between the discharge of the fuel gas compressor and that valve. The principal volume in the latter section was that of the fuel gas receiver. Dr. Robinson said that it should have been about twice as large as it was. He said that the pressure dip should have been a fundamental item for a design check. Snamprogetti had completely failed to check the volumes. But he also said that a larger receiver would not have fitted within the fuel gas skid area.

Mr. Tomlinson accepted that Snamprogetti had been at fault, but the error lay in an excessive volume of pipework downstream of XV 924. He explained the matter thus (Day 40, p.146):

The fault was indeed remedied by altering the start-up procedure so that valve XV 924 opened before the fuel gas compressor started up. That was recommended by the Baker Jardine Hazop and, according to response sheet for node 2.3.7, carried out by 31^st July 1996.

This matter arises out of Baker Jardine Hazop node 2.4.1. It relates to the lubrication of the crankcase. The lubricating oil system took too long to fill upon start-up of the fuel gas compressor. On start-up the system took 40 to 60 seconds to fill the oil cooler and then lubricate the crankcase. The crankcase needed to be lubricated within ten seconds. Mr. Sylvester-Evans explained that the design problem was essentially the back-siphoning of oil caused by the elevation of the oil cooler above the crankcase. That led to problems of priming the pump and filling the oil cooler within the desired ten seconds from starting the compressor.

The relevant Hazop worksheet, dated 23^rd May 1996, recorded:

On the same day, 23^rd May, Mr. Thomson of Snamprogetti telephoned the package vendor, George Meller, and received the following information by fax:

The source of the information contained in the Hazop worksheet is evidently a communication from George Meller Ltd. to Mr. Thomson dated 14^th May 1996. Only part of that communication is before the court. It includes a diagram showing the oil cooler blocked off and a bypass to be installed at the same level as the crankcase. It also shows a sump heater with the legend “Heater maintain between 5C and”, and in the text it states “GML/Burton Corblin confirm that for this particular compressor duty the oil cooler is not required provided lube oil temperature is maintained between 5C and”. The legend in full must have read “Heater maintains temperature between 5C and 15C” or words to that effect; and the second extract must end “between 5C and 15C”. There is no argument about that.

It is clear from the later communication that the proviso in the communication of 14^th May must be read as referring to the fixed points of the sump heater. Thus the record quoted above in the Hazop worksheet was based on a misunderstanding. The recommendation in the same worksheet was answered in the later communication from George Meller: namely, that the temperature of the oil should not exceed 70C. It was also stated that with ambient temperatures within the range specified (minus 6.5C to plus 26C) the oil temperature would not exceed that limit.

Mr Thomson nevertheless wrote back to George Meller on 24^th May asking for an assurance based on thermodynamic calculations that the oil could be maintained at its operating temperature. He pointed out that with normal mineral oils there was a danger of destroying the additives with temperatures in the region of 70C. He required an assurance that the oil stipulated for the project could operate at 70C without deterioration.

After further correspondence, it was not until 8^th August 1996 that a copy of a letter from the manufacturers of the fuel gas compressor, Burton Corblin, to George Meller was forwarded to Snamprogetti. By that time, the fuel gas compressor was treated by Phillips as lower priority work (see paragraph 96 above). The letter said:

Snamprogetti had not received the calculations Mr. Thomson had asked for, and the letter from Burton Corblin was not conclusive. On the one hand, Burton Corblin’s own experience allowed an ambient temperature exceeding the specified limit by 14C without use of an oil cooler. On the other hand, the ambient temperature limit applied where the absorbed power did not exceed 37 kW, whereas the power absorbed by the fuel gas compressor could be 45 kW.

Mr. Sylvester-Evans expressed the view that the cooler was not required. Mr. Thomson said (Day 37, p.141) that the oil cooler could be bypassed by a simple modification in the pipework. Mr. Sylvester-Evans said that there were a number of potential solutions. If a cooler were retained, the pipework could readily be modified to break a siphon. I accept that evidence.

Mr. McMaster submitted that at the date (17^th June 1996) when a decision had to be made whether or not to instal a fuel gas heater, it was not clear whether the remedy was to remove the oil cooler or instal a different system. But in my judgment this was not a major problem, and it must have been obvious that it could readily be solved.

Of these alleged items, only two additional vents were in fact required. I am satisfied that those vents ought to have been included by Snamprogetti by way of notes on the P & IDs (process and instrumentation diagrams) and isometrics stating that the vents should be provided by the installation contractor. The vents were added by KYE. The work took two days and was completed on 1^st September 1996.

The piping on the discharge side of the fuel gas compressor leading to pressure relief valves PSV 919A and B contained a pocket, i.e. a low point where liquid and detritus could accumulate. I accept Dr. Robinson’s evidence (Day 60, p.66) that that pocket potentially compromised the reliability of the relief valve system. He continued:

I am satisfied that Snamprogetti were in breach of contract in this respect. Snamprogetti had suggested heat tracing and insulating the line, but that would not have avoided the possibility of the lodgement of detritus.

Although it appears that it was not easy to devise a solution in the abstract, there is no evidence to suggest that there was any particular difficulty to a skilled person on the platform in finding a solution.

I am satisfied that these five items involving no monetary claim are together wholly insufficient to give rise to a significant risk, viewed as of 17^th June 1996, that any necessary work to solve the problems could not be completed by the end of October 1996.

In paragraph 28 of his supplementary report, dated 22^nd March 2002, Dr. Robinson listed the items that in his opinion affected the relevant risk. He pointed out that uncertainty of success might arise for several reasons, e.g. unknowns in the whole procedure, non-delivery of equipment when required, and weather conditions limiting the working period. He gave what he described as his optimistic assessment of the probability of the completion of each task by 21^st October 1996, to allow for commissioning by the end of that month. His assessment was as follows:

In his second supplemental report, dated 28^th August 2002, Dr. Robinson recognized that having a suction pressure control valve set at 99 psia would prevent the fuel gas compressor from supplying sufficient fuel gas to the main compressor turbine to provide the power required in the early years. The fuel gas compressor would manifestly have to run faster and the frequency of the vibrations would change. He said (paragraph 7.4):

Mr. Sylvester-Evans recognized that it might take some time to find the right valve to replace the recycle valve PV 923, but he was in no doubt that it would have been resolved well before October 1996, applying a sensible level of resources of Phillips, Snamprogetti and the vendors. I accept that evidence. Viewed as of June 1996 the same conclusion would have been reached.

There are two further items relied on by Phillips in this connection. The first is the drainage of suction pipework to the fuel gas compressor. The relevant work took six days and was completed on 30^th August 1996. It was manifestly not a serious problem. The other is the updating of the shutdown logic to meet the situation where the turbine and main compressor trip. I have already said that there is insufficient evidence to support this claim. The same applies to the assessment of risk.

I conclude that none of the matters relied on by Phillips as increasing the risk, as of June 1996, that the fuel gas compressor would not be capable of being put into operation by the end of October 1996 have, individually or collectively, any significant effect on that risk.

There were severe overall electrical power constraints on the 52/5A platform. Minutes of a meeting held between representatives of the parties at Bacton on 23^rd August 1994 include the following item:

The reference to load is a reference to electrical power load. The reference to the turbine is a reference to the turbine to drive the main compressor. The expression “generation” refers to the generation of electrical power on the platform.

The suggestions made by Phillips were not pursued at the time. But the minute indicates the situation concerning electrical power.

It is common ground that the electrical power consumption of the fuel gas heater installed was less by 20 kW than that of the fuel gas compressor.

A document emanating from Phillips dated 26^th March 1996 giving justification and premises for an authorization for expenditure on a power generation upgrade on platform 52/5A states the following. The power generation equipment consisted of two gas turbine generator sets rated at 300 kW each and one diesel generator set rated at 100 kW. The performance of the gas turbine generators had deteriorated. Their present output was 260 kW. The estimated platform load during normal [main] compressor operation was 253 kW. If two generators were to be run together, they would have to be kept synchronized. That could not be done without manned attendance on the platform.

I conclude that the electrical power generation capacity on the platform was sufficiently constrained for the power consumption of the fuel gas compressor to be a material consideration in considering the tenders for the fuel gas compressor.

In their particulars of claim, Phillips pleaded that by reason of the vibration problem, the fuel gas compressor was ‘perforce’ abandoned. Phillips’s pleading was amended. By re-amendment, it was pleaded as follows. The claimant decided to instal a fuel gas heater because it was unlikely that works to remedy vibration problems or other remedial works for the fuel gas compressor system could be completed before the end of October 1996, if at all. The fuel gas compressor was eventually abandoned and the heater used in its place. The claimant did not see fit to incur further cost and effort in attempting to resolve the vibration problems. In that the compressor would never be used for the purpose for which it was designed it was perforce abandoned.

I do not accept that it was perforce abandoned. It was abandoned because in the event, unforeseen by either party at the material time, it was never needed.

Mr. McMaster submitted that Phillips were entitled to recover from Snamprogetti, as wasted expenditure, the expenditure they had incurred on the fuel gas compressor. He based his argument on the following contentions:

In that the fuel gas compressor was never used, the expenditure on it was indeed wasted. But it is in issue whether the fuel gas heater was installed because of problems with the fuel gas compressor. Mr. King gave the following evidence about that in his first witness statement in a section beginning at paragraph 15. He first became involved with the work on platform 52/5A in early January 1996 on his appointment as Manager, Southern Area Projects. Given the need for simplicity and reliability, he was surprised that it had been felt necessary to use a fuel gas compressor, because the Zechstein wells appeared to provide the potential to supply fuel gas at an adequate pressure for the foreseeable future. I comment that even in January 1996 the possibility (to put it at its lowest) that the fuel gas compressor might turn out to be unnecessary was in the mind of Phillips. Its views on the subject may have been developing since 1994 (see the extract from the minutes of a meeting held on 23^rd August 1994 quoted at paragraph 416 above). The vibration was first noted on 11^th February (see paragraph 76 above). Mr. King went on to say that he requested Mr. Mobbs to commission a vibration survey by Acoustic Technology Limited to establish in more detail the magnitude and extent of the vibrations affecting the structure and equipment. Mr. Mobbs sent him a memorandum dated 28^th February 1996 stating that the vibration was not severe. Mr. King was aware that others disagreed, and his own subsequent experience on the platform showed that Mr. Mobbs’s assessment was wrong. By mid-March, no consideration had been given by himself or by the Phillips team to the elimination of the fuel gas compressor as no reason had arisen to doubt its effectiveness from a process standpoint. He regarded the vibration problem as very serious and knew that it had to be resolved before the system could be handed over to Operations. But he felt confident that subject to resolving the vibration problems commissioning could begin in earnest. Mr. King went on to mention the problems of starting the turbine and tripping of the fuel gas compressor. On 18^th April 1996 he was expecting a planned attempt to be made to run the main compressor to export sour gas to Bacton for the first time. Then he heard of the release of sour gas from the platform vent that had occurred on that day. He decided that commissioning should be suspended pending further investigation. At that stage, he felt that there was a need to understand what had happened. He recommended that a design audit be carried out. He wrote to Snamprogetti’s managing director on 23^rd April. The Hewett co-venturers were notified in writing on 24^th April by letter drafted by Mr. King.

Mr. King said that the factors he had considered in deciding to suspend commissioning were

Mr. King’s evidence continued as follows. Over the next few days, he discussed with Mr. Alan Wells the best way to approach the design review. It was following Mr. Wells’s suggestion that the Baker Jardine Hazop was set up. Some time before 3^rd May 1996, Mr King had discussions regarding the fuel gas system with Mr. Wells and with Mr. R. M. Gray. Mr. Gray was the senior process engineer within the process engineering department of Phillips at Bartlesville, Oklahoma. Mr. King became concerned that the fuel gas system was too complex and that it would not be possible to solve the vibration problems in the time available. Mr. Gray attended a meeting held on 3^rd May between representatives of Phillips (not including Mr. King) and of Snamprogetti. At that meeting, Mr. Wells, at the request of Mr. King, raised the issue of an alternative to the fuel gas system, eliminating the need for a fuel gas compressor.

Mr. Gray sent an e-mail dated 7^th May 1996 to Mr. King. The material part of that e-mail reads as follows:

The minutes of the meeting of 3^rd May contain the following entries:

Mr. Gray evidently visited the platform on 26^th May. He sent an e-mail to Mr. King dated 29^th May. The following is an extract from that e-mail:

On 25^th July 1996 Phillips, by letter enclosing written reasons for their request, sought from the co-venturers approval for further expenditure. The document stated that approval by the Hewett Unit Owners of the funding requested was urgently required to allow recommencement and completion of essential construction and commissioning activities to ensure that the gas turbine/compressor was proven available for production by end October 1996 latest. Under the heading “Fuel gas system” it was stated:

Miss Boswell submitted that at the end of March and up to the middle of April 1996 Phillips did not perceive that the vibration was such a serious problem that it could not be resolved by normal measures in the summer of 1996, if anything needed to be done. She relied on the following evidence of Mr. King and Mr. Halliwell. Mr. King gave this evidence in cross-examination (Day 22, p.79):

Mr Halliwell gave this evidence in cross-examination (Day 45, p.109)

I accept that the position in the middle of April 1996 (or specifically, 15^th April) was that Phillips, having received the second ATL report dated 19^th March, did not believe that vibration was such a serious problem that it required the abandonment of the fuel gas compressor, either for 1996 or for any time. Miss Boswell submitted that nothing happened in relation to vibration after that date which altered that position.

Miss Boswell submitted that the fuel gas compressor was abandoned because of the problem over electrical power generation on the platform. She said (Day 90, pp.69 to 72):

Mr. Constable was not a witness in this case. I am satisfied on the totality of the evidence that the decision to instal a fuel gas heater was not made in order to save electrical power. The position as of mid-April 1996 was that the outstanding vibration problems, which I have considered in more detail above under the heading Vibration, could have been solved and the condition rectified well before October 1996. The escape of gas that occurred on 18^th April 1996 was not itself of great importance, but it triggered Mr. King’s decisions to suspend commissioning and to have a design audit. The latter decision led to the setting up of the Baker Jardine HAZOP. Mr. Gray, possibly with input from Mr. King, expressed the opinion on 7^th May that there was a risk that the vibration problems would prove so difficult to rectify that a major fix such as purchasing another fuel gas compressor might be required. If Mr. King shared that view, he had changed his mind since the middle of April for no clear reason that appears on the evidence. Mr. Gray expressed a similar view on 26^th May, viz. that there was a risk that the fuel gas compressor would not be usable because of excessive vibrations. Thus time to rectify the vibration problem had been lost between 18^th April and the end of May, by which time the view was taken that the fuel gas heater should be progressed.

Miss Boswell has not made out her submission that Phillips acted in bad faith. The loss of time that occurred between 18^th April and the end of May was the result of a number of steps taken by Phillips, none of which in my judgment can be said to be unreasonable. Indeed the decision to instal the fuel gas heater, though by no means inevitable, was in my judgment reasonable. But the principal cause of that decision was the vibration problem, which I have found not to have involved any breach of duty on the part of Snamprogetti. In the absence of the vibration problem, any decision to abandon the fuel gas compressor because of the other problems would have been unjustified.

I am satisfied that when the fuel gas heater was installed the expectation of Phillips was that pressures would fall to a level requiring the future use of the fuel gas compressor. In the event, the fuel gas compressor was not needed because the pressures held up and the fuel gas heater had been installed.

Thus the expenditure on the fuel gas compressor was wasted because of the combination of circumstances that the Zechstein gas pressures held up and the fuel gas heater had been installed. It is true that the latter circumstance was triggered by the escape of gas that took place on 18^th April, but in the absence of a breach of duty in relation to the vibrations the consequences of that circumstance could not be laid at the door of Snamprogetti.

Thus I accept paragraphs (b), (c) and (d) of Mr. McMaster’s submission mentioned in paragraph 424 above. But the claim for wasted expenditure fails because the waste was not caused by any relevant breach of duty on the part of Snamprogetti.

Phillips claim the difference between the cost of retro-fitting the fuel gas heater and the cost that would have been incurred if it had been fitted in the first place.

The basis of this claim must be either that a fuel gas heater should have been specified by Snamprogetti originally, or that it was installed by reason of a breach of duty on the part of Snamprogetti. I have already given my reasons for finding that Snamprogetti were not in breach of duty in specifying and recommending the purchase of the single-cylinder reciprocating compressor that was installed on the platform. They were not in breach of duty in not specifying a fuel gas heater. I have also given my reasons for finding that there was no breach of duty on the part of Snamprogetti that gave rise to a necessity to instal the fuel gas heater.

It follows that this claim fails.

Moreover, the decision to instal the fuel gas heater before solving the vibration problems was made in order to have the main compression facilities running by the end of October 1996. That was done in order to avoid loss of production and loss of revenue. It constitutes an action taken to mitigate consequential loss. Damages to reflect the cost of such an action are not recoverable (see paragraph 15 above).

Snamprogetti’s design provided for blowdown of the main compressor on all trips. The Baker Jardine HAZOP response sheet (action 3.1.4) stated that that caused unnecessary loss of gas, environmental damage and increased recovery time from a minor trip. The recommendation was

On 2^nd March 1994 there was a meeting between representatives of Phillips and Snamprogetti. The minutes of that meeting record that it was Phillips’s practice to blow down compressors on vibration and other trips. At the Snamprogetti design HAZOP held in September 1994 and attended by representatives of both parties a proposal was made to provide automatic blowdown after a cooldown period following shutdown of the main compressor. The reason that is recorded on the relevant Hazop Action/Response sheet was

Mr. McMaster criticized that change as a departure from the standards of design that Snamprogetti had promised to bring to its work. That was because blowing down the compressor on all shutdowns involved loss of gas, introduction of sour gas to the atmosphere and an increase in the time it took to start the compressor again after a minor trip, hence potential loss of production.

In his witness statements, Mr. Tomlinson said this:

Mr. McMaster cross-examined Mr. Tomlinson on this topic at some length (Day 41, pp.72 to 87). I quote from the transcript of that cross-examination as follows (pp.76 to 80). Mr. McMaster’s first question relates to the Baker Jardine HAZOP and to Mr. Tomlinson’s evidence that there had been a change to the philosophy of Phillips:

I am satisfied that this change represented a change of mind on the part of Phillips. Snamprogetti are not to blame for the original design and are not liable to pay Phillips for the change.

Phillips expended money in carrying out a survey of the sea bed for the purposes of the Seafox operation. I am satisfied that it was reasonable to carry out that survey. Phillips also spent money in stabilizing the sea bed. It is unnecessary for me to decide whether that stabilization was necessary for the Seafox operation. Phillips have classified all that expenditure under a single heading. In expressing, under the heading Quantum below, the finding that expenditure under this heading was reasonable, I am not making any finding as to the reasonableness of the amount incurred.

I begin consideration of the quantum of damages with the following propositions:

There were agency staff engaged by Phillips for the purpose of working on the 52/5A compression project. Such proportion of their pay as was referable to the remedial works resulting from breaches of duty on the part of Snamprogetti is to be reflected in the damages, since that would not otherwise have been paid to them. But Mr. McMaster submitted that Phillips were entitled to recover a proportionate part of the pay of Phillips’s permanent employees who were engaged on the relevant work. He argued that they would have been working productively on something else if they had not been working on this project. It would be wholly unrealistic to assume that no additional managerial time was expended. There were accurate records to show what that management time was. He invited me to take the common sense approach of the well-known case of Tate & Lyle Food and Distribution Ltd. v. GLC [1982] 1 WLR 149. The first part of Mr. McMaster’s argument raises the question of opportunity cost, which I have already answered. As to the second part, extra managerial time would not sound in damages unless it were associated with extra managerial cost. Moreover, an accountant might reasonably apportion a cost by reference to management time, but that does not mean that extra management time was expended. The argument confuses costs and damages. In the Tate & Lyle case, (Hugh) Forbes J. said (ib., p.152):

Thus Forbes J’s dictum was strictly obiter. And it is to be borne in mind that the case originated as an Admiralty case, where there was a practice of allowing Agency as a single sum without strict proof: see p.151 of the report. In the circumstances, with great respect to Forbes J., I do not feel impelled to follow his dictum.

I have already (paragraph 15 above) indicated that the costs of Seafox are not recoverable as damages because they were incurred specifically to avoid loss of production and loss of revenue. But there is a second reason, namely that the costs of Seafox were not incurred, and could not justifiably have been incurred, in consequence of the breaches of duty that I have found to have existed. They were incurred principally because of the vibration and the desire to have a fuel gas heater installed and commissioned by the end of October 1996.

The two expert witnesses whom I heard on the quantum of damages were Mr. Anthony Farrow, for Phillips, and Mr. Leslie Charters, for Snamprogetti. They both gave their evidence by reference to costings of the remedial works. The expression remedial works was used by Mr. McMaster, and I adopt it here, as a convenient expression to identify the works carried out to implement the recommendations made at the Baker Jardine HAZOP. Those costings were prepared by Mr. Morrell and Mr. Abernethy who prepared a data base from the vouchers. Various revisions were made. The version used as the basis of Phillips’s claim was called Rev6last.

In his first report, dated 31^st January 2002, Mr. Charters wrote in paragraph 3.17:

In order to calculate the additional cost, Mr. Charters worked on the basis that retro fitting was 2.5 times more expensive than onshore fitting, and in consequence he deducted 40 per cent. of the retro fitting cost in order to calculate the additional cost. For reasons that he explained in sections 4.36, 4.37, 4.43, 4.45 and 4.46 of his report, he assessed the appropriate credit as £31,358.

Mr. Farrow, in paragraph 1.2.7 of his supplementary report dated 22^nd March 2002, mentioned the retro fit credits noted by Mr. Charters and said:

On 13^th September 2002, Mr. Charters produced a second supplemental report in the light of evidence that had been given. In that report, he expressed the view in relation to a number of items of claim that Phillips had suffered no loss.

On 18^th September 2002, Mr. Sylvester-Evans introduced an addition by way of an amendment to his report. He had been asked to consider the potential consequences of Phillips’s allegations on the basis that they were correct. He stated that he had considered what logically was required to remedy the alleged errors in the light of his knowledge and understanding of the work required and how that work could have been and was carried out. He proceeded on the assumption that the actual cost of works and hardware and so on was the same in 1995 and 1996. That assumption has not been in issue. In paragraph 4.18.2 Mr. Sylvester-Evans wrote this:

Dr. Robinson was taken in cross-examination (Day 65, pp.142 to 165) through Mr. Sylvester-Evans’s report in relation to all of those items. He was in total agreement with Mr. Sylvester-Evans’s evidence on most of the items and was in substantial agreement with his evidence on the remainder.

I am satisfied that Mr. Sylvester-Evans’s approach is the proper basis on which to assess the loss suffered by Phillips. Mr. Farrow gave no evidence in accordance with that approach. He stuck to the retrofit credit method throughout. For example, in his first supplemental report dated 22^nd March 2002 he exhibited a table which showed a figure of £14099.99 as his allocation to HAZOP node 2.1.1 (Zechstein metering skid blowdown) of the costs of KYE labour. (His retro-fit credit on that item was £1192, which he later increased to £3,919). He still relied on the figures in that table when he gave evidence in December 2002, although Mr. Sylvester-Evans had introduced his categorization on 18^th September 2002 and Dr. Robinson had been cross-examined on it on 31^st October 2002. Mr. Sylvester-Evans had written in relation to this item in his report of 18^th September 2002, paragraph 4.18.29:

Dr. Robinson agreed with that evidence (Day 65, p.162):

Mr. Farrow did modify his figures, producing in his third supplemental report of 11^th September 2002 a figure of £17,022 of the cost of KYE labour and material for this item. (That figure was further reduced in Mr. McMaster’s final submissions to £14,039 “in the light of certain aspects of Mr. Charters’s and Mr. Farrow’s evidence which Phillips accepts slightly change what it is able to recover”). But Mr. Farrow retained his retro-fit credit method of assessing the loss. And his retro-fit credit did not include overheads. He expressed his reasons in paragraph 1.2.6 of his first supplemental report as follows:

In his third supplemental report, paragraph 2.3, Mr. Farrow said this:

Mr. Farrow adhered to his view when cross-examined about it. The following is an extract from his cross-examination (Day 81, pp.17 to 19):

The burden of Mr. Farrow’s argument, to put it shortly, was that the marginal cost that would have been incurred if Snamprogetti’s contract had been performed is represented by the direct labour hours and does not carry any overheads. But by the same token, the marginal cost of carrying out those items of the remedial works that I have found to be due to breach of duty on the part of Snamprogetti would carry little or no overheads, since those works, too, are a small part of the whole remedial works. Mr. Farrow seems to have had in mind that Snamprogetti were liable for the whole, or substantially the whole, of the remedial works. At any rate on the facts as I have found them, his argument is seriously flawed. He himself expressed the view in the passage quoted above that it was a question of the extent of the works.

Mr. McMaster made the following submissions in relation to category 2(C) items:

Mr. McMaster’s argument is attractive, but I cannot accept it in its entirety. I have dealt with points (a) and (h). The factual basis for saying that (b), (c), (f) and (g) occur for the second time is not clear to me. But if they do occur for the second time, they must have been required for other purposes by AMEC and may well have been required for other purposes in relation to the rest of the remedial works. The factual basis of (e) is not clear to me. Point (d) may be correct. It is not clear what AMEC costs are said to have been wasted or how they are additional to the other points. Nevertheless, I accept that there may have been some loss.

In his final submissions Mr. McMaster claimed damages in the sum of £17,648 in respect of the Zechstein metering skid, having taken into account the retro-fit credit of £3,919. (In addition to the individual HAZOP items, he submitted an overall claim, not allocated to individual HAZOP items, for fixed costs in the sum of £1,478,109). He submitted that it was impossible to provide a precise figure for each HAZOP node. I accept that if so, that is a good reason for attempting to estimate the loss approximately on the evidence available. But in the case of the Zechstein metering skid, I am satisfied that the loss, if any, can only be a small part of the sum Mr. McMaster claimed.

There are other instances of category 2(C) items where Dr. Robinson was in entire agreement with Mr. Sylvester-Evans. It is clear that Phillips have suffered little or no loss in relation to any of those items. Yet in reliance on the evidence of Mr. Farrow Mr. McMaster submitted that considerable sums were due to Phillips from Snamprogetti for those items. They are the following:

Mr. Morell and Mr. Abernethy, who had checked KYE’s timesheets and job cards for the purpose of allocating the KYE work to the various HAZOP nodes, were unable to identify the work in relation to 5451.50 hours. Mr. Farrow spread those hours, which he called “All HAZOP” hours, across all the HAZOP nodes pro rata to the number of hours found to have been spent on each HAZOP node. Thus hours for unidentified work were included in the claim as formulated by Mr. Farrow. Since Snamprogetti’s liability for each of the HAZOP items was in issue, and also because not all of the all-HAZOP hours were necessarily spent on any of the HAZOP nodes, I find that to be an unhelpful method of costing the work for the purpose of assessing damages. The effect of adding in the all HAZOP hours is, on its own, sufficient to add more than 50 per cent. to the number of hours attributed to each HAZOP node.

In the light of Mr. Sylvester-Evans’s amendment of 18^th September 2002, Mr. Charters amended his second supplemental report on 5^th December 2002. By that amendment, he gave his opinion as to the loss suffered by Phillips in the light of Mr. Sylvester-Evans’s categorization and description of the works. And in his second supplemental report, Mr. Charters was able to allocate some of the all-HAZOP hours to HAZOP nodes; the rest he excluded from his costings of the remedial work. I use Mr. Charters’s second supplemental report, as amended, as the starting-point for my assessment of the damages.

Before considering the assessment of damages item by item, I shall assess the costs properly chargeable to the items of remedial work. I shall then apply Mr. Sylvester-Evans’s criteria in order to arrive at an assessment of the relevant loss suffered by Phillips. The distinction between the former and latter processes is not clear-cut, as will appear.

In the assessment of costs, I must consider certain points of general application which it is claimed should be apportioned to the various items of remedial work. I consider first management costs.

Phillips claim £185,143.56 by way of management costs, said to have been incurred in relation to the remedial works. That figure is broken down as follows:

Mr. Charters considered that nine persons who worked at Woking would have been employed there in any event and did not represent an additional cost. I accept the evidence of Mr. King, however, that eight of them (the exception being Mr. Viganego) were agency personnel. Mr. King said that Phillips would not have required those agency personnel but for the need to carry out the remedial work. They were an additional cost which arose from the need to carry out the remedial work. Whilst Mr. King’s evidence on the point is not particularly clear, Mr. King may have been intending to convey the information that all the agency fees claimed were incurred in relation to remedial work. Mr. Charters, however, considered that that was not so. He had checked, or spot-checked, the time sheets. He reached two conclusions, expressed in his supplemental report dated 22^nd March 2002, paragraphs 2.18 and 2.19. The first conclusion was that work to the value of £33,170.70 had been carried out after 10^th October 1996 through to May 1997. He considered that apart from the cost of finalizing paperwork, that work could not be linked to any default on the part of Snamprogetti. He had in mind, of course, all the alleged defaults. I reject that conclusion. Mr. Charters’s second conclusion was that the claim for management expenses was a global claim in the sense that it related to all the work done on the platform at the material time, not simply the remedial work. I accept Mr. Charters’s second conclusion. In his second supplemental report, dated 13^th September 2002, Mr. Charters expressed the view at paragraph 2.165 that it was reasonable to assume that time related charges such as Phillips’s management were directly related to the labour resources employed. By way of amendment dated 5^th December 2002 to paragraph 2.165 of his second supplemental report, he resiled from that view on the ground that Phillips had made no attempt to identify what management time was devoted to any specific breach of contract. In spite of that, I think it right to allow the variable management costs (i.e., those of the agents) pro rata to the KYE direct hours worked. On the evidence of Mr. King, I find that all the expenses mentioned above against the word “Offshore” were agency fees.

I conclude that the proper sum to allow for management charges is £153,368.

There is a claim in respect of Bacton contractors which Mr. Farrow eventually assessed at £136,504. Bacton contractors were said to be contractors employed by Phillips at its onshore base at Bacton.

Mr. Charters made the point that of the sum claimed, £30,232 represented payments to Phillips’s employees. That was not denied. There is no suggestion that it would not have been paid to them in any event, even though they may have been occupied on the remedial works. Thus this part of the claim fails.

There is a claim, originally in the sum of £104,958, in relation to Grootcon. Grootcon were contractors used by Phillips at Bacton for maintenance services. Grootcon supplied, in shifts, for work on the 52/5A platform, two crane drivers, an electrical technician and an instrumentation technician. The crane drivers also acted as labourers and watchmen. The electrical technician was present on the platform to act as a responsible person to issue permits for work, to deal with isolation work and any interface with the platform equipment. The instrumentation technician did similar work in relation to work carried out on the instruments. On the evidence of Mr. Abernethy, all of them were required for the remedial works. Since much of the expenditure was incurred after 10^th October 1996, after the fuel gas heater was operating, Mr. Charters considered that that expenditure was probably for work for Bacton or was a continuation of the power generation upgrade. He also considered that some of it was for supervision not related to the remedial work, since KYE had their own supervision. On 22^nd March 2002, in his first supplemental report, Mr. Farrow stated that Phillips had now located more invoicing records, and in consequence he reduced his valuation of the amount of the Grootcon claim to £88,978. I am satisfied on the evidence contained in Mr. Farrow’s third supplemental report, dated 11^th September 2002, that the claim is well documented. Mr. Charters’s supposition is not borne out. I assess this claim in the amount of £88,978.

£2443 of the claim represented payments to a company called Sutton Electronics. On the evidence of Mr. Abernethy, I am satisfied that those payments were to communications operatives for connecting and interfacing the Seafox telephones and computers to the 52/5A platform. That part of the claim falls with the claim for the expenses of Seafox.

A further £12,012 was claimed as payments to Antech Engineering. Mr. Abernethy said this in paragraph 57 of his witness statement:

Restbury was a vendor of equipment used in the remedial works. Mr. Charters said that Antech was partly for Seafox (£1,612) and partly for commissioning (£8,524). In his final submissions Mr. McMaster said that in so far as that was correct, then it should not be excluded from the claim but would only be recoverable (in full) in the event that Seafox costs were recoverable. Mr. McMaster evidently had no instructions whether Mr. Charters was correct about Antech. I am satisfied that he was. The commissioning was not complete when the Baker Jardine HAZOP was held. Mr. Abernethy’s evidence does not go so far as to say that any extra money was paid to Antech on account of the remedial work. The totality of the evidence on this point is insufficient to establish whether any of the money, other than £1,612, was spent on the remedial works. The claim for £1,612 falls with the Seafox claim, and I reject the whole claim in respect of Antech.

There is a claim for £478.40 in relation to Oilfield Medical Supplies. It is clear from the cross-examination of Mr. Charters (Day 87, p.106) that this claim falls with the Seafox claim.

There is a claim for £2,360.67 for OTS (Oilfield Testing Services). After much initial confusion, it emerged that this related to non-destructive testing of welding. Mr. Abernethy, in his second witness statement, confirmed that the work was essential for the project. Mr. Farrow was able to say in his supplemental report that Phillips had now identified an invoice for Messrs. OTS in the sum of £2,360.67. It thus appears that this money was paid to an independent contractor for work relating to the remedial works. It is thus properly included in the claim under this head.

I thus assess a total of £91,339 under the heading of Bacton contractors: £88,978 for Grootcon and £2,361 for OTS.

Mr. Charters said in his amended second supplemental report (paragraph 2.59) that the witness evidence and a further consideration of the global nature of the claims had enabled him to allocate the claimed all-HAZOPS hours with reasonable accuracy. I consider the individual items below.

I accept a conclusion of Mr. Charters, based on the evidence, that setting up the job was required for the whole of the work on the platform. That included the power generation upgrade, Bacton and AMEC. (“Bacton” refers to non-remedial work carried out by KYE for Phillips. AMEC refers to outstanding work not completed by AMEC when work was stopped by Mr. King). I accept that the relevant number of hours was 639. Mr. Charters allocated that number of hours to the remedial work in proportion to the numbers of offshore hours allocated to the remedial work and the non-claimed work. The numbers of those hours were respectively 30,158 and 19,542, totalling 49,700 hours. The respective proportions are 61 per cent. and 39 per cent. Before making his amendments to the second supplemental report, Mr. Charters had arrived at figures of 12,604 and 20,688 hours for the remedial work and non-claimed work respectively. Those figures lead to the respective percentages of 62 and 38, which Mr. Charters did not amend. Since this is not an exact science, the discrepancy is unimportant. He arrived at figures of 396 and 243 hours for the allocation of temporary supplies between claimed work and non-claimed work respectively.

Mr. McMaster submitted that there was no evidential basis for deducting 243 hours from the direct hours allocated by Phillips as all-HAZOP temporary supplies, apparently simply on the basis that there was power generation upgrade work on the platform. That argument presupposes that Phillips’s allocation was correct in the first place. I accept Mr. Charters’s conclusion on this point.

Mr. Charters explained that ‘Operations’ was a department of Phillips that was working on the platform dealing with the power generation upgrade (among other things) and whenever the platform was producing gas. It is clear on the evidence of Mr. Abernethy (Day 26, p.158) and of Mr. Morrell (Day 28, p.122) that the category of Assist Operations could include power generation upgrade and other work falling outside the remedial work. 1486 hours out of 1638 hours claimed against Snamprogetti under this heading related to the period October to December 1996. Of those 1486 hours, 565 hours were direct hours included in all-HAZOPS, and 921 hours were indirect hours. Mr. Charters allocated those 1486 hours to non-claimed work, removing them from the all-HAZOPS and indirect categories, stating that those incurred in the period from late October to December 1996 were “clearly not connected with the HAZOP node remedial work”. Mr. McMaster submitted that there was no reason to re-allocate any of those hours. Mr. Charters may have gone too far in saying that the hours were clearly not connected with the HAZOP node remedial work. But it is by no means clear that they were connected with that work, and in my judgment he rightly removed them into the non-claimed category. Phillips’s case that those hours ought to be charged to the remedial work, let alone that they ought to be charged to Snamprogetti, is not made out.

Mr. Charters said that 462 direct hours for commissioning taking place between 30^th September and 20^th October 1996 were included in the all-HAZOPS claim. He considered that the cost of commissioning work outstanding when work was suspended in April 1996 should be removed from the claim. That principle is not in issue, but the amount is in issue. On the basis of evidence contained in Mr. Morrell’s second witness statement, dated 14^th March 2002, Mr. Charters assumed that operations and vendor assistance would be required for two men for seven days (196 hours). He accordingly transferred 196 direct offshore hours out of all-HAZOPS hours.

Mr. McMaster quoted Mr. Morrell’s evidence. He submitted that the effect of it was that there would only be two vendor representatives who would stay on the platform for around three of the seven days, subject to how the run went. Thus the re-allocation of hours should not be made.

Mr. Morrell’s evidence on the point is this. I quote the relevant paragraph (paragraph 13) in full. Mr. Morrell was referring to Mr. Charters’s first report, dated 31^st January 2002, paragraph 4.47.6:

It is clear from that evidence that the 148 hours did not relate to commissioning. It is also clear that the operations team was required for the final seven-day run, which was part of the commissioning. Mr. Charters’s assumption, in my judgment, is well justified by the evidence.

Of the remaining 266 hours in all-HAZOPS, Mr. Charters re-allocated 200 arbitrarily. I do not accept that that is justified.

Mr. Charters gave the composition of the electrical all-HAZOPS claim. The total number of hours was 514. He transferred 77 hours to all-HAZOPS temporary supplies. That figure is included in the figure of 639 hours mentioned in paragraph 490 above. There was no dispute about that transfer. There was no dispute about a further 77 hours, relating to general earthing. The relevant job card identified the specific instruments in question. He re-allocated those hours from all-HAZOPS to the relevant baker Jardine HAZOP nodes as follows:

Mr. McMaster submitted that the assumption that the earth bosses were for the future general use of the platform was incorrect. He drew my attention to some drawings which showed that all the relevant bosses were in the areas where the remedial works were carried out. He added that earth bosses were not in normal practice fitted for future general use of any facility but were items for a discrete purpose. As to the last proposition, I am not aware that it was in evidence. But if it is true, Phillips could have adduced evidence showing the purposes of the earth bosses. In the absence of such evidence, I am not prepared to draw an inference, from their location, that they related to the remedial works. Even if Mr. Charters’s assumption was wrong, there remains his opinion that the installation of the bosses did not seem to be linked to the HAZOP work, from which he made his assumption. Phillips have failed to prove their case on this point. In my judgment, Mr. Charters was right to remove the 31 hours in question from the all-HAZOPS category.

Mr. Charters transferred 15 hours for testing all opened junction boxes to unclaimed work. His reason was that the work did not seem to be linked to the Baker Jardine HAZOP report, and presumably would have been necessary in any event due to the power generation upgrade.

Mr. McMaster submitted that Mr. Charters was wrong to assume that the testing of the opened junction boxes was necessitated by the power generation upgrade. He drew my attention to a wiring diagram index sheet that had been prepared by Snamprogetti for the compression project which listed the destinations of the cables running from the various relevant junction boxes. Those destinations were HIPS, emergency shutdown and marshalling cabinet. Mr. McMaster submitted that the work had obviously nothing to do with power generation upgrade and everything to do with the HAZOP works.

The wiring diagram index sheet had not been put to any witness. According to the relevant test pack (number 2276-45-003), the work of testing the junction boxes that had been opened appears to have been completed by 28^th September 1996. That, I think, is some slight evidence supporting Mr. McMaster’s submission. On the other hand, the source or sources of the electrical power coming into the junction boxes was not stated. I reject the submission that the work had obviously nothing to do with the power generation upgrade. It may well have had something to do with it. It may well have been part of the remedial work. This is a matter on which Phillips could, I assume, have adduced clear evidence. They have failed to do so. In my judgment, Mr. Charters was right to remove the 15 hours in question from the all-HAZOPS category and place it into Bacton.

The largest item under this heading related to the erection and striking of scaffolding. The hours charged were respectively 143 and 171½. Mr. Charters said that there was no link to the HAZOP work on the job card. It was possible that it was for the power generation upgrade because it seemed to him that that work would require scaffolding but Grootcon (who carried it out) did not have any scaffolding facilities. In view of the doubt, he said, he transferred the 313½ [sic] direct hours from all-HAZOPS to non-claimed work.

Mr. McMaster submitted that reference to the electrical job card index for the relevant workpack showed that the scaffolding was not power generation upgrade work. All that Mr. McMaster succeeded showing in fact was that a few of the many other items mentioned in the workpack related to matters other than the power generation upgrade.

In my judgment, the purpose of this scaffolding is not at all clear. Mr. Charters was right to exclude it from the claimed work.

This matter relates to two job cards. One of them, representing 31 hours, was for connections to an instrument workshop on Seafox main deck. Mr. Charters transferred that to All HAZOPS: temporary supplies. It forms part of the 639 hours mentioned above. Mr. Charters said that it had no connection to the claimed nodes.

Mr. McMaster submitted that it was not correct that the connections to the instrument workshop had nothing to do with the claimed nodes, and that Mr. Charters lacked the expertise to make such a comment. He referred me to the purchase order for hire of the workshop, which showed that the delivery point was the Seafox. He said that the point had not been put to the relevant witnesses.

The only witness who mentioned this job card was Mr. Charters. The matter was put to him by Mr. McMaster. I quote one question and answer from the relevant cross-examination (Day 87, p.32):

The effect of transferring the hours to temporary supplies is that they are spread over claimed work and non-claimed work. I am satisfied that Mr. Charters was right to do that.

The other job card, for 24 hours’ work, was for the installation of a temporary electrical power supply to a flushing rig. Mr. Charters said that the flushing rig was for the fuel gas heater and the fuel gas compressor. It did not properly fall within all-HAZOPS. He accordingly transferred 12 hours to the fuel gas heater and 12 hours to the fuel gas compressor. He was not cross-examined about that. I am satisfied that his allocation was reasonable.

Isolation work was carried out as preliminary work from 18^th to 31^st July 1996. 441 direct hours were claimed. Mr. Charters, having researched the files, concluded that it seemed evident that isolation of contactors 1 to 5 was required primarily to instal the fuel gas heater. (There was a typographical error in paragraph 2.80 of his amended report, but from the context I infer that it was intended to refer to contactors 1 to 5). There was also isolation for vessel inspection work. Mr. Charters allocated all the 441 direct hours to the fuel gas heater, although, he wrote, it was possible that isolation would have been required for Bacton [i.e. non-claimed] work.

De-isolation work was carried out from 11^th September to 7^th October 1996. The time on the job card was 818 direct hours. He added that it seemed evident that that time included time for re-connecting contactors 6, 7 and 8 that had previously been isolated for reasons not connected to the HAZOP nodes. In the absence of further information he allocated the 818 hours in proportion to the numbers of contactors involved. He allocated three-eighths (306 direct hours) to Bacton work and five-eighths (512 direct hours) to the fuel gas heater.

Mr. McMaster submitted that Mr. Charters lacked the expertise necessary to make the comments that he did. Mr. Charters, he said, appeared to be suggesting that over 1,200 hours (441 direct hours for isolations and 818 direct hours for de-isolation) should be deducted. All deductions he sought to make should be ignored. The points did not appear to have been put to the relevant witnesses.

In fact, Mr. Charters was deducting 306 hours, not 1,259 hours. Mr. Charters was scarcely cross-examined about platform process isolations. He wrote about them in paragraphs 2.78 to 2.81 of his second supplemental report. He wrote about instrumentation (which I have considered above) in paragraph 2.77. The cross-examination in question was as follows (Day 87, p.29):

Mr. Charters’s evidence on this point is the best available. I accept that he has made a reasonably fair allocation of these hours.

Mechanical all-HAZOPS is related to two job cards. One of them is for additional scaffolding work at 96 hours. Mr. Charters wrote that he had been unable to locate that job card and that there was no evidence to identify the work or to link it to any HAZOP. He transferred the hours to non-claimed work. Mr. McMaster did not draw my attention to any other evidence about the job card or the work. But he submitted that Mr. Charters was acting as a judge assessing the evidence rather than as an expert witness. I interpret Mr. Charters’s evidence as meaning that he had been unable to find anything in the documents disclosed or in the witness statements or in the transcripts of the evidence that enabled him to identify the work. It is true that he exercised his judgment as an expert witness in transferring the 96 hours out of all-HAZOPS. But in so doing he was merely stating his opinion, as he was entitled to do. In my judgment, he was right.

The other job card was for fire watch offshore: 1,693 hours. Mr. Charters allocated those hours to claimed and non-claimed work in the proportions (62 per cent. and 38 per cent. respectively) mentioned in paragraph 490 above. He accordingly allocated 1,050 hours to claimed work and 643 hours to non-claimed work (power generation upgrade in this case). Mr. McMaster submitted that none of the various tentative bases put forward to justify this re-allocation stood up to scrutiny.

The relevant part of the cross-examination of Mr. Charters (Day 86, pp.110 to 116) is as follows:

It appears that Mr. Charters may have been under a misapprehension as to the meaning of the reference to welding sockets. But there was other Bacton work that might have involved welding. Mr. Charters was also confused over his explanation of his apportionment. If only 10 per cent. of the KYE work required fire watch, then his apportionment represented an assumption that 10 per cent, and not 6 per cent., of the non-claimed work required fire watch. Nevertheless, I am satisfied that Mr. Charters well understood what he was doing.

I am satisfied that Phillips’s allocation of the whole of the fire watch hours to claimed work was unjustified. Mr. Charters’s re-allocation was reasonable, and I accept it.

Mr. Charters referred to three job cards relating to general scaffolding. They amounted to 1,373 direct hours. He considered that those job cards probably related to the dismantling of the existing scaffolding left by AMEC when the commissioning of the compression facilities was halted. Accordingly, he transferred those hours from all-HAZOPS to unclaimed work. Mr. Farrow agreed that they included dismantling of the scaffolding left by AMEC, but thought that they also included work modifying it. He said this in the course of his cross-examination (Day 83, p.6):

There was documentary evidence that one of the job cards in question related to scaffolding work of 47 days’ duration “Ex AMEC”. Mr. Farrow gave evidence that there were Grootcon scaffolders employed in AMEC carry-over works whose hours were not charged against Snamprogetti.

Mr. Morrell gave evidence about job card PH426, which was missing and which he had never seen. The following exchange took place during his cross-examination (Day 28, p.27):

Mr. Charters’s conclusion was that since Phillips had not been able to allocate the general scaffolding to any specific work it seemed reasonable that the claim should be adjusted.

It is clear that a substantial part of the relevant hours related to the dismantling of AMEC scaffolding, which by common consent is not chargeable to Snamprogetti. In my judgment, the amount of the balance, and whether it was used for the HAZOP items, is so uncertain that Mr. Charters was right to adjust the claim by removing the whole amount of 1,373 direct hours from the all-HAZOPS category.

Mr. Charters expressed his view that the figures adduced by Phillips as the direct labour costs of KYE were too great in that they made insufficient allowance for the travelling time of members of the labour force to and from the platform at each end of a work period. The mobilization allocation varied between 1 and 12 hours, and averaged 1.18 hours. Mr. Abernethy gave evidence that perhaps 2 or 2.5 hours per man were lost when the crew changed. I accept that evidence. Mr. Charters used a figure of 2.25 hours, and reached the conclusion that the time for mobilization plus demobilization had been under-estimated by 558 hours. I accept his calculation and conclusion. He then deducted a total of 588 direct hours (in error for 558) pro rata from the figures of direct hours for each item as adjusted by him. The effect of the error is negligible, and the indeed the effect of the whole adjustment is small.

These charges are properly treated, by common consent, as variable costs since they formed part of the charges amounting to £139,740 made to Phillips by KYE. Mr. Charters considered that so far as the claim against Snamprogetti was concerned there was a cost overrun of £32,885 which ought to be deducted from the figures under this head. His reason was that the remedial works were completed by 10^th October 1996, whereas the charges ran on to 15^th July 1997.

The works were not in fact completed by 10^th October, but they were substantially completed by 18^th October. All the work was completed by 2^nd December. The works outstanding on 18^th October were those dealing with HAZOP items 2.2.2, 2.2.5, 2.3.11, 2.5.1, 3.1.3, 3.2.3, 3.2.5, 4.1.3, AMEC, Bacton and Seafox work. By far the bulk was Bacton work. Outstanding works on items 2.2.2, 2.2.5, 2.3.11, 2.5.1, 3.1.3, were in each case a small proportion of the whole item.

In my judgment, it is reasonable for costing purposes to include the KYE Head Office charges up to 2^nd December 1996. On the basis of evidence given by Mr. Charters in paragraph 2.12 of his first supplemental report, I find that at the material time the charges of KYE for its head office expenses were about £3,200 a week. Thus I allow £26,000 for those charges. I conclude that the deduction should be reduced to £6,885, leaving a charge of £132,855.

The claim for KYE materials includes costs of purchase of materials, costs of hire of plant and certain other things. The method by which Phillips attributed KYE material costs to the individual HAZOP nodes was explained by Mr. Morrell in his first witness statement as follows:

Mr. Morrell expanded upon his explanation in his evidence in chief (Day 27, p.127 to p.154). Materials that were identified as relating to AMEC, Bacton or Seafox works were so classified. As to the rest of the materials, the process can be described as being in three stages. First, those materials that could be identified as applying to a particular HAZOP node were allocated to that node. Second, of the remainder, those that could be identified by the discipline within which they would have been used, e.g. electrical, mechanical, instruments, scaffolding, were apportioned in each case across those HAZOP nodes that had involved the relevant discipline in proportion to the numbers of hours that had been spent on the relevant discipline in those nodes. Finally, the balance of the material costs was apportioned to the HAZOP nodes proportionately to the direct hours allocated to those nodes. AMEC and Bacton materials were not included in the claim.

Mr. Charters criticized the way that Phillips had apportioned the costs of KYE materials across the HAZOP nodes. He pointed out that an apportionment of £6,534 to summer storage (HAZOP node 3.2.13) should have been reduced to £691 when a correction in the number of hours of direct labour attributable to that node was made. The error was a large one: 1,225 hours had been taken, whereas the corrected figure was 130 hours. An amendment to the particulars of claim which corrected the number of hours did not amend the material costs claimed for that node. Mr. Charters concluded that that showed that Phillips’s method of apportionment was unreliable. I reject that conclusion. The failure to make the appropriate amendment to the material costs claimed was inadvertent.

Mr. Charters apportioned time related costs, most bulk purchases and lorry and container hire arbitrarily. On the whole, I find that Phillips’s method of apportionment is preferable, though the figures at which Phillips arrived are varied by my decision on the apportionment of KYE direct hours between the nodes.

Mr. Charters carefully considered a major purchase of materials under purchase order 29620 in the sum of £38,986. The whole of that sum had been allocated to specific nodes in Rev6last, most of it to node 2.3.11. Mr. Charters divided the amount in accordance with the destined location of the materials as shown on the drawings and job cards. The main practical difference between his approach and Phillips’s is that he divided the amount attributed to node 2.3.11 between the fuel gas heater and the fuel gas compressor. Whilst I accept that division, since it is convenient, I am not satisfied that Mr. Charters’s allocation is otherwise an improvement on Phillips’s original. The differences between the two sets of figures are comparatively small. I should, if necessary, apportion all costs directly allocated to node 2.3.11 in Rev6last between that node and the fuel gas heater in proportion to the numbers of KYE direct offshore man-hours respectively attributable to those items. But since I find that the claim in respect of both of those items fails, it is unnecessary for me to do so.

Mr. Charters removed two items. One was for a Tangye pump in the sum of £3,573, on the ground that there were no supporting invoices. The other was supplies from GY Fire and How Fire, amounting to £4,562, since the relevant HAZOP node had not been pursued as a ground of claim against Snamprogetti. The latter deletion is common ground. As to the former, Mr. McMaster referred me to a purchase order number 29872 together with invoices adding up to slightly more than £3,573. Mr. Charters was not cross-examined about this matter, but I find that he must have been mistaken in removing this item.

The claim included a claim in the sum of £23,730 for the hire of scaffolding. Mr. Charters gave evidence about that in his second supplemental report (p.223, paragraphs 2.111 and 2.112) and in evidence in chief (Day 84, p.110). It is clear from schedule 7 to his report that he had looked at the vouchers in detail. He concluded, and I accept, that the total sum should have been £22,508. Of that sum, £6,875 was for insulation materials and not for scaffolding. I accept his evidence that the time allocated to insulation was 341 hours for the fuel gas heater and 190 hours for miscellaneous pipework, and that the cost of the insulation materials should be allocated in the same proportion. Accordingly, I allocate £4,415 to Bacton and £2,460 to the fuel gas heater.

Mr. Charters pointed out that of the hire charges for the scaffold, £6,475 represented charges for the period 1^st June to 26^th July, before the remedial work started. He re-allocated that sum to AMEC. He also arbitrarily re-allocated the remainder of the hire charges. Mr. McMaster accepted that half the hire charges should be disallowed. The hire charges were £15,633, of which I disallow one-half, say £7,800. I do not accept that the hire charges should be re-allocated. The result is that I allow £7833 out of the sum of £23,730.40, and thus deduct £15,897.40 from the claim.

The claim for KYE materials includes an amount of £14,000 for courses undertaken by KYE personnel. Mr. Charters said this about it (second supplemental report, p.223, paragraphs 2.113 and 2.114) and Day 84, p.111):

Even if Phillips were justified in paying this sum, which is not clear, it represents a fixed or capital cost which will not be reflected in the damages. I thus exclude it from the calculation of costs. However, part of the amount was allocated to Bacton in Rev6last. Only £12,517.40 was claimed.

This claim is for the cost of transporting personnel to and from the platform by helicopter based on Great Yarmouth.

It was, eventually, common ground that the cost per passenger-hour for transport by helicopter was £1,190. A higher figure was given by Mr. Farrow for sole use helicopters, however. Mr. Charters said, and I accept, that that higher figure was subsumed within the average of £1,190. Mr. Charters initially stated that the average number of passengers per flight established from the helicopter manifests was 9.5. Mr. Farrow took a sample of the helicopter manifests and worked out an average of five based on the sample (Day 83, p.81). On making another check for the entire month of September 1996 he arrived at an average of 5.7 (Day 83, p.85). In a set of calculations that he prepared overnight during his cross-examination, in answer to a question that had been put to him, Mr. Charters produced a table of dedicated flights from July to October 1996 that showed the average number of passengers per flight as 7.92. He produced that set of calculations at the end of his cross-examination, and was not questioned about them. Mr. McMaster submitted that the true average was 6.2 or 6.4.

The time taken by a single flight, outward or return, was 20 minutes. Using his average of 9.5 passengers per helicopter journey, Mr. Charters calculated the cost per passenger-journey one way as £41.78. Mr. Farrow agreed with that calculation save for the average number of passengers per journey (Day 83, p.89). The calculated cost of £1,190 an hour included the cost of a standing charge which was charged whether or not the helicopter was used. In the further calculation that he produced at the end of his cross-examination, Mr. Charters deducted that standing charge on the basis that it was the marginal cost of the journeys that was relevant. Applying an average of 8 passengers per helicopter journey, he arrived at a new cost per passenger journey of £34.49. Mr. McMaster accepted that the standing charge must be excluded from the calculation.

To resolve the difference of opinion as to the average number of passengers per helicopter journey, I adopted and expanded the method used by Mr. Farrow. Mr. Farrow said this (Day 83, p.85).

Mr. Charters calculated from study of the helicopter manifests that there were 1573 person-flights between June and December 1996. Of those, 755 were for personnel other than KYE personnel travelling between July and December 1996. (Mr. Charters gave a figure of 745 in the body of his second supplementary report, paragraph 2.143, but a figure has been mistranscribed from his schedule 9. The figure in schedule 9 appears to be a correct figure). That leaves 818 person-flights attributable to KYE. (In the absence of evidence I take all the June flights to be flights of KYE personnel). The relevant cost of helicopter journeys is 818 times £44. That equals £35,992. The cost must be closely related to man-hours worked on the platform. Thus I conclude that the cost falls to be apportioned to the claimed and the non-claimed work of KYE in proportion to the relevant direct offshore labour costs.

This claim is for the transport of materials to the platform for the remedial works. The charges are for space on vessels and for quayside assistance. I accept Mr. Farrow’s assessment of the relevant costs at £16,402, based on the invoices and on the evidence of Mr. Brighton. That sum should be apportioned in proportion to the apportionment of the material costs. I am not satisfied that Mr. Farrow’s figure (said to be attributable to AFE 3461, the authorization for expenditure in question) is such that the apportionment should be confined to the HAZOP nodes, as submitted by Mr. McMaster.

This part of the claim relates to sums paid for the attendance on the platform of representatives of vendors of items of equipment for overseeing the installation and commissioning of those items. The relevant vendors and the respective amounts claimed are as follows:

The Delaval Stork claim relates to HAZOP node 3.2.3. Delaval Stork carried out modifications to the anti-surge control system at their factory, and the claim is for the attendance of their engineers to re-commission it on the platform.

The EGT claim relates to HAZOP nodes 2.3.11 and 2.5.1. As to the former, the claim relates to work by EGT’s technicians in integrating the new fuel gas heater system into the gas turbine. As to the latter, EGT were also involved in the commissioning of the propane start-up system.

The Heatex claim also relates to HAZOP nodes 2.3.11 and 2.5.1. Heatex supplied the new fuel gas heater. The work appears to have related to modifications to the control panel and the provision of a new room thermostat for the propane heater cabinet.

The Hima-Sella claim relates to HAZOP node 3.1.3. The work involved modifications to the HIPS logic system.

The Restbury claim involves HAZOP nodes 1.1.6, 2.2.2, 2.2.4, 2.3.9, 2.3.10, 2.3.11, 3.1.3, 3.1.4 and 3.1.5. Restbury International were involved in modifications to the ESD and SCADA systems. The ESD is the equipment shut-down system. The SCADA system sends signals from the platform to the control room on the 48/29A platform so that the 52/5A platform can operate as not normally manned. The work carried out related to software modifications which their operatives carried out on the platforms.

The Foxboro claim is said in Mr. McMaster’s submissions to relate to HAZOP nodes 2.2.5, 2.3.11 and 3.2.3. It concerns a site visit to reconfigure controllers numbered 761 and 762 on the gas turbine. That was nothing to do with HAZOP node 2.2.5, which related to the fuel gas compressor suction pressure control valve.

The amounts of the above claims can be allocated to individual HAZOP nodes. Accordingly, I need not further concern myself with those claims that do not arise out of breaches of duty on the part of Snamprogetti that I have found to have occurred. On that basis, I am concerned only with Hima-Sella and Restbury.

In relation to Hima-Sella, I am satisfied that the figure of £13,818 is justified as a cost under HAZOP node 3.1.3.

As to Restbury, I am satisfied that the amount of £29,363 was incurred. Evidence before me of the details of the work done is sketchy. The only three HAZOP nodes with which I am here concerned are 2.3.10, 3.1.3 and 3.1.5. HAZOP node 2.3.10 relates to the defective or unsuitable recycle valve PV 923 in the fuel gas compressor system. That valve was not replaced. I am not satisfied that Restbury did any work in relation to the software concerning that valve or its potential replacement. Mr. McMaster in his submissions has apportioned the Restbury claim equally among those nine nodes to which he allocated it. I accept that that is a reasonable way to apportion these costs. I disallow the apportionment to HAZOP node 2.3.10. That leaves eight nodes among which the claim should be equally apportioned. I accept also that Restbury would have done software work on the main compressor HIPS, item 3.1.3, and on the main compressor blowdown sequence, node 3.1.5. Thus one-eighth of the Restbury claim should be apportioned to HAZOP node 3.1.3 and one-eighth to node 3.1.5. That amounts to £3,670 in each case.

In paragraph 495 above I have quoted evidence of Mr. Morrell that the final seven-day run of commissioning was outstanding when the works were suspended on 18^th April 1996. I accept his evidence that that run would involve the presence of two representatives for about three of the seven days. Mr. Charters said, and I accept, that a credit for this must be allowed against the claim in respect of vendors’ representatives. He used the prices charged by EGT. He gave it as his opinion that the sum of £11,396 should be allowed, calculated on the basis of two representatives attending for seven days. Using the period of three days in place of seven days, and substituting my figure of £44 per passenger-journey by helicopter for his figure of £41.78, his figures become:

The figure of £23 per man-day for catering does not represent the average implicit in the total of £143,669 used in paragraph 559 below. I have adopted it for simplicity, the total amount being small. Phillips have not challenged the figure; Mr. McMaster’s submitted figure for the total that I have found to be £5,482 was £5,400.

I find £5,482 to be the correct figure to use as the basis of the allowance. However, I have allowed only a small proportion of the whole claim of £186,183. The allowance of £5,482 should be reduced pro rata. The effect of that is that the figure of £13,818 for Hima-Sella is reduced to £13,411, and the figure of £3,670 for Restbury is reduced to £3,562 in relation to each of the two relevant nodes.

Phillips were charged a total of £143,669 for catering for personnel working on the platform during the relevant period. That figure is agreed. It includes £121,451 for catering for personnel on Seafox. I have disallowed the other costs of Seafox, but catering charges would have been incurred if Seafox had not been used, and there is no reason to suppose that the charges would have differed substantially from those imposed by Seafox. The experts were not agreed as to the amount that should be credited in respect of work that was not remedial work. In my judgment, the simplest way to treat this is to apportion the catering costs pro rata to offshore man-hours worked. In my judgment that is a fair way of apportioning the cost to the relevant HAZOP nodes. I adopt it here.

This claim is for materials directly ordered by Phillips. The total amount, which is not in issue as a figure, is £107,341. All these expenses can be allocated to individual HAZOP nodes. The allocation of these expenses to individual HAZOP nodes is not in dispute. Thus I need not consider further those that do not relate to HAZOP nodes which arise out of any breach of duty on the part of Snamprogetti. It follows that I need deal only with the following:

The claim in respect of Restbury applies to the same HAZOP nodes as in the case of the vendor representatives. I reach the same conclusion as before, namely that I allow one-eighth of the claim to each of the HAZOP nodes 3.1.3 and 3.1.5. That is £5,151 each.

Neils-Jamesbury supplied pressure control valves. This claim, in relation to HAZOP node 2.2.5, is for the cost of the new suction pressure control valve PV 1020 for the fuel gas compressor. That valve was introduced to produce a suction pressure of 99 psia. I have found that that was not appropriate, and a valve providing a pressure of 175 psia ought to have been supplied. There is no evidence as to the difference, if any, in price between the two articles. I shall assume that £5,031 represents a fair estimate of the cost of a suitable valve.

As to Foxboro, the evidence of Mr. Abernethy is that Foxboro supplied controllers. Two controllers were supplied, one for the fuel gas system (HAZOP node 2.3.11) and one for valve PV 1020 for HAZOP node 2.2.5. I accept that evidence. There is however, some confusion about this. The schedule that sets out the list of materials refers to HAZOP nodes 2.2.5, 2.3.11 and 3.1.3 in this connection. Mr. McMaster in his submission divided the sum claimed equally among those three nodes (£1,545 each). There is no evidence before me as to the price of each controller. I shall divide the sum of £4,635 equally between the two nodes mentioned by Mr. Abernethy.

The costing exercise starts with the direct hours (i.e., man-hours) worked by KYE labour.

The starting-point is the allocation of KYE labour hours to HAZOP nodes and other work as set out in an amended version of Rev6last. As a starting-point, those figures are not controversial. They appear in Appendix 1 to this judgment.

Mr. Charters, having studied the direct and indirect hours claimed for each trade, reached the conclusion that some of the direct hours in question ought to be, but had not been, allocated to AMEC and Bacton work. The persons concerned were foreman insulator, instrument fitter, instrument technician, insulator, logistics clerk, storeman, superintendent and welding inspector. Their direct hours totalled 1,423. Mr. Charters said, and I accept, that Phillips had failed to allocate any of that time to AMEC and Bacton work. I accept that it is highly improbable that none of those hours represented AMEC and Bacton work. As appears from Appendix 1, the number of offshore direct hours said to have been expended on AMEC and Bacton was 1948. That represents 11.44 per cent. of the sum of HAZOPS and AMEC and Bacton hours. Mr. Charters considered that that proportion of the 1,423 hours, namely 163 hours, should be apportioned to Bacton. (The distinction between AMEC and Bacton is not relevant for this purpose). Not knowing to what nodes any of the 1,423 hours related, Mr. Charters transferred the 163 hours from the All-HAZOPS category. I accept that he was right to do so.

The Phillips allocation had treated all work to the fuel gas heater as being remedial work to the fuel gas compressor. Most of that work appeared under HAZOP node 2.3.11. That node was concerned with advising on remedial measures concerning the vibration of the fuel gas compressor. Some of the work to the fuel gas heater was included under HAZOP node 2.2.5. That node was concerned with the fuel gas compressor suction pressure control valve PV 1020, but the job cards in question related both to that work and to valve PV 1021, which related to the fuel gas heater. Mr. Charters created a new category of quasi-node, Fuel gas heater. He transferred the relevant hours, 4,271 hours, that had been included under node 2.3.11 to the new category. He also apportioned half of the hours in the job cards that related both to node 2.2.5 and to the fuel gas heater to the fuel gas heater category. It is manifestly convenient to have the new quasi-node, and I accept Mr. Charters’s method of dealing with this matter.

In spite of what appear to me to be errors in amended schedule 5 and amended schedule 10 to Mr. Charters’s second supplemental report, I am satisfied that the effect of those changes is to transfer 19 hours from node 2.2.5 and 35 hours from node 2.3.11 to the new fuel gas heater node.

On the evidence of Mr. Morrell (Day 28, p.136), I am satisfied that 38 hours relating to fuel gas washers should be transferred out of node 2.3.11. Mr. Charters transferred those hours into Bacton as not being the responsibility of Snamprogetti. Whatever the logic of that process, in my judgment the conclusion is correct since Rev6last was designed to assess the hours of work rendered necessary by the faults of Snamprogetti.

Of the hours claimed under node 2.3.11, 122 related to the replacement of existing contactor spools. Those hours had been included under node 2.3.11 because the contactor spools were the main feed for the fuel gas heater (evidence of Mr. Abernethy, Day 26, p.101). Mr. Charters transferred those 122 hours from node 2.3.11 to Bacton on the ground that the work in question was maintenance work. In my judgment, the work was clearly maintenance work, and Mr. Charters was right to do so.

HAZOP node 3.3.1 for heat tracing of seals comprised 257 direct offshore hours. Of that time, 190 hours came from job card 0689. Job card 0689 was for insulation to miscellaneous pipework. It is uncertain to what that work relates. Mr. Charters transferred 190 hours from node 3.3.1 to Bacton. He was right to do so.

The effect of the changes mentioned in paragraphs 490 to 525 and 566 to 571 above appears in column (2) of Appendix 2 to this judgment.

The total offshore KYE man-hour costs are agreed at £605,016.48. It is common ground that they should be apportioned at an equal average hourly rate across all the items in question. Thus they will be apportioned in proportion to the direct man-hours allocated to each item. The all-HAZOP hours are similarly distributed among all the other direct-cost items. The allocation which I find to be appropriate appears from column (2) of Appendix 2.

The transfer of 558 hours from direct costs to indirect costs (which is not reflected in column 2 of Appendix 2) has no effect on the total costs of the individual items, since both the direct costs and the indirect costs are distributed across the items in the same ratios, namely in proportion to the direct offshore hours appearing in column (2) of Appendix 2. I have used the expression indirect costs here to comprise the three categories Head Office, Indirect and Non-productive shown in that appendix.

The apportionment of the total KYE offshore labour costs, direct and indirect, appears in column (3) of Appendix 2.

The total onshore cost attributable to KYE labour is also agreed at £181,818.63. For 8789.5 hours (see Appendix 1) that represents an average of £20.69 an hour. I calculate the onshore costs of KYE labour for each item as follows. First, I have discounted the head office costs of £139,740 by £6885, leaving £132,855 (see paragraph 529 above). That figure I have apportioned to the items by reference to the offshore direct hours. The onshore non-productive hours I have treated in the same way. They amount to £14,265.76, being 689.5 hours at £20.69 an hour. The total sum so apportioned is £147,120.76. Finally, I have taken the onshore direct hours applicable to each item and applied the average rate of £20.69 an hour. As regards the fuel gas heater, I have taken the figure of 815 onshore hours from Rev6last, which attributed those hours to HAZOP node 2.3.11. HAZOP node 2.3.11 recommended obtaining advice and report on the vibration of the fuel gas compressor. Manifestly, it was not in pursuance of that recommendation that the time in question was spent. Mr. Charters transferred those hours to the fuel gas heater. He was right to do so. I have adopted that approach. The result appears in Appendix 3.

The total KYE labour costs, offshore and onshore, are the sum of the entries in the final columns of Appendices 2 and 3. They are summarized in Appendix 4, column (4).

In my judgment, the helicopter and catering costs should be apportioned by reference to the KYE offshore hours (see paragraphs 544 and 559 above). The relevant figures appear in Appendix 5.

The (variable) costs of Phillips’s management and Bacton contractors should, in my judgment, be apportioned in accordance with the KYE labour, offshore plus onshore. The apportionment is set out in Appendix 6.

I am satisfied that the KYE material costs directly allocated to HAZOP nodes in Rev6last can stand. They appear in column (2) of Appendix 7. Those allocated to nodes by way of the relevant disciplines are in the all-HAZOPS category and ought to be spread over AMEC and Bacton in addition to the nodes. I reduce each apportionment in proportion to the numbers of direct offshore hours spent on the nodes and on the nodes plus AMEC and Bacton. There is no evidence upon which any other adjustment could be made. The resulting figures appear in column (4) of Appendix 7. I have done the same in the case of those all-HAZOPS materials that were apportioned to the nodes without reference to disciplines, save that I have first reduced the total for reasons given in paragraphs 537 and 539 above. The materials in question all fell within the all-HAZOPs category. This part of the claim amounted to £58,558.65, of which I have disallowed £28,414.80, leaving £30,143.85. The relevant figures appear in Appendix 8. The reduced all-HAZOPs materials costs appear in column (3) of that appendix, and are reproduced in column (5) of Appendix 7. The total allocation of materials costs appears in column (6) of Appendix 7.

The cost of transporting materials to the platform in vessels was £16,402. That is to be apportioned in accordance with the apportionment of the cost of materials. The cost of materials charged by KYE was £246,000, as appears in the back-up material to Rev6last. The cost of materials purchased by Phillips was £107,341. It does not appear what proportion of the materials was transported to the platform. Not all the materials were physical objects. One, for example, was training courses. Thus this is a rough and ready exercise. I shall take the denominator as £350,000. The KYE material costs appear in column (6) of Appendix 7 and are reproduced in column (2) of Appendix 9. The Phillips materials costs (see paragraphs 561 to 563 above) appear in column (3) of Appendix 9. The apportionment of the vessel costs appears in the last column of that appendix.

In Appendix 10 I set out my conclusions as to the costs of the remedial works on the relevant items, together with an apportionment of the costs of running the HAZOP which more properly belongs to the next paragraph.

The sum of £7165 is claimed as the cost of the Baker Jardine HAZOP. It is said that that would not have been necessary but for Snamprogetti’s breaches of duty. However that may be, the costs of the HAZOP were probably increased by those breaches of duty. No doubt the variable costs are only a fraction of the total costs. Nevertheless, I am satisfied that Phillips have suffered some small damage in this connection by reason of the breaches of duty. In my judgment, a fair estimate of the damage is to be arrived at by apportioning the whole cost equally over all the HAZOP nodes, including those involving no claim against Snamprogetti. The expression HAZOP node strictly identifies an item designated by two numbers, each recommendation being identified by a third number. However, in these proceedings the word node has been universally used to refer to a recommendation identified by three numbers. In counting nodes, I use the expression in the latter sense. There were 69 nodes. Each one for which I have found Snamprogetti liable thus attracts £103.84 in damages under this head, save for the water wash drainage. The water wash drainage (HAZOP nodes 3.7.1, 4.1.1) is a matter which constituted a purely technical breach of contract, for reasons given in paragraph 321 above. In my judgment, it should not attract more than nominal damages of £2. On the other hand, item 2.3.10, concerning valve PV 923, though leading to no costs, as I have found, did justify investigation by way of HAZOP.

HAZOP node 2.1.1, the Zechstein metering skid blowdown, is in Mr. Sylvester-Evans’s category 2(C). There may have been some loss, probably a small fraction of the cost that I have attributed to the item. But there is no material before me on which I can form even an approximate estimate of its amount. This claim must fail for want of proof of quantum, if any, save to the extent of £103.84. That sum represents a share of the cost of the Baker Jardine HAZOP, and constitutes a genuine loss. Phillips are entitled to damages of £103.84 for this breach.

HAZOP node 2.2.5 is the fuel gas compressor suction pressure control valve. Mr. Sylvester-Evans, with whom Dr. Robinson agreed, said the following in his consolidated report, paragraph 4.18.4:

HAZOP node 2.3.10 relates to the fuel gas compressor recycle valve PV 923. The recycle valve PV 923 was unsatisfactory. But it was not replaced, no doubt because the fuel gas compressor was abandoned. Phillips claim £7,976 for this item. That sum is made up as follows. £3,263 is claimed as one equal ninth part of the fees, £29,363, of Restbury International spread over nine HAZOP nodes. £88 has been allowed against that sum as a commissioning credit. There is no evidence what Restbury International did in relation to this recycle valve. I am not satisfied that anything was done. £223 is claimed as a share of the cost of supply vessels. The balance of £4,578 is claimed as one ninth of the cost of materials charged for by Restbury International. I am not satisfied that any materials were supplied in relation to this item. Thus this claim fails in the absence of proof of any loss, save for a share of the HAZOP costs. Damages: £103.84 for the HAZOP.

HAZOP node 3.1.3 relates to the HIPS for the main compressor. The work may be considered in two parts: removal of the HIPS from one location downstream of the compressor to another: and installation of a new HIPS upstream of the compressor. The former is category 1 and the latter is category 2(B). Mr. Sylvester-Evans wrote the following, with which Dr. Robinson agreed:

Mr. Charters considered that the KYE direct hours could reasonably be divided equally between each system. I accept that approach. I consider it reasonable to treat the attributed costs of helicopters, catering, Bacton contractors, vendor representatives, Phillips management and the HAZOP in the same way. Materials and supply vessels, however, must be attributed entirely to the suction HIPS, upstream of the main compressor. Phillips have failed to satisfy me that they have suffered any loss in relation to the suction HIPS save for the half share of the attributed HAZOP cost. As to the downstream HIPS, I find that the damages are one half of the expenses that I have mentioned. That amounts to £45,974.32. Allowing for the other half of the HAZOP expense, the damages under this head amount to 46,026.24.

HAZOP item 3.1.5, relating to the blowdown sequence, is category 1. The loss sustained by Phillips was £9058.23.

HAZOP item 3.2.5 concerns the main compressor start-up sequence. Mr. Sylvester-Evans said this in his consolidated report, paragraph 4.18.24:

Dr. Robinson said this in examination in chief (Day 61, p.166):

I do not accept that this item required the movement of a gauge. The HAZOP recommendation is:

Mr. King gave evidence in his first witness statement, paragraph 88, that a new pressure transmitter needed to be purchased and integrated into the start-up logic. Mr. Abernethy said in his first witness statement, paragraph 21:

I am not satisfied that there was any relocation of an instrument or cabling, as stated by Dr. Robinson. It is not clear on the evidence how much of the work claimed for was extra work over and above what would have been needed anyway. The cost figure of £6357 is, in my judgment, no measure of the loss. This claim fails except to the extent of £103.84.

HAZOP items 4.2.1 and 4.3.1 are both category 1 items. I am satisfied that Phillips have suffered loss in the sum of £20715 in the case of the former and £20835 in the case of the latter.

It follows that the total sum that I find represents the loss suffered by Phillips by reason of Snamprogetti’s breaches of duty is £97,052.

Snamprogetti counterclaim £321,143 from Phillips. Of that sum, £138,710 is claimed as an entitlement under milestone provisions of the contract. Liability for that sum is admitted, subject to set-off. A further sum of £144,188 is counterclaimed for 4270 additional design hours spent by Snamprogetti. Finally, the sum of £38,245 is counterclaimed as damages for failure on the part of Phillips to make an assessment pursuant to a risk and reward scheme provided by the contract.

It is common ground that Snamprogetti spent a further 4,270 man-hours on the project after (or after the beginning of) the Baker Jardine HAZOP. That appears on the pleadings. It is not possible for me fully to identify particular hours with particular design work. Mr. McMaster submitted that I should allow the proportion of the claim for design hours that reflected my findings on liability. I accept that submission. I shall assume that the design hours spent on each item were in proportion to the KYE offshore direct hours spent on that item. The KYE offshore direct hours spent on the HAZOP items plus the fuel gas heater are to be found in Appendix 2. They amounted in total to 9241.5, excluding the 1548 hours of all-HAZOP items. The latter were distributed pro rata among the former and do not affect the ratio. I found Snamprogetti liable to Phillips in relation to eight HAZOP nodes, of which six attracted KYE offshore hours. The offshore hours spent on the HAZOP nodes which I have found to be the responsibility of Snamprogetti amounted in total to 2077, again excluding the all-HAZOPS apportionment (Appendix 7). The balance is 7164.5. That is 77½ per cent. of the total. I find Snamprogetti to be entitled to 77½ per cent. of its claim in respect of additional design hours. That amounts to £111,745.

Mr. McMaster fairly described his attitude to the apportionment of Snamprogetti’s fees as a reasonably tolerant one. What he said was this:

The last element of the counterclaim arises under paragraph 3.5.1.2 of the price schedule to the contract. That paragraph is headed “Scheme 2 – Cost of Quality”, and so far as material reads as follows:

Snamprogetti claimed that Phillips, in breach of contract, had failed to make a fair and reasonable assessment, or any assessment, under the paragraph headed “Reward”. The sum Snamprogetti claim is £20 times 6 per cent. of 31,871, viz. £38,245, the maximum payable under that paragraph. Phillips, on the other hand, made a claim under the paragraph headed “Risk”. The threshold of rework hours qualifying Phillips for a payment was 6 per cent. of 31,871 hours, namely 1912 hours. Mr. McMaster submitted that Phillips were entitled to £10 for each rework hour spent by Snamprogetti post-HAZOP in excess of 1912 hours.

I have, in effect, assessed the rework hours at 22½ per cent. of 4,270, i.e. 960.75. Thus the threshold of Phillips’s claim has not been passed.

Phillips have made no assessment under the paragraph headed “Reward”. Given the existence of these proceedings, I am satisfied that if Phillips had made such an assessment it would have been for a nil value. Miss Boswell invited me to substitute my discretion for that of Phillips. I decline to do so. Assuming Phillips to have been under a duty to exercise their discretion reasonably, I am satisfied that a nil assessment would not be beyond the bounds of reason. This part of the counterclaim fails.

Snamprogetti are entitled to £250,455 on their counterclaim.

On the claim form it was stated that the claimant made the claim on its own behalf and on behalf of all co-venturers in the exploitation of the Hewett field as is provided for by clause 48 of Contract UK0601. In further information served pursuant to a request made on 9^th May 2002, after the start of the hearing, Phillips said, among other things:

By letters dated May and June 2002 ARCO British Ltd, Fina Exploration Ltd, Superior Oil (UK) Ltd and Petro-Canada UK Ltd confirmed that they were made aware of the proceedings before the Particulars of Claim were served and approved their service.

The re-amendment to the prayer stated this:

Miss Boswell accepted that Phillips can pursue a claim on behalf of its principals, the other co-venturers, if it has authority to do so. But she submitted that Phillips were not entitled to recover more than the combined share of Phillips and the authorizing co-venturers, namely 72.66 per cent. Thus the discount sought by Phillips was not a voluntary concession.

It is clear that in the circumstances neither Phillips nor any co-venturer will be entitled to bring a claim against Snamprogetti for the whole or any part of the balance of 27.34 per cent. Thus the question whether the concession is voluntary is purely academic. I asked Miss Boswell whether she had some practical reason for inviting me to decide the point, and she gave none. I decline to decide this point. There are other points: whether the loss is Phillips’s loss, and whether Phillips has suffered loss to the extent that it has been reimbursed. It is unnecessary for me to decide those points.

There was a claim against Snamprogetti International S.A. under a guarantee dated 28^th February 1994. The claim was for £5,394,747.42, or alternatively £4,032,612 plus an unquantified element not exceeding £75,000, or alternatively such sum as Snamprogetti was liable for in damages to Phillips. There was a dispute about the construction of the guarantee. But it was conceded that the liability of the second defendant did not exceed that of the first defendant.

Since I have found a net sum to be due to the first defendant, there is no liability on the second defendant.

I find the loss to be £97,052. The damages under the claim will be 72.65225 per cent. of that amount, i.e. £70,510. The sum due under the counterclaim is £250,455. There will be judgment in favour of Snamprogetti for the balance, £179,945. The claim against the second defendant is dismissed.

KYE LABOUR HOURS – Rev6last

Allocation and apportionment of KYE offshore man-hour costs.

Allocation and apportionment of KYE onshore man-hour costs.

KYE labour costs (£)

Paragraph 578 refers

Paragraph 579 refers