Associated British Ports v Hydro Soil Services NV & Ors

CONTENTS

These proceedings arise out of a contract made in March 2002 between the claimant (“ABP”) and the first defendant (“HSS”) under which HSS was engaged to strengthen the quay wall of berth 205 at the port of Southampton to allow that berth to be dredged from its existing depth of -12.8mCD (i.e. 12.8 metres below Admiralty Chart Datum) to -16.0mCD plus 0.5m overdredge allowance. The contract was a lump sum contract in the sum of £3,856,465.00. The contract was based on the ICE Conditions of Contract, 6^th edition, as amended by the parties.

On 29^th May 2002, ABP consented to the assignment of the benefit of the contract by HSS to the second defendant (“DIUK”). DIUK has been represented before me jointly with HSS. In this judgment I shall use the expression HSS to refer to the first defendant or both defendants indiscriminately.

Under the contract, HSS was responsible for the design and construction of the strengthening works. It subcontracted the design of the works to the Part 20 defendant (“Haecon”). Where appropriate, I use the expression “the defendants” to include Haecon.

The quay wall that required strengthening was a steel sheet pile wall (sometimes called the SSP wall) at the water’s edge topped by a coping beam. It was propped horizontally at the top at intervals by vertical triangular counterforts. The piles were about 27m in length. In cross-section, the piles were approximately U-shaped, alternate piles being oriented in opposite directions, and adjacent piles being interlocked. Behind the wall, at a depth of about 6m, was a relieving platform consisting of a concrete slab. The relieving platform was supported by H-piles (piles with an H-section). Its purpose was to bear the weight of the overlying soil and the cranes and other things on the dock and to transmit that weight to a low level, thereby reducing the lateral forces acting on the wall. The quay wall was constructed on reclaimed land in the early 1970s.

The design of the strengthening works incorporated the installation of ground anchors attached to the steel wall and the installation of very high pressure (VHP) vertical grout columns along the length of the berth. The VHP columns were installed by a process known as jet grouting. That process involves pumping grout, a mixture of cement and water, down a pre-cut hole so that it jets out of nozzles in a rotating “monitor” at the bottom of the hole and penetrates and mixes with the surrounding soil. The monitor is gradually drawn upwards while the jetting proceeds. The mixture of soil and grout makes a material commonly called soilcrete, which hardens forming the VHP columns. The columns are commonly called the grout wall or the VHP wall, but the expression refers to a series of discrete columns, whose average separation as designed was approximately 1.7m. Each column was in the form of a pair of cylindrical columns each of design diameter 900mm. The axial separation between the members of each pair was designed to be 700mm. The second was constructed immediately after the first and before the first had hardened, so the two members of the pair were designed to function as a single column.

The strengthening works took place between March and November 2002. During the course of the works the sheet piles cracked and bulged outwards by an amount up to one metre. It is common ground that further remedial works are now required to enable berth 205 to be dredged to -16mCD plus 0.5m overdredge allowance. The principal dispute is as to the cause of failure of the strengthening works and which party bears responsibility for payment for the necessary remedial works.

ABP claims that the strengthening works were not fit for their purpose. It is said that remedial works costing more than £11 million are required. ABP claims against HSS a declaration that HSS is responsible for defective works and for the cost of remedying them. ABP also claims damages against HSS for breach of contract. HSS counterclaims extra payment for the necessary remedial works on the basis that those works have been rendered necessary by an unforeseeable physical condition, namely that there were plastic hinges (Footnote: ¹) in the existing sheet pile wall. In the event that HSS is found liable to ABP by reason of defects in the design of the strengthening works, HSS seeks from Haecon an indemnity, contribution or damages for breach of the design contract.

Under its counterclaim, HSS also has two claims against ABP which have been entitled the coping beam claim and the grout disposal claim. I consider those claims below, at paragraphs 299 et seq. and 314 et seq. respectively.

The two major disputes are (1) ABP’s claim against HSS for the relief mentioned above on the basis that HSS was in breach of contract in that the works were not fit for their purpose, and (2) HSS’s claim against ABP for extra payment on the basis that HSS encountered an unforeseeable physical condition. I shall consider the latter first, but before doing so I shall identify the expert witnesses in this case.

The following expert witnesses gave evidence before me. Professor Milija Pavlovic Sc.D (Cantab.), Professor of Structural Engineering and Mechanics and Head of Section (Concrete Structures) at Imperial College, Fellow of the Institution of Civil Engineers, Fellow of the Institution of Structural Engineers, Fellow of the Association for Consultancy and Engineering; Mr. Anthony Bracegirdle M.Sc (Imperial College), Member of the Institution of Civil Engineers; Mr. William M. Reid, Fellow of the Institution of Civil Engineers, Fellow of the Institution of Structural Engineers, Fellow of the Royal Academy of Engineering; Dr. David Greenwood Ph.D., Chartered Engineer, Fellow of the Institution of Civil Engineers; and Mr. Anthony O’Brien M.Sc. (Imperial College), Chartered Engineer and Member of the Institution of Civil Engineers. Professor Pavlovic and Mr. Bracegirdle were instructed by ABP; Mr. Reid and Dr. Greenwood were instructed by HSS; and Mr. O’Brien was instructed by Haecon.

References in this judgment in the form (day d, page p) are to a daily transcript of the proceedings commissioned by the parties.

It is common ground that the effect of clause 12 of the contract is that if during the execution of the Works HSS has encountered physical conditions (other than weather conditions or conditions due to weather conditions) or artificial obstructions which conditions or obstructions could not reasonably have been foreseen by an experienced contractor, then HSS is entitled to extra payment for extra work rendered necessary thereby. In this case HSS rely upon a physical condition, namely the physical state of the sheet pile wall. They say that it was overstressed and contained plastic hinges. That meant that it was unable, without bulging, to resist the pressure of the grout applied at the back of the sheet pile wall when the VHP columns were made. HSS contend that it was reasonable to expect that the sheet pile wall would have had a factor of safety of about 2, and that if it had had such a factor of safety it would not have bulged or cracked when the grouting process took place.

Mr. Bowdery submitted that the condition of the sheet pile wall was not a physical condition falling within clause 12, since a contractor should not have the benefit of a clause 12 claim if his strengthening works fail because of some pre-existing condition in the works which he has undertaken to strengthen. Clause 12 was not intended to apply to the condition of the very thing that was to be strengthened. In my judgment, the concept of encountering something during the execution of the works tends to import something adjectival that hinders the work, rather than a substantive element of the works themselves. Nevertheless, the words are wide enough to cover the latter, and the clause must be construed against the person (here, ABP) who offers the document for agreement. Accordingly, I reject Mr. Bowdery’s argument. The questions remain, what was the condition of the sheet pile wall, and was it unforeseeable by an experienced contractor. I shall first consider the first of those questions.

In outline, the position is this. It is common ground that before the works were carried out on berth 205, berths 204 and 205 were in the same condition, and that berth 204 is still in the same condition as it was. The present condition of berth 204 has been investigated in order to determine the condition in which berth 205 was before the works were carried out. Surveys of berth 204 above the level of the river bed carried out before the trial showed that most of the sheet piles were not vertical. A small degree of displacement from the chord at mid-depth was to be expected, but generally the displacement from the vertical (where the chord would be expected to lie) was greater than expected. It was said that the maximum bending moments in the sheet piles would be proportional to the maximum displacements from the chord, a displacement of about 300mm indicating a bending moment at relieving platform level sufficient to create a plastic hinge there.

Mr. Reid, pointing out that in berth 204 there were many instances of displacements from the vertical of 300mm and more, expressed the opinion that in consequence there must be many instances of plastic hinges in the sheet pile wall. Mr. Bracegirdle considered that the piles had been driven consistently with a rake, so that the toe of each pile would be seaward of the vertical drawn from the top of the pile. Thus the displacement from the chord would be less than the displacement from the vertical, and one could not conclude that there were plastic hinges in the wall.

With the object of resolving this issue, a magnetic sonde was used in three locations to measure the position of the sheet piles both above and below the level of the river bed. The defendants submitted that the sonde survey showed conclusively that the piles had been driven vertically, and that the displacements from the vertical were indeed displacements from the chord. However, the position is not so simple as that. In more detail, it is as follows.

In the experts’ joint statement, the following propositions appeared:

In his expert’s report, prepared by 9^th September 2005, Mr. Bracegirdle put forward the thesis that the steel sheet piles had systematically been driven with a seaward rake, i.e. so that the toe was seaward of the head. Thus the displacements of the piles from the vertical would be greater than the corresponding displacements from their chords. He illustrated his thesis with diagrams in his report, contained in figure 24. He said that it was evident from the archive material that the reclaimed bank moved seawards during the driving of the sheet piles to an extent of about 2 inches (50mm). That was confirmed by the witness statement of Colin Gilbert (Footnote: ³). That movement was consistent, said Mr. Bracegirdle, with lateral loading imposed by the reclamation bank. As outlined in figure 24, movement of the reclamation bank and loading of the sheet piles would have caused curvature of the sheet piles. Seaward movement of 50mm at the top of the reclamation bank would be sufficient to induce a curvature of about 1/360 per metre in the sheet piles beneath the relieving platform. In reality, it would not be possible to carry out a definitive analysis of the complex issue of pile wandering. While the analysis presented was simplistic, it was sufficient to illustrate the mechanism of seaward wandering of the piles during driving.

Mr. Reid did not agree with the above analysis. He gave a graphic illustration of how the sheet piles could bow outwards below relieving platform level yet have their toes vertically beneath the positions of the piles at relieving platform level. I return to this point in paragraph 82 below.

In order to determine whether the toe of the sheet piles was located vertically below the position of the sheet piles at the level of the relieving platform, a magnetic sonde was used in October 2005. The magnetic sonde determined the horizontal distances of a pile from the measuring equipment at different levels at three locations in Berth 204. Those locations have been identified by the references X1, X2 and X3.

In the magnetic sonde, an electric pulse is passed through a coil. The pulse creates a magnetic field. The creation of that field produces a response from ferromagnetic objects (e.g. steel) within range of the coil. That response creates a signal in the coil which is detected by the electronics controlling the coil.

A survey using the magnetic sonde was carried out by GB Geotechnics Limited on berth 204 on 22^nd and 23^rd September 2005. The work was carried out by cantilevering a drilling rig over the side of the quay wall and drilling into the seabed about 18m below the rig. The drilling rig was used to instal a thick-walled cylindrical plastic tube. The plastic tube contained internal grooves which facilitated the orientation of an inclinometer and the sonde. The inclinometer was used to assess the position of any point in the plastic liner in a plane at right angles to the line of the sheet pile wall and in a plane parallel to the sheet pile wall. The sonde was restrained by the grooves so as to maintain the axis of the sonde coil at right angles to the sheet pile wall.

Three boreholes designated X1, X2 and X3 were made at different places along the length of the wall. Verticality of the boreholes was assessed by divers using spirit levels and tape measures at various levels from sea bed upwards. Verticality was also assessed by the use of the inclinometers. Inclination was measured by the force necessary to keep a pendulum in the borehole in an axial position. The inclination was a measure of the rate of change of the horizontal position of the instrument with depth. That was integrated to give the horizontal displacement of the instrument from a given datum, in this case the zero datum being the horizontal position of the inclinometer at the level of the bottom of the sheet pile wall.

Mr. George Stephen Ballard gave evidence before me. He founded the practice of GB Geotechnics and was Principal and Director of GB Geotechnics Limited. He stated that the magnetic mass sonde had been developed by his company, originally to measure the depth of iron caissons sunk into river beds in the nineteenth century to support railway bridges. It was not then its primary purpose to measure the distance of the caisson from the sonde. He said in his first witness statement, dated 25^th August 2005, that his firm was satisfied that the out of plumbness measured by the system in relation to berth 205 was accurate to better than 5 per cent. of the recorded value.

In a further witness statement, dated 29^th September 2005 and relating to the survey of berth 204, he said that both the magnetic sonde and the inclinometer data from borehole X2 showed strong evidence of contamination by wind, tide and wave, in both distance and tilt: the data had therefore been rejected for combination as a tilt corrected profile. He said that data from boreholes X1 and X3 also showed marginal influence from wave and tide action, but the evidence was that any error was considerably less than the amplitude of the real variations recorded. The data could be combined as reasonably valid tilt corrected pile profiles. He wrote:

The curvature of the sheet pile is the rate of change of direction (in a vertical plane perpendicular to the line of the wall) with depth. It may be measured in radians (circular measure of angle) per metre, expressed simply as per metre. The word curvature in the last paragraph of the above quotation clearly means displacement from the datum.

At location X3 the toe of the pile was indeed shown to be vertically below the upper position. The pile bulged seaward in between. At location X1, the toe was about 600mm landward of the upper point. The evidence from location X2 is less clear. Mr. Williamson relied on the information encapsulated in an exhibit (X30, page 2) that he put forward, which shows the toe of the pile slightly seaward, say 50mm, from the upper position. Exhibit X30, page 2 shows an average profile from the magnetic sonde results for location X2. I accept that X30, page 2 is the best estimate of that profile available to the court, but the confidence level as to its accuracy is low.

Regarding the shape of the sheet pile at location X1, Mr. Bracegirdle wrote in a report:

A witness statement from Mr. Colin Gilbert was put before me. He did not give oral evidence, but his written evidence was unchallenged. He was the senior engineering assistant to the resident engineer for the supervision of the “new” (early 1970s) quay wall works to provide four additional berths, namely Berths 202 to 205. In his statement, he said this:

Mr. Reid explained that when the piles were driven in pairs the hammer impacted on the clutch between the two. He gave some evidence of which the following is an extract (day 16, page 112):

British Standard BS EN 12063:1999 was put to Mr. Reid. That standard says at paragraph 8.6.1:

Paragraph 8.7.1 says:

Mr. Reid said this (day 22, page 44) in answer to questions from Mr. Bowdery:

Despite that evidence of Mr. Reid and Mr. Gilbert, the measurement of the position of the toe of the pile at location X1 convinces me that the piles cannot all have been driven vertically. Some may have been driven at a rake. Some may have wandered during the course of driving. It is, however, evident that the toes of adjacent piles could not have been widely separated.

I thus cannot accept that the horizontal displacement of the piles is a reliable indicator of the stresses or bending moments to which they have been subjected.

Given the existence of the profiles, it seemed to me (and in my judgment it is the case) that an obvious way to check whether the piles contained plastic hinges was to determine the curvatures of the piles from the profiles. The bending moment is a function of curvature, depending on the bending stiffness of the piles.

The bending stiffness is the factor, EI, by which the curvature has to be multiplied to arrive at the bending moment. Professor Pavlovic gave evidence that the bending stiffness of the steel sheet piles was 212,000 kNm²/m (Footnote: ⁵) and that the bending moment capacity of the sheet piles (i.e. the bending moment where the section becomes fully plastic) was 2500 kNm/m. There was a conflict of evidence about that. Mr. Reid presented calculations that what he called the plastic moment of the sheet piles (that is, the bending moment where the outer fibres reach their elastic limit: I shall use the expression maximum elastic moment) was 1776 kNm (by implication, 1776 kNm/m) and that the bending moment capacity was 2103 kNm/m. It was implicit in those calculations that EI was 226,600 kNm²/m. In his finite element analysis, Mr. Bracegirdle used a bending stiffness of 220,000 kNm²/m.

Professor Pavlovic said that the value of 212,000 kNm²/m was calculated by both Haecon and ABP, and that the moment capacity of 2500 kN-m/m was calculated by Haecon. Mr. Reid evidently relied on an undated page from a Larssen catalogue. Both witnesses used a figure of 3.55E5 kN/m² as the yield stress (elastic limit) of the relevant steel. (In the expression 3.55E5 the letter E refers to the exponent: it is not to be confused with the element E in the expression EI. The notation 3.55E5 means 3.55 X 10 i.e. 3.55 X 100,000; 6.5E-3 means 6.5 X [10 raised to the power -3], i.e. 6.5 divided by 1000, i.e. 0.0065).

The documents relied on by Professor Pavlovic were not identified, but I have researched the files of papers put before me. The figure of 212,000 (2.12E5) for EI appears on page 18 in section 5.2.2 of a document emanating from Haecon which is an as-built design report identified by the code USR2334 637 and marked revision 2 22-08-03. On the same page the figure of 2537 kNm/m appears as the bending moment capacity of the wall. Although I have not had the benefit of oral evidence or submissions on this point it appears that the figure of 2.12E5 kNm²/m is erroneous. The area A per unit length is stated as 4.2488E-2 m²/m, and EA is stated to be 8.92E6 kN/m (i.e. kN/m² times m²/m). From that, E can readily be calculated as 210E6 kN/m². I is given as 1.00952E-3 m. Multiplying that by E does indeed give a figure of EI equal to 2.12E5. But the units must be kNm². The units in which EI is stated are kNm²/m. It is stated on page 9 in section 5.1.1 of the same report that I is 1.00952E-3 m for the combined section, i.e. an inpan and an outpan. It is also there stated (and corroborated by Mr. Reid’s page from the Larssen catalogue) that the length of a combined section is 0.84m. Thus I, expressed in m/m, would be the figure stated divided by 0.84, i.e. 1.20E-3. EI, expressed as bending stiffness per metre, is thus 2.52E5 kNm²/m. Those are indeed the units in which EI is wrongly expressed to be 2.12E5 on page 18 of the report.

The figure of 2537 kNm/m for the bending moment capacity of the sheet piles is stated to allow for a safety factor. Given the yield stress of 3.55E5 kN/m², also expressed on the same page, 18, of the report, and information given as to the axial force, it is possible readily to calculate that the safety factor is 1.1. Thus the bending moment capacity implied is 2537 kNm/m multiplied by 1.1. That bending moment capacity is arrived at by multiplying the sectional area by the yield stress and by an undefined factor 2α which is evidently 0.185m. (The whole is then divided by the safety factor of 1.1 to arrive at the figure of 2537 kNm/m). To arrive at the bending moment capacity, that is the bending moment which causes (in theory) all the fibres at the relevant section to be subjected to their yield stress, it is necessary to integrate the elementary moments of the forces in the fibres about the neutral axis of the curved sheet pile. The elementary forces are the product of the yield stress (force per unit area) and the area of the element in question. The lever arm is the distance of the elementary area from the neutral axis. Thus the dimensions of the elementary moment (and hence of the integrated moment) are stress times area times distance. Hence the factor of 2α measured in metres. I can find no source which explains how the figure of 0.185m was arrived at. The page from the Larssen brochure gives data from which the factor 2α can be calculated. The figure is the ratio of the plastic section modulus to the sectional area. Both those figures are given on that page of the brochure. The plastic section modulus is stated to be 5.924E-3m³/m; the sectional area is stated to be 4.19E-2 m²/m (I have converted the units). The ratio is 0.141m. If that figure is substituted for 2α, the bending moment capacity comes out at 2127 kNm/m without a safety factor.

E is an elasticity modulus which applies in the elastic regime, where stress (force, here longitudinal force, per unit area) is proportional to strain (here, fractional increase in length). Where that regime applies, the ratio is a constant E known as Young’s modulus. The value 210E6 kN/m² for E is implicit in the section of the Haecon as-built design report to which I have referred above. The same value is explicit in an earlier design report of Haecon, identified by the reference USR2334 31 and dated 08-11-01, where it appears at page 14. Mr. Reid has taken the value 206E6.

For the foregoing reasons, I cannot accept Professor Pavlovic’s figures of 2.12E5 kNm²/m for EI or his figure of 2500 kNm/m for the bending moment capacity of the sheet piles. I accept the figures used by Mr. Reid. There are, it is true, minor discrepancies in the parameters A, I and W (see below) between the page of the Larssen brochure that he relied on and the information contained in an undated Larssen Steel Sheet Piling Handbook which was exhibited to the witness statement of a witness who was not called to give evidence before me, and between those sources and the Haecon calculations. In fact, the differences between Mr. Reid and the Haecon reports as to the values of E, I and EI make little difference to the outcome that is relevant for present purposes, that is the curvatures necessary to produce the maximum elastic moment and the moment capacity of the sheet piles. I set out in the table below the figures, implicit or explicit, in the various sources. The parameter W is the so-called section factor or elastic section modulus. The last line of the table gives the curvature required in each case to give the maximum elastic bending moment.

Since I have not accepted the value of 2500 kNm/m as the bending moment capacity of the piles, the only acceptable evidence of the ratio between the maximum bending capacity and the maximum elastic moment is that contained in Mr. Reid’s extract from the Larssen brochure. That ratio is 5924/5005. It leads to the value of 2103 kNm/m for the maximum bending capacity on Mr. Reid’s figures, which I have accepted. A figure of 2100 kNm/m was spoken to by Mr. Bracegirdle. If the moment remained proportional to the curvature beyond the value of 1776 kNm/m, the constant of proportionality being EI, the curvatures which give rise to the capacity bending moment would be respectively 9.27E-3, 9.12E-3, 9.09E-3 and 9.09E-3.

The Haecon calculations which I have referred to for the purpose of resolving this conflict of evidence between Professor Pavlovic and Mr. Reid were not put to any witness nor were they the subject of submissions. I disregard them when I come to consider the fitness of the design.

The bending moment that mobilizes the elastic limit of the steel is 1776 kNm/m. That causes a curvature of 7.83E-3 per metre. At that bending moment, the outer fibres of the steel are stretched to their elastic limit. They can be stretched further, but offer no increase in resistance. At that point, a plastic hinge starts to form. The hinge is fully plastic at a bending moment of 2103 kNm/m. The moment increases less than proportionally to the curvature as the plastic hinge forms. Thus the curvature of 9.27E-3 given in paragraph 43 above must represent a lower limit. Once a plastic hinge is fully formed over a significant length of the sheet pile, 2 metres or so, the pile, if unrestrained at its sides or adjacent to piles in similar condition, can be pushed out further with little increase in the force acting on it. The same applies to the whole wall, with adjacent piles interlocked, if similar forces are acting along all of it.

I invited the experts through counsel to draw smooth curves representing the best fit of a smooth curve to the profiles. From those curves one can determine the curvature. A smooth curve must be drawn since there will have been random errors in the measurements. The sheet piles could not have been in a shape other than a smooth curve: both Professor Pavlovic and Mr. Reid gave evidence to that effect. A profile of a pile shows a vertical section of the wall, perpendicular to the line of the wall. It expresses the relationship between y and x, where y is the horizontal position of the pile in question, measured at right angles to the line of the wall, at any given depth x.

Mr. Reid and Professor Pavlovic responded to that invitation. Both found the smooth curves by computer programs which found the best fit of a profile, subject to the constraint that y was a polynomial function of x of a given degree. The degree of a polynomial is the highest power of x that it contains. Broadly speaking, the higher the degree the more closely can the curve follow the vagaries of the points. But following the points too closely will give a false picture since the shape of the piles is smooth but there are random errors in the positions of the points. Mr. Reid calculated a quartic polynomial as a best fit for the profile of X3. Professor Pavlovic calculated a number of polynomials. He calculated polynomials of all degrees from the third to the ninth for each of X1, X2 and X3. Both gentlemen specified the polynomials in question by reference to their coefficients of the powers of x. Professor Pavlovic also calculated polynomials to fit data presented by Mr. Reid in his report. Those data related to six other piles in berth 204 which had not been examined by the magnetic sonde but had been measured by divers on behalf of HSS. Mr. Reid had drawn profiles of them above the level of the sea bed. Professor Pavlovic used the polynomial curve-fitting procedure in relation to four of those six profiles. He did not specify the relevant polynomials. He concluded that away from the ends of the range of measurements the highest curvature shown by any of the profiles was 6.5E-3.

Both Professor Pavlovic and Mr. Reid entered serious caveats against the use of polynomials to determine curvature. Professor Pavlovic said this with reference to the sonde data and the divers’ measurements:

Professor Pavlovic made a second point. He said this:

Those objections involve two distinct matters. To take Professor Pavlovic’s second point first, there is the error in curvature which best-fit polynomials are prone to exhibit towards the ends of the ranges that they are designed to cover. I accept that that error exists, and that the curvature figures generated for the points towards the ends of the respective ranges are useless. Second, there is the proposition that differentiation of a function, especially double differentiation, is liable to produce wholly unreliable values of curvature, whatever the function, polynomial or other. I shall consider that point further.

The essence of the argument is this. Suppose there is a relationship between two physical quantities that can be represented graphically by a smooth curve. Suppose measurements are taken of those quantities in a real experiment or observation. Suppose the measurements are subject to random errors and are plotted as a graph. A “best fit” curve will be a smooth curve showing the trend line, but a curve that passes through all the points will be a squiggly line. It is of the nature of squiggly lines that their curvature rapidly varies from point to point on the line. The values of the curvature will change between high positive values and high negative values, necessarily passing through zero in between. Thus if one measures, or calculates, the curvature at any point it is likely to have a value far removed from the average value in the vicinity of that point, i.e. from the value of the curvature of the trend line in that vicinity. If the squiggly line is represented by a mathematical function, polynomial or other, the curvature can be calculated by differentiation and double differentiation of that function. The answer will be correct, but useless, since it not the curvature of the squiggly line, but that of the trend line, that is desired.

The polynomials that have been generated by Professor Pavlovic are (I accept) the best-fit polynomials of the given order in each case that can be found to fit the data, at any rate to as high a degree of accuracy as is necessary for present purposes. They thus represent the smooth curve, not the squiggly line. That the objection to double differentiation is in most cases negligible for present purposes is clear from the actual results. In appendices 1, 2 and 3 to this judgment I set out figures that I have calculated from Professor Pavlovic’s polynomials relating to the sonde measurements using boreholes X1, X2 and X3. I have calculated both the y (horizontal position) value for each given value of x (the vertical position) and the second differential of y with respect to x for each given value of x. I have also shown the measured y value for each value of x. Thus it is possible to see how far the smoothed-out curves depart in their y values from the measured values. The second differential of y with respect to x gives a sufficiently accurate value of the curvature of the polynomial function, given that the gradient (the first differential of y with respect to x) is everywhere small. I have not calculated results for every polynomial. But the consistency of these results along each curve is proof enough that the objection to double differentiation does not apply for present purposes. If it did, one would expect rapid oscillations in the values of curvature as the x value gradually changed.

However, there remains the point mentioned by Professor Pavlovic that curvatures are extremely sensitive to even small errors in the local values of displacement (i.e. y). I accept that important caveat.

The higher polynomials, the seventh to ninth, show increasingly more oscillation than the lower polynomials in the magnitudes and signs of the curvatures. It is the higher degree polynomials that more closely follow the vagaries of the individual readings.

The quintic polynomial representing the profile derived from the X2 borehole shows a curvature of 8.36E-3 in the region of x equals -2.25mCD to -5.25mCD. Mr. Ballard rejected the results from borehole X2 since they were highly erratic as a result of stormy weather, which interfered with the measurement of the y-values. Mr. Williamson submitted that an averaging of the various results obtained from borehole X2 gave a good match to the netsurvey results (see paragraph 69 below) for the same sheet pile. Thus, he submitted, they were reliable. There was in my judgment sufficient force in that submission for me to have thought it worth while to calculate, as I have, the curvatures represented by that average. But the profile is far from a smooth curve, and I cannot regard the resulting curvatures as calculated from the polynomial of orders 5 and above as reliable in that case. Further, there is a kink in the y value from borehole X2 at about -15mCD which clearly cannot represent reality (nor has it been suggested that it does) and which is reflected increasingly in the higher polynomials.

Apart from the exceptions that I have mentioned above, no curvature greater than 5.42E-3 appears at any point between the relieving platform level (0mCD) and minus 16.5mCD for any of the polynomials up to those of the sixth degree that I have used to calculate the curvature of the sheet piles measured by means of the sonde.

Mr. Reid, in a document handed up to me, made a number of criticisms of the method of calculation of curvature by curve-fitting. I consider three of them here. I set them out below together with my comments:

In the same document handed up to me, Mr. Reid expressed the view that curve fitting to a fourth degree polynomial is probably preferable to curve fitting to higher degree polynomials. I have mentioned (Footnote: ⁶) Professor Pavlovic’s view that higher order polynomials will not improve the accuracy of the curvatures. I accept both of those views. The question arises why Professor Pavlovic and Mr. Reid arrive at different values of the curvature represented by the profile of X3. Mr. Reid’s quartic polynomial shows a maximum curvature of 7.36E-3/m. Professor Pavlovic’s quartic shows a maximum curvature, ignoring the five x-values at each end of the range, of 5.38E-3/m. Professor Pavlovic has put forward reasons for that discrepancy. But in my judgment there is another reason, which has not been mentioned.

Professor Pavlovic derived his data from a document identified as X22. That was put forward by Mr. Williamson as a representation of the profile obtained by the sonde. The original profile obtained by the sonde was exhibited to Mr. Ballard’s witness statement. It was in the form of a graph of horizontal location y against vertical position x. Thus it showed the shape of the profile, modified by the difference in scale between the horizontal and the vertical axes. The scales were small, and the plot was not on graph paper. Thus it was impossible to read off the precise data from the graph. That was no doubt a perfectly reasonable way to represent the data, given the wide margin of error to which Mr. Ballard spoke. The horizontal scale of X22 was even smaller, and it was even more difficult to read off the horizontal positions accurately. When I read off the data in the two cases, I obtained different results. Mr. Reid’s quartic and Professor Pavlovic’s quartic do indeed yield different y values for the same x values. I have carried out a least-squares exercise, and I find that Professor Pavlovic’s quartic represents the y values as I have read them off X22 better than Mr. Reid’s quartic represents the data as I have read them off Mr. Ballard’s graph. But in my judgment, if there is a discrepancy between the two sources, Mr. Ballard’s is to be preferred to X22, being the original source and easier to read. Mr. Reid’s quartic better represents my reading of Mr. Ballard’s profile X3 than does Professor Pavlovic’s.

I have also considered the distance of the bulge from the chord line, where that distance is a maximum. I have taken the chord line from x=0mCD (not the precise position of the relieving platform) to the toe of the pile, as shown on the profiles. The distances, as I have measured them, are 283mm for Mr. Ballard’s profile, and 271mm from X22. The corresponding value taken from the curve produced by Mr. Reid’s quartic is 279mm from the chord of that curve, that for Professor Pavlovic’s quartic is 255½mm from the chord of Professor Pavlovic’s quartic. Thus Mr. Reid’s quartic appears to be the more accurate.

The foregoing discrepancies between the two quartics illustrate the sensitivity of the curvature to the data. The data in this case are not claimed to be accurate.

Mr. Reid’s quartic intended to match the profile of the wall as measured from borehole X3 showed a maximum curvature of 7.36E-3, correct to 3 significant figures. That was at x=15.1m. Mr Reid calculated a bending moment of 2072 kNm (per metre run) at x=15m, where the curvature calculated to 3 significant figures was the same. The value of EI implicit in the maximum elastic moment capacity of 1776 kNm/m which I have accepted on the evidence of Mr. Reid is 226,600 kNm²/m. The calculated curvature of 7.36E-3 then implies a bending moment of 1668 kNm/m. At the same time as he adduced the figure of 2072 kNm/m, which he did by way of correction to a previous calculation of his, Mr. Reid abjured the use of polynomials to determine curvature, as indicated above.

Mr. Reid went on to criticize a response that Professor Pavlovic had given to the same question from me. He said (as I have accepted (Footnote: ⁷)) that Professor Pavlovic had quoted too high a value, 2500 kNm/m, for the bending moment capacity of the steel sheet piles. Mr. Reid said this:

Mr. Reid’s allowance for corrosion misses the point. Mr. Reid’s reduction of 10 per cent. for corrosion implies a corresponding reduction in EI. In that case, the bending moment implied by a given curvature is reduced correspondingly. As I have said above (Footnote: ⁸), I accept Mr. Reid’s figure of 1776 kNm/m as the limit of elastic flexural capacity of the steel. In Mr. Reid’s calculation of that figure there is implicit the value of 226,600 kNm²/m for the bending stiffness EI, and he correctly states the radius of curvature as 128m, which is equivalent to the curvature of 7.83E-3 per metre mentioned above (Footnote: ⁹).

The sheet piles located at X1, X2 and X3 constitute a very small sample of the sheet piles composing berth 204. A survey of the sheet piles of berth 204 was carried out in July 2005. As Mr. Reid explained, the survey was conducted under water using a spirit level equipped with a device to measure offsets from the vertical over one metre lengths. Mr. Reid concluded:

As mentioned in paragraph 47 above, Mr. Reid appended to his report of September 2005 a copy of output from that survey. I have examined all his plots of the output, six in number. They show “deflection” (horizontal position) against level. At location 20, readings were taken from +1mCD to -12.8mCD (sea bed level). There was a bulge with a maximum at -8mCD. The displacement between relieving platform level (0mCD) and the maximum of the bulge was about 95mm. I have made a rough calculation of the average curvature in the region where it appears greatest by measuring the gradient of the plotted curve at two points and calculating the average rate of change of the gradient per metre change in level. The curvature as so calculated is less than 5E-3 per metre. At location 25, where readings go down to -11.72mCD, there is no bulge. The maximum displacement is 12mm and the curvature obviously negligible. At location 160, the graph goes down to -8.37mCD. There is the top part of what may be a bulge at -7 to -8mCD. The maximum displacement is 160mm. The curvature, as calculated by me, is 7E-3. At location 185, where readings go down to the level -9.29mCD, there is a bulge at -5mCD. The maximum displacement is about 65mm. My calculation gives a curvature of 5.5E-3. The plot of location 215 goes down to the level -8.68mCD. No bulge is reached by that level. The maximum displacement shown is 260mm. The curvature is clearly negligible. Finally, location 310 goes down to -11.42mCD. The maximum displacement shown (though apparently increasing with depth at -11.42mCD) is 425mm. The maximum curvature below the relieving platform is negligible.

Location 160 is close to location X2. Location 310 is close to location X1.

I have mentioned the Netsurvey above. Two such surveys were carried out, one in 2004 and one in 2005. I shall confine my attention to the former, since there is no dispute as to the reliability of its results, but the latter is controversial. The Netsurvey was a multibeam system of measurement of distance from a boat, using a global positioning satellite system and with corrections for movement and angle of the boat.

In the Netsurvey, points were plotted which showed the profiles of the piles above seabed level at 5 metre intervals along the wall in several berths including berths 204 and 205. I have looked at all the profiles covering berths 204 and 205. Berth 204 extends from bollard 216A to a point close to bollard 236; berth 205 extends from that point to bollard 258. Bollard 258 is the zero datum of the survey for measurement along the quay in metres. Bollard 236 is close to the 330m point; bollard 216A is close to the 630m point. There is a striking difference between the profiles for berth 204 and those for berth 205. That is scarcely surprising, since, as is common ground, the piles in berth 205 have been pushed out by the grouting works. The curvature and displacement of the piles is apparent, with the maximum displacement generally at about the level of -9 to -10mCD. The profiles of berth 204, by contrast, are comparatively straight and show less displacement. It is not possible accurately to tell the curvature of the piles from these small-scale profiles. But the curvatures of the vast majority appear to be negligible for present purposes. Those which appeared to me to embody the greatest curvatures are identified in the table below.

The profiles which appeared to me to embody the greatest curvatures near the maximum of the bulge were at chainages (metrages) 335, 345, 375, 515, 525, 530 and 570. The respective horizontal displacements from the positions at 0mCD are also set out in the table. I made a rough estimate of the maximum average curvatures over a few metres’ depth in each of those cases by measuring and calculating the change in direction of the profile in radians per metre depth. The curvatures I obtained were respectively 5E-3, 5E-3, 8E-3, 3E-3, 3E-3, 5E-3 and 5E-3 per metre. Only the third of those represents a bending moment in excess of the maximum elastic moment. The horizontal displacements could not be measured with any precision; the stated curvatures are even less reliable. Nevertheless, it is clear that in the vast majority of these numerous profiles no curvature at the bulge and above the sea bed is shown which approaches that representing the maximum elastic bending moment, let alone a plastic hinge.

The profiles which appeared to me to embody the greatest curvatures near the relieving platform were at chainages 340, 350, 355, 360 and 375. The curvatures at chainages 340, 360 and 375 suggest the presence of plastic hinges. But the Netsurvey readings in the vicinity of the relieving platform may be affected, as pointed out by Professor Pavlovic, by the presence of low water corrosion plates attached to the sheet piles by way of repair or reinforcement in that vicinity. Corrosion had to be particularly guarded against in the intertidal zone, which extended down to -0.6mCD. If such plates appear in the Netsurvey profiles, they will have the effect of increasing the apparent curvature of the sheet piles near the relieving platform. Thus I cannot regard the curvatures that I have estimated at that level as reliable.

With the single doubtful exception of metrage 615, which possibly shows a small landward offset, all the profiles that show non-zero offsets show seaward offsets. Several piles which appear to be straight have substantial offsets. The following are examples. I give the offset at the lowest point recorded: 390m, offset 380mm at -11mCD; 435m, 500mm at -10mCD; 485m, 250mm at -11mCD; 500m, 350mm at -10½mCD; 595m, 580mm at -12¾mCD.

In my study of the Netsurvey of berth 204, which I made without reference to the other evidence, I found seven locations (and only seven) out of the total of about 60 where the visible curvature in the vicinity of the bulge seemed to be significant. One of those, at metrage 375, turned out to be close to X1 (which was at about 380m) and to Mr. Reid’s location 310. Another, at 515m, was close to X2 (which was at about 507m) and to location 160. A third, at 570m, was close to location X3 (which was at about 573m). Thus, whether by chance or by design, X1, X2 and X3 were located at or near points of exceptionally high visible curvature. The profile that I have noted in paragraph 69 above at metrage 530 is, I think, the one that appears as having a 650mm displacement, the maximum shown, in figure 4 of Mr. Reid’s report.

The location mentioned by Mr. Reid as location 185 is situated at about 475m on the Netsurvey metrage. The curvature shown on the Netsurvey in that vicinity is negligible. The profile that Mr. Reid exhibited in his appendix is anomalous in that the maximum is shown at -5mCD. I regard that as implausible. I accept the Netsurvey result.

Mr. Williamson relied on a finite element analysis produced by Mr. Bracegirdle which showed a bending moment of 600 kNm/m at -10mCD corresponding to a bulge of 100mm at that point. He invited me to conclude that X3, showing a bulge of 300mm, must have been sustaining a bending moment of 1800 kNm/m, a moment in excess of the maximum elastic moment. The sonde profile shows a deviation from the chord of 283mm. If the proportionality held exactly, and the figure of 600 kNm/m were exactly accurate, then the bending moment at the maximum of the bulge on X3 would be 2.83 times 600 kNm/m, i.e. 1698 kNm/m. These figures are not, of course, at all as accurate as their apparent precision. But Mr. Bracegirdle’s finite element analysis has the pile almost vertical at 0mCD, and its tangent rotates outward before it starts to rotate inward as one runs one’s eye down towards the maximum of the bulge. That may produce a different curvature, and a different calculated bending moment, at the maximum of the bulge, than would be the case for the shape of X3 where (as in the case of X2) the pile is raking seaward even at relieving platform level. It is nevertheless noteworthy that the bending moment of 1698kNm/m is close to the figure of 1668kNm/m arrived at in paragraph 62 above.

The above considerations in my judgment tend to confirm rather than otherwise the usefulness of estimating curvature to determine bending moment. Both methods are affected by errors in the original profile measurements. Apart from that, inaccuracies in determining the curvature by using polynomials arise principally, in my judgment, not at all from the process of differentiation but from non-conformance of the polynomials with the smooth curve represented by the true profile. But to extrapolate from the displacements shown in three instances, X1, X2 and X3, (which themselves show that the piles have not in all cases been driven vertically) to the conclusion that the bending moment is in all cases proportional to the displacement of the pile from the vertical passing through the pile at the relieving platform level is in my judgment wholly unjustified.

Mr. Bracegirdle said in his report of September 2005, at paragraph 11.11.14, that finite element analyses of the sheet pile wall in berth 204 suggest that the maximum displacements occur at about -10mCD; this, he said, is about the mid-point of the span below the relieving platform and the logical position for maximum displacement.

It is common ground that the bending moment at the toe of a sheet pile will be zero. Mr. Reid expressed the view mentioned in his proposition 3 (Footnote: ¹⁰) that the maximum curvature will be either at the tie level (i.e., at or near the relieving platform level) or at a depth of about -8mCD.

I am thus satisfied that the maximum curvature will generally occur above sea bed level. Thus the points of maximum curvature below the tie level should have been visible in the generality of the Netsurvey results.

The nearest candidate to represent X1 that I can find from the Netsurvey is the profile taken at 385m. That shows a maximum offset of 100mm at -7.5mCD and an offset of 50mm at -11.5mCD. If the toe really was at 600mm landward, as measured by the sonde, that would, I think, require a curvature below the sea bed (and thus not visible to the Netsurvey) involving a bending moment in excess of the maximum elastic bending moment. However, the sonde produced no reliable readings for the levels between -4mCD and -14mCD. That fact, coupled with the stated inaccuracy in its measurements, leads me to a state of uncertainty as to the degree of curvature of X1 below the sea bed.

If the evidence of curvature stood alone, I should reach the following conclusions. In the case of locations X1, X2 and X3 and their immediate vicinities, I would not be satisfied that there were any bending moments as great as the maximum elastic bending moment, let alone any plastic hinges. Except in the cases of X1, X2 and X3 and their immediate vicinities, I would conclude that the curvature in the sheet piles of berth 204 below the tie level nowhere exceeded about 5E-3/m, thus yielding a factor of safety of about 1.5 in relation to the maximum elastic moment, and of about 1.9 in relation to the formation of plastic hinges. I would conclude that in most places, the factors of safety were substantially greater.

When the excavation for the relieving platform was carried out, the pressure of the ground and water on the seaward side of the sheet piles above the level of the relieving platform caused the sheet piles to tilt above that level so that the top of the sheet piles moved landward. Mr. Bracegirdle and Mr. Reid were agreed about that. Mr. Reid expressed the opinion that the sheet piles had been driven vertically, but that when the land on the seaward side of the sheet piles was excavated to -12.8mCD, the piles bulged outwards with a maximum bulge at about -10mCD by reason of the bending moment caused by the tilt above the relieving platform. That movement would reduce the bending moment that had caused it. No calculations were proffered in support of Mr. Reid’s theory, but he made a model which graphically illustrated it. It involved rotation of the sheet pile wall outwards “through” the passive soil below -12.8mCD (i.e. moving the passive soil outwards and compressing it) during the course of the excavation to -12.8mCD, with the toe of the pile stationary. Mr. Reid’s theory may have some element of truth in it, but I cannot accept it in its entirety, for reasons that follow.

Mr. Bracegirdle expounded a theory that because of seaward movement of the heads of the piles while they were being driven, the piles would be driven with a seaward rake, i.e. so that the toe was seaward of the head. It is not controversial that such seaward movement of the heads of the piles occurred. So far as I understand it, Mr. Bracegirdle’s theory works as follows. The seaward movement of the top of the pile in the landfill above the Valley gravels at an early stage of driving induced curvature in the pile since the bottom part of the pile was more or less vertical in the Valley gravels. The curvature in the sheet piles was maintained as driving continued, so that the toe of the pile described an arc as it descended, and the upper parts of the pile followed in its track. Thus when the pile was driven it bulged landward. Subsequent excavation of the landfill to construct the relieving platform caused the head of the pile to move landwards, as mentioned above. That more or less relieved the bulge and left the pile more or less straight but raked, with its toe seaward of its head. That appears to me to require lateral movement (rotation) of the pile “through” the Valley gravels at that stage, and to a lesser extent “through” the Selsey sands. When the berth was dredged to -12.8mCD the sheet piles bulged outwards, but retained their general seaward rake. I cannot accept (since I can see no reason for it) that a pile retains its curvature while driven. There is no suggestion that it does not remain elastic: its curvature at any given time depends upon the stresses to which it is subjected at that time. Thus there seems to be no reason why the toe of the pile while being driven should carve out for itself a path with a curvature similar to that of the part of the path which it has already traversed.

I prefer Mr. Reid’s theory to that of Mr. Bracegirdle. But the actual measurements convince me that neither represents the complete truth. I am satisfied on the evidence of X1 that the piles were not all driven vertically. And the seaward offset of the piles is not always explicable by a bulge. Piles that are straight as far down as the Netsurvey could detect their position show an offset in several cases. Only one shows a landward offset. Thus there was a tendency for the piles to be offset seaward regardless of bulge.

In soil mechanics, parameters used in calculations are frequently represented by symbols which consist of a letter, say c, (sometimes primed, say c´) with a following subscript (say u). In this judgment, I shall write such symbols as a pair of letters in brackets, say (cu) or (c´u).

Mr. Bracegirdle caused to be carried out a finite element analysis designed to show, among other things, the effect of soil pressure on the SSP wall. The validity of that analysis was challenged by reference in particular to a learned article by D.M.Potts entitled Numerical Analysis: a virtual dream or practical reality?: (2003) Géotechnique53, No. 6, 535-573. The article pointed out (p.570) that a significant amount of human error is involved, and that implied that the users had insufficient training and/or experience in the use of advanced numerical analysis. It was not put to Mr. Bracegirdle that the persons carrying out the numerical analysis he used had insufficient training or experience. The program he used was called ICFEP: the Imperial College Finite Element Programme. That program had been developed by the same Professor Potts. Mr. Bracegirdle had derived advice and assistance, particularly in relation to small strain analysis, from Professor Richard Jardine of Imperial College. Professor Jardine had developed the non-linear soil model and had applied it to soil mechanics problems for a period in excess of 20 years. Variations in the model and soil parameters had been applied to assess the sensitivity of the model to wall friction and dilation. It was found that the outcome of the model was insensitive to those parameters. More to the point was the question of the validity of some of the other data used in the program. Mr. Bracegirdle’s use of a value of 35° for φ´ was criticized. Mr. O’Brien said, and I accept, that the model did not allow for variations in the geology along the wall. The analysis modelled only the locations where there was a deepened channel of valley gravel. Elsewhere, the Marsh Farm formation was higher. The model was a 2-dimensional model for a 3-dimensional situation. Mr. O’Brien considered that the assumptions and idealizations made for the ICFEP analysis used by Mr. Bracegirdle meant that the result was of little value in identifying the main issues: there was insufficiently detailed knowledge of the relevant soil parameters to give a reliable outcome. There were other criticisms, e.g. that the grid used was not fine enough.

Mr. Reid made a specific criticism of the finite element program by reference to a graph of shear force in the grout wall as a function of depth (run 200) appended to Mr. Bracegirdle’s report. Mr. Reid said that it was self-inconsistent. He said that by differentiating the function produced by the program as representing the shear force as a function of depth one would obtain the (horizontal) loading. If that were done, it would appear that the loading below –23m CD was in the same direction as the loading between –7 and –13m CD. He said that that showed an inconsistency. He pointed out (correctly) that Mr. Bracegirdle had said that the force below -23mCD was in a direction from seaward to landward, whereas the loading from –7 to –13m CD must be from landward to seaward. Mr. Bracegirdle in reply stated that above about –14m CD locked-in stresses on the seaward side of the grout column exceeded the stresses on the landward side. It was below that level that the stresses on the landward side exceeded those on the seaward side. Thus the stresses which Mr. Reid said were wrongly shown to be in the same direction were indeed in the same direction, from seaward to landward. Mr. Reid also made the criticism that differentiating the function produced by the program as representing the bending moment as a function of depth would produce a shear force function which differed from that produced by the program. Mr. Bracegirdle pointed out that since the grout wall was subject both to compression and to bending, differentiation of the bending moment plot to derive shear forces was incorrect and would produce erroneous results. I reject both of Mr. Reid’s criticisms of the finite element analysis.

Mr. Williamson submitted that Mr. Bracegirdle’s explanation was untenable for three reasons. His first was this:

I shall consider that passage proposition by proposition. (1) The first sentence I accept. (2) “This will generate strains in front of the wall”. The strains are the movement, and the movement of the walls will generate strains in the soil. (3) “Strains.....that will mobilize passive pressures”. There will already be passive pressures. The strains will tend to counteract the reduction in passive pressures caused by the dredging. (4) “These will act beneath a level of -16.5mCD in a seaward to landward direction”. The passive pressures will act on the grout wall in a seaward to landward direction. What is “mobilized”, however, an unfortunate word, is a reduction in such pressures. If considered separately as a pressure on its own, it will act on the grout wall from landward to seaward. This is not, of course, a sensible way of looking at the matter. The pressure in the soil between the grout wall and the sheet pile wall will actually operate in an opposite direction on the landward face of the sheet pile wall to that in which it will operate on the seaward face of the grout wall. But I do not doubt that Mr. Williamson was referring to the seaward faces of both walls.

Mr. Williamson’s second point was this:

Mr. Williamson’s third point was this:

As Mr. Bracegirdle explained, and as Mr. Reid showed in his calculation of the shear force that the grout wall had to resist, there is a net landward to seaward force on the grout wall at the bottom. Mr. Bracegirdle said that that would occur below -23.0mCD.

I accept that the results of the finite element analysis will inevitably be inaccurate. I accept that the reliability of the input data must affect, and can seriously affect, the reliability of the outcome. However, no other such analysis has been put before me, and it is the best evidence of many of the matters which are the subject of its outputs.

Bending moments to be expected in the sheet pile wall arising from soil pressure have been calculated by Mr. Bracegirdle using the ICFEP finite element analysis and by Mr. O’Brien. Mr. O’Brien’s method was, as he stated in his report,

The ICFEP analysis by which the initial state of the sheet pile wall in berth 205 (i.e. before the strengthening works were carried out) was calculated (“run 200”) showed a maximum bending moment of 760 kNm/m, which was at the relieving platform level, and a maximum deflection (distance from chord, assumed vertical) of 96mm. The maximum deflection, and the corresponding local maximum bending moment of 580 kNm/m, both occurred close to a depth of -10mCD. The value of (K0) used in the calculation was 1.5. Although the finite element analysis was criticized, Mr. O’Brien considered that apart from a number of factors that he mentioned, the calculation (or “base case” as he called it) was likely to produce a maximum bending moment in the sheet pile wall of something between 700 and 900 kNm/m.

The factors which Mr. O’Brien considered would lead to an increase in bending moment over the base case were these. First, the factor (K0). Second, that the stress relieving platform was not wholly effective. Third, that there had been over-dredging of the berth.

As to the factor (K0), it seems to me that Mr. O’Brien has conflated two matters. The first is that the (K0) value chosen for the Marsh Farm deposits of 1.5 was not necessarily realistic: it could be higher, and would certainly vary with depth (his report, paragraph 12.2(ii)). The second matter is that the driving of the H-piles would have increased the value of (K0) (ib., paragraph 12.2(vii)). But in my judgment there is an important difference between the two uses of the parameter (K0). That parameter is a characteristic of the soil, and represents the ratio of the horizontal stress to the vertical stress. In a large mass of soil, the greater the depth, the greater the pressure from the overlying soil, hence the greater the horizontal pressure for a given value of (K0), other things being equal. That pressure applies through the whole mass of the soil at the relevant depth. But the effect of driving the H-piles is different, since it is local to the H-piles. The H-piles bear the weight of the soil above the relieving platform and the surcharge loads on the ground at the top principally by reason of the frictional grip on them of the soil surrounding them. The driving of the H-piles in those circumstances is reflected not only in a vertical but also in a horizontal stress in the soil, especially in the layer of very dense sand, which lies typically between -10 and -11mCD. That horizontal stress is resisted by elastic forces in the soil which produce a horizontal pressure gradient so that the extra horizontal force produced by the H-piles becomes negligible after some distance from them.

What Mr. O’Brien said in paragraph 12.2 (vii) of his report is this:

Mr. O’Brien said in the course of his cross-examination by Mr. Bowdery (day 27, page 44) that he could provide some imprecise calculations of the stresses that would develop as the H-pile was driven into the ground. He was asked to do so. The calculations were duly produced. In those calculations, relying on a book by Jardine and Chow (ICP Design Methods for Driven Piles in Sands and Clays, 2005), Mr. O’Brien calculated that the force applied to the sheet piles was about 1.6 times to 5.3 times the capacity of the clutch. I think that those calculations must represent earlier calculations on which Mr. O’Brien based his report, since they were dated 16^th December 2005, after the end of the oral evidence.

I think it is fair to say that Mr. O’Brien was not very happy about those calculations. He had said (transcript, day 27, page 44):

Mr. O’Brien’s calculation shows the force said to be applied to the sheet pile wall by reason of the driving of the H-piles. It does include a calculation of the strength of the clutch, but it does not show the response of the sheet pile wall to the applied force. Until the berth was dredged to -12.8mCD, the sheet pile wall would have been supported by the soil on the passive side of it. It may be that a small movement of the wall would have relieved the net force on the wall. Whilst there may have been some difference between the actual situation and a situation where the H-piles were driven before the sheet piles were driven, no evidence as to the amount of any such effect has been offered.

The passages from Mr. O’Brien’s report quoted above suggest that the calculation is a calculation of permanent effects of the driving of the H-piles, not of transient effects, even though some of the causes appear to be transient. Part of the force calculated is a force existing below the level of the toe of the H-pile. That force at any given level of the sheet pile wall is transient, occurring only before or at the time when the toe of the H-pile passes the relevant level as it is driven. And in the calculation, the rate of hammering of the H-piles is a relevant factor. That again is clearly a transient cause. (Mr. O’Brien stated in his calculation that at the current state of the art it was not possible to carry out an appropriate dynamic calculation. He applied a factor of 1.2³ [equal to 1.73] to the calculated stresses to allow for the dynamic effect of the hammering. That figure was derived from what appear to have been empirical published figures).

I cannot accept Mr. O’Brien’s evidence that the effect would not have occurred if the H-piles had been installed before, rather than after, the installation of the sheet piles. It is of course true that the calculated forces would not act on the sheet pile wall if the sheet pile wall were not there. But the stresses calculated to have been in the soil were not dependent upon the presence of the sheet pile wall. If the sheet pile wall were driven subsequently, the forces each side of the wall resulting from those stresses in the soil would be substantially equal and opposite, and hence balanced. They would cause negligible relevant stress to the sheet pile wall. But once the dredging to -12.8mCD took place, the balancing passive force from the soil would be removed. The situation would not then be substantially dependent on the order in which the two sets of piles had been driven.

If I am wrong, and the effects that Mr. O’Brien was calculating were transient effects, then the same conclusion applies, since in that case the sheet pile wall would have been supported at the material time by the soil on the passive side. The stresses induced in the sheet pile wall by such transient forces have not been calculated, and there is no material before me to suggest that they might have been significant.

There is a further, related, consideration. The H-piles cause a local increase in the horizontal stress in the soil by reason of their displacement of the soil as they are driven. I accept evidence of Mr. Bracegirdle that that cause has an effect on the sheet pile wall which is negligible for present purposes. Whether that consideration is included in the Jardine and Chow method does not appear. But the conclusions expressed in paragraphs 103 and 104 above still apply.

Thus I am satisfied that this (K0) effect does not depend to any significant extent on the order of driving the H-piles and the sheet piles.

Another point pertaining to the fact that the H-piles were driven after the sheet piles was made by Mr. Reid. He said in his report, paragraph 29(3), that during the driving of the H-piles unwanted horizontal displacement of the sheet piles would occur. The effect would have been exacerbated by the system used to support the piling rig. One end of the piling rig was supported on the ends of the cantilevered sheet piles. Unfavourable forces and vibrations would thereby have been imparted to the sheet pile wall. He explained in oral evidence (day 22, page 102) that the vibrations caused by the hammer and by the horizontal movements of the piling rig would loosen the soil and reduce its capacity to resist the landward movement of the tops of the sheet piles. He said (report, paragraph 29(9)) that the combined influence of all those secondary effects was impossible to assess with accuracy, but it could be deduced that the first plastic hinge that could develop in the sheet pile wall after dredging would occur close to the river bed, rather than at relieving slab level as would otherwise be the case.

I need not consider that last deduction. As to the movement of the sheet pile wall caused by the movement of the massive piling frame and cranes, the forces generated would depend on the acceleration of those objects as well as their location. No figure for their acceleration, or indeed any calculation, was offered. I regard this evidence of Mr. Reid as speculative and I attach little weight to it. I am not satisfied that the order of driving the H-piles and the sheet piles had any significant effect on the state of the sheet pile wall before the grouting works were executed.

Mr. O’Brien considered that a change in the value of (K0) from 1.5 (as in Mr. Bracegirdle’s ICFEP run 200) to 2.5 (Mr. O’Brien said that the K value might be close to 2.5) would increase the maximum bending moment in the sheet piles by 10 to 25 per cent. He did not say how he arrived at the amount of that increase. Mr. Bracegirdle (ICFEP run 206) calculated the maximum bending moment on the basis of a value of (K0) equal to 2.4 at all strata as 850 kNm/m, an increase of 12 per cent. over his figure of 760 kNm/m for (K0)=1.5.

Mr. O’Brien’s second factor was that the stress relieving platform was only partially effective, because of the relatively stiff competent materials underlying it. He considered (paragraph 12.3(iii) of his report) that the top of the soil below platform level would bear 15 to 30 per cent. of the overburden at platform level.

I digress to consider the relieving platform.

The purpose of the relieving platform was to reduce the seaward stresses on the sheet pile wall imposed by the soil on its landward side. Those seaward stresses arose from the vertical pressure on the soil caused by the weight above it. Some of the weight of the soil and surcharge loads above the platform and resting on it was to be borne by the H-piles, which would transfer that load to lower levels.

Mr. Bracegirdle referred (report, section 19.3) to an assumption that the relieving platform would transfer all the soil self weight to the toe of the sheet pile wall. The expression “all the soil self weight” clearly refers to the loads on the platform that I have mentioned above. Mr. Bracegirdle said that that assumption was incorrect and unjustified. However, he did say (report, paragraph 6.11.7) that some small settlement of soil could be expected beneath the relieving platform. In that case, there would be no load on the top of the soil beneath the relieving platform. I thus interpret Mr. Bracegirdle’s denial of the assumption to his recognition of the fact that the H-piles would support the load above the relieving platform in part by friction in such a way as to increase the stress in the soil above the level of the toe of the sheet pile wall.

Mr. Reid estimated that the percentage of the load transferred directly to the soil underlying the platform was approximately 50 per cent. (report, paragraph 34). He reached that conclusion in part in reliance on his “bulking” theory. His bulking theory is to be distinguished from dilatancy, which is the volume increase of soil caused by shearing forces. The latter is well established; the former appears to be unique to Mr. Reid. As I understand it, the bulking theory asserts that when a load which has compressed soil is released, the soil loosens and expands. Thus when the ground is excavated for the purpose of casting the relieving platform, the soil beneath loosens and expands by heave. When the ground is restored above the platform, the soil immediately beneath the platform bears some of the load. Not all is transferred to the H-piles. In Mr. Reid’s own words (report, paragraphs 34 to 38):

Mr. Reid explained the matter further in evidence in chief (day 17, pages 36, 37):

Although Mr.Reid is reported as having said “wall”, I think he must have said, or intended to say, “slab”, the point that he was answering being that if the soil did not bulk but filled the bow of the sheet pile wall, the level of soil under the relieving platform would fall.

Mr. Reid calculated the percentage of the load on the relieving platform carried by the H-piles in this way. Assuming a certain percentage load carried by the H-pile, he calculated the consequent shortening of the pile. That would involve the lowering of its top. He then computed the heave that would generate the assumed percentage. He used assumed characteristics of the heaved soil to determine the proportion of the load that would bear directly upon it and the proportion that would be supported by the H-pile. Mr. Reid said (day 21, page 128) that in his view the soil would move back down, upon loading of the relieving platform, only by the amount it had heaved when the soil had been excavated. Having assessed the heave, by other means, at 2mm, he then arrived at his figure of 50 per cent. I take it that he used the results of the research he mentioned in order to assess the heave. He said (day 21, page 128) that it was impossible to calculate the heave. However, it appears from the appendix to his report, sheet 17, that the heave that according to his calculation produces the 50 per cent. division is not 2mm but 4.444mm. 2.222mm is the shortening in the length of the H-pile.

I am far from saying that bulking cannot happen because nobody has written a paper on it. But Mr. Reid’s assessment of 2mm heave is not supported by evidence that has been put before me, and appears to be inconsistent with his own calculation of 50 per cent.

Mr. O’Brien did not adopt the bulking theory. He put the matter this way (report, paragraph 12.3(iii)):

Thus Mr. Bracegirdle considered that the relieving platform imposed no load directly on the soil beneath it, but accepted that there would be an effect on the sheet pile wall by reason of the shaft resistance of the H-piles lower down. Mr. Reid and Mr. O’Brien considered that both effects would exist.

I am concerned with the magnitude of the effect, whatever the mechanism of its cause. I revert to Mr. O’Brien’s evidence.

In table 12.1 in paragraph 12.3(iv) of his report he postulated an increase in the stress borne by the top of the soil below platform level from 0 (the base case) to 40kN/m². The latter figure represents approximately 30 per cent. of the load on the relieving platform (130kN/m²: see paragraph 119 above). He considered that that increase would cause an increase in bending moment in the sheet pile wall of 5 to 15 per cent. He did not say how he arrived at those figures. Mr. Bracegirdle’s figure of 760kNm/m was based on the assumption that there was no stress on the top of the soil below the relieving platform. He calculated that if the whole of the load on the relieving platform were borne by the soil beneath it, the maximum bending moment would be increased by 45 per cent. to 1100kNm/m. On the assumption of proportionality, which is the best one can do, that implies that the maximum bending moment if the relieving platform allowed 15 to 30 per cent of the load on it to be transferred to the top of the soil beneath it would be increased from 760kNm/m by 6¾ per cent. to 13½ per cent.

Mr. O’Brien considered that the effect of over-dredging from -13.1mCD to -13.7mCD would be to increase the maximum bending moment by 5 to 10 per cent. He did not say how he arrived at those figures. Mr. Bracegirdle, using run 220 for some reason that is not apparent, calculated that over-dredging from -12.8mCD to -14.1mCD would increase the maximum bending moment from 727kNm/m (cf. run 200: 760kNm/m) to 750 kNm/m, an increase of 3 per cent.

No calculation has been done of the combined effect of all these changes. The best that can be done to give a rough estimate is to treat them as mutually independent. Applying Mr. Bracegirdle’s estimates in that way, the combined effect would be to multiply the figure of 760kNm/m by 1.12 X 1.0675 X 1.03 (=1.23) to 1.12 X 1.135 X 1.03 (=1.31) giving a maximum bending moment of 936 to 995kNm/m. Mr. O’Brien’s estimate of the combined effect is 1.2 to 1.5. Thus there is not a wide degree of dispute between Mr. Bracegirdle and Mr. O’Brien on the combined effect of these matters, on the assumption that Mr. O’Brien is right as to the change in the factors mentioned in paragraph 96 above.

Taking the maximum figure of Mr. O’Brien’s base case, 900kNm/m, and the figure of 2100kNm/m as the bending moment capacity of the sheet pile wall, 43 per cent. of the structural capacity of the sheet pile wall would have been utilized. That accords with his opinion that “one would normally expect the applied bending moments to be a low percentage of the wall capacity, with perhaps less than 50% of the structural capacity being utilized” (his report, conclusion of section 12.3). He concluded that the changes in parameter which he considered appropriate, and which I have considered above, could have increased the utilization factor to about 80% to 90%. Those figures represent increase factors of 1.87 to 2.1 from a base of 900kNm/m, and of 2.4 to 2.7 from a base of 700kNm/m. He concluded that the maximum bending moment moves perilously close to the bending capacity of the sheet pile wall, giving a far lower margin of safety than would normally be expected. But those factors go well beyond the rest of his evidence. If one takes his maximum estimate of the base case, namely 900kNm/m, and his maximum estimate of the uplift factor, namely 1.5, one arrives at a utilization factor of 64 per cent., giving a margin of safety of 55 per cent.

I accept Mr. Bracegirdle’s evidence as to the individual factors representing the effect of the changes in parameters proposed by Mr. O’Brien. That is to say, I accept that they correctly reproduce the effects of the changes as calculated in accordance with the ICFEP analysis. Mr. O’Brien himself said that ICFEP was a highly respected software package (section 12.2 of his report). It was the assumptions and idealizations that he criticized. In so far as the assumptions did not meet Mr. O’Brien’s criteria and gave rise to his suggested increases in bending moment, they have been amended in the manner indicated above. There remain criticisms, including the fact that the analysis is a 2-dimensional analysis of a 3-dimensional problem. But in my judgment, those criticisms are likely to be of lesser force when applied to adjustments than when applied to the original calculations. Thus in my judgment Mr. Bracegirdle’s factors of 1.23 to 1.31 are more reliable than Mr. O’Brien’s effective factors of 1.87 to 2.7. Mr. O’Brien does not differ from Mr. Bracegirdle in his base case, though he gives a bracket. If one takes Mr. O’Brien’s median base case of 800kNm/m and applies a factor of 1.3 to it, one arrives at a figure of 1040kNm/m as the maximum calculated bending moment. That gives a factor of safety of 2 in relation to the formation of plastic hinges.

There will be variations along the wall since, as Mr. O’Brien pointed out, there are variations in soil strata along the wall. However, the bending moments calculated from consideration of soil pressure are consistent with the observed curvatures, except in the immediate vicinities of X1, X2 and X3.

The defendants contend that the reason why the sheet pile wall came to bulge excessively when the grouting process took place was that there were existing plastic hinges in the sheet pile wall. That proposition was somewhat widened to the proposition that the sheet pile wall was in a state of excessive and unforeseeable bending before the grouting took place. The possible effect of the grouting is thus relevant to determination of the initial state of the sheet pile wall. Before considering that question, I shall give here a fuller description of the grouting process than I have given above. The following is a description given by Dr. Greenwood.

Mr. Bracegirdle’s ICFEP analysis (Run 200) showed that in the absence of grout loading the sheet pile wall, assumed to have been driven vertically, would have an offset of 96mm from the vertical at a depth of about -10mCD. The maximum bending moment in that vicinity would be 580 kNm/m at approximately the same depth. There would also be a maximum bending moment of about 760kNm/m in the opposite direction at the level of the relieving platform.

When the grouting process was working as intended, the high pressure causing the high velocity of the jet of grout was dissipated as the jet was absorbed in the surrounding soil. The pressure that continued in the soil while the grout was liquid was simply the hydrostatic pressure of the grout, plus the pressure difference required to drive the excess grout up the annular tube back to ground level. If the grout return was lost because the annular tube was blocked, then the high pressure of the jet would extend to the liquid grout in the soil for so long (if at all) as the jetting process was continued. The grout return could also be lost because grout was finding its way to contact with the sheet pile wall, forming a lens behind the sheet pile wall and leaving no excess to pass up the annular tube. I derive the foregoing propositions from the evidence of Dr. Greenwood.

I have been unable to reconcile two pairs of different statements made by Mr. Bracegirdle in relation to the finite element analysis of the effect of grout pressure on the wall. The differences were not put to him. The first was this. In section 12.8 of his report he said:

I understood that passage to mean that the assumption underlying the calculation was that half the area of the sheet pile wall was covered by the liquid grout. (The calculation was a two-dimensional calculation, i.e. it considered a small length of the wall in isolation, or possibly with defined restraints). But in the course of cross-examination by Mr. Williamson (Day 10, page 111) Mr. Bracegirdle said:

The other difference is this. Figure 72, mentioned above, shows the pressure of the liquid grout as being substantially greater at all depths than the active soil pressure in the absence of the grout. Indeed, the passage from Mr. Bracegirdle’s report mentioned above continues

I cannot resolve those apparent discrepancies. However, I conclude, in particular from what Dr. Kovacevic is recorded as having said (transcript of the meeting, pages 17 to 19), the following. At increment 45 the system (including the state of the sheet pile wall) had been modelled on the assumption of no grout pressure. Dr. Kovacevic established the difference between the total horizontal stress which was acting at each level of the wall at increment 45 and the hydrostatic pressure due to liquid grout at the same level. He then applied, using the program, successive increments of 10 per cent. of that difference, assumed to be acting on the wall wherever the grout pressure exceeded the calculated horizontal soil pressure. That showed increasing bending moments in the sheet pile wall and corresponding increasing bulging in that wall.

The following table, derived from graphs contained in Mr. Bracegirdle’s report, shows the maximum bending moments and displacements of the bulge from the vertical calculated by him using ICFEP on the assumption that the piles were driven vertically.

Mr. Bracegirdle said that in the event of hydrofracture the comparison of stresses would be considerably more onerous. He said that loss of slurry return to the surface would cause the immediate outward displacement of the sheet piles and hydrofracture. I accept that if loss of slurry return through blockage of the annular cylinder occurred the grout could apply pressure to the wall far greater than the hydrostatic pressure. Whether it would produce an effect on the wall greater than that calculated by the analysis depends on the extent of the wall subjected to the pressure. The analysis itself assumes a substantial coverage by grout at hydrostatic pressure. But whether or not the process was hydrofracture strictly so called is not of significance. Dr. Greenwood considered that a more likely process was that the hydrostatic pressure of the grout could be transferred either directly or through a few centimetres of sand or gravel to deform the wall. He wrote in his report (paragraph D33):

Mr. Bracegirdle said, and I accept, that an exact analysis of the effect of liquid grout on the sheet pile wall was not possible. One reason he gave, which I accept, was that the restraining effect of the neighbouring sheet piles was highly uncertain. He said that the calculations nevertheless showed the sensitivity of the sheet pile wall to grout pressure.

Dr. Greenwood said this in his report:

I accept that evidence except in so far as it expresses a limit on the area of the sheet pile wall which could be covered by the liquid grout. It is apparent that the grout could spread widely behind the sheet pile wall. There is uncontroverted evidence that grout was found no less than 16 metres from the edge of the works. Mr. Bracegirdle said that the grout could form a lens behind the wall which would expand as it filled with more grout. Mr. O’Brien said this in the course of his cross-examination by Mr. Bowdery:

I find the pressure head required to drive the excess grout, or soil/grout fluid, up the annular tube to have been about half a metre when the grouting was being carried out in a single phase, and about one metre when it was being carried out in two phases. Thus in normal operation that extra pressure head would have to be added to the relevant depth to obtain the pressure of the grout while it remained liquid. I obtain the figures of one-half and one metre as follows. The pressure head h is given by

Dr. Greenwood said that his figure A5-1 showed that the viscosity was very sensitive to change in shear rate at low values of that rate. I do not accept that his figure A5-1 shows that, though the scale is so small as to make estimation difficult.

I find the pressure head required to drive the fluid through the broader section to be negligible. All these conclusions are based on the assumption that the annular cylinders are not obstructed.

Dr. Greenwood said that with flowing suspensions of particles tending to pastes we do not know how the velocity difference is distributed across the section as viscous fluid flow transforms into plug flow with reducing liquid phase. The calculations do not account for any difference in resistance between viscous and plug type flows. I accept that evidence. Whilst the figures of ½ metre and 1 metre are in my judgment the best available estimates, they cannot be regarded as reliably accurate.

Increments 45 to 50 assume a certain percentage of the excess of the hydrostatic pressure over the geostatic pressure applied over the whole wall. In fact, on the above evidence, the whole of the excess could act, and could well act over a height of about 12 metres and a comparable width. While the effluent was flowing up the annular cylinder, the pressure would be greater than the hydrostatic pressure.

A device known as a Lutz registration device was used by HSS to monitor the installation of the grout columns. According to the unchallenged evidence of Mr. Nico Claessens, who acted as works manager for the installation of the columns, and whose evidence was admitted under the Civil Evidence Act, that device registered all relevant grouting parameters, namely depth, penetration/lifting speed, rotation speed, grout pressure and grout flow. The device showed the data in a digital form and also produced a paper record. It kept a continuous record except on occasions when it malfunctioned; on those occasions, operatives checked the relevant gauges, manometer and flow meter. A visual inspection of the grout return was continuously undertaken. If loss of grout return was experienced, the VHP installation was stopped. Reduced flow rates and pressures were applied, and the rods were moved up and down irregularly, a procedure known as “dancing”. The problem was overcome by gradually building up pressure and flow to the pre-determined rates.

Mr. van der Eecken, who gave evidence before me, was Product Manager for HSS with responsibility for quay wall projects. He said in his witness statement that when he attended on site, he observed the works including the registration of the data by the Lutz device. He also reviewed the records for the days when he was not on site in order to ensure that relevant grouting parameters and work procedures had been maintained.

A site diary of Vanessa Bellardinelli was admitted in evidence under the Civil Evidence Act. Ms. Bellardinelli was a geotechnical engineer, qualified in Italy as a civil engineer. She was site engineer for the quay wall strengthening works. Her site diary entry for 25^th June 2002 stated:

On 13^th July 2002 Mr. van der Eecken sent an e-mail to Marc Voorhuis, Bob Michta and Eric le Moine, with a copy to his manager Lucas Bols, asking them to take certain action and foreshadowing two alternative proposals of action to deal with the problem, the second of which was in the event adopted. The relevant passage for present purposes is this:

Earlier (day 3, page 112) Mr. van der Eecken’s e-mail of 13^th July 2002 was put to him. The following exchange took place (day 3, page 113):

In spite of Mr. van der Eecken’s subsequent denial quoted above (day 3, page 127) I am satisfied that there were, and continued to be, occasions of blockage of the effluent by gravel. I have not been shown, let alone offered an explanation, of the Lutz data claimed to be relevant. Mr. Bracegirdle’s statement was correct.

Whether blockage was partial or complete, while the jetting process continued the sheet pile wall would have been subjected to higher than hydrostatic pressures of grout wherever (if at all at the material time) liquid grout was in contact with the back of the wall and was liquid continuously between there and the jetting monitor. Partial blockage could raise the excess of pressure over hydrostatic substantially above the extra head of 1 metre of grout that I have mentioned above. Total blockage could raise it to VHP level. To avoid that, the operatives would have had to react quickly to loss of grout return caused by obstructions in the annulus.

The higher the pressure of the grout on the back of the sheet pile wall, the less the area of wall that would have to be covered by liquid grout to cause plastic hinges to form in the wall, other things being equal. The amount of wall that would have to be exposed to the grout pressure would also depend on the location of the exposure. It would also depend upon the magnitude of the pre-existing bending moments in the wall. The greater that magnitude, the less the area required, again, other things being equal.

Whilst I am not satisfied that there were occasions when the full pressure (i.e.VHP) of the grouting process was applied to the back of the sheet pile wall, I am well satisfied that there were numerous occasions when the hydrostatic pressure was applied over substantial areas.

The curvature of the sheet piles before the works were carried out was greater in some parts of the wall than others. Yet the incidence of the damage that occurred was widespread, not confined to limited areas along the length of the wall. No correlation between the locations of the damage and the places of higher curvature has been suggested.

I should answer a particular rhetorical question raised by Mr. Williamson. It was this. The finite element analysis was said to support a conclusion that merely 50 per cent. of the grout loading was sufficient to cause an immediate 1 metre displacement of the wall. That conclusion was in no way qualified. If 50 per cent. of the grout loading instantaneously could cause such a massive displacement (1 metre), why did not 100 per cent. of the grout loading, applied repeatedly and over long periods, cause the wall to fail, or, indeed, cause anything like this displacement? My first point in answer is that the figure of 50 per cent. relates to the excess over the geostatic pressure, and is more than 50 per cent. of the hydrostatic pressure of the grout. However, that does not meet the gravamen of the question. The second point is that Mr. Bracegirdle made it clear that the analysis had been carried out in two dimensions (his report, appendix 16, first paragraph). That was a relevant qualification. The question of the extent to which support to the section of the wall in question was afforded by neighbouring sections was not addressed. Third, 100 per cent. of the grout loading was not necessarily ever applied to the wall: the precise area of the wall covered does not appear. Finally, 1 metre displacement and more did occur at many places, as the Netsurvey of 2004 clearly shows. Mr. Williamson submitted that Mr. Bracegirdle sought to evade the question, and he suggested to Mr. Bracegirdle that he was fencing. On the contrary, I have the impression that Mr. Bracegirdle was doing his best to answer the relevant questions and was not fencing.

I am satisfied that the normal grouting process was perfectly capable of acting on the sheet pile wall if the latter had an existing bending moment of about 760kNm/m in such a way as to cause it to form plastic hinges. A priori, the grouting process would of course have been more likely to make the wall bulge out if the wall had already been in a state where it had, or was close to having, plastic hinges. The greater the initial bending moment, the less the increment required to reach plasticity.

The jet grouting process for strengthening quay walls is a novel process. It was, however, used four times before it was used at Southampton. Haecon have mentioned the following designs which include deepening and strengthening of steel sheet pile quay walls by use of VHP columns and ground anchors:

Since berths 204 and 205 must be taken to have been in the same condition before the works were carried out, I conclude, on the whole of the evidence, that berth 204 is and was at the material time in the condition described in paragraph 81 above. That is to say, I reach the same conclusion as is expressed provisionally in that paragraph. The experts’ agreement item 17 that the state of berth 205 before the grouting was carried out was the same as that of berth 204 must mean that the two states were comparable in this sense, that the states of bending of the two quays were similar. Thus the reference in paragraph 81 to X1, X2 and X3 must be read, in relation to berth 205, as a reference to three unidentified lengths of about 20m each.

I am also satisfied that the state of the sheet pile wall in berth 205 in the locations corresponding to the vicinities of X1, X2 and X3 would, standing alone, be wholly insufficient to account for the widespread cracking and bulging that took place during the grouting process.

Amended clause 11 of the contract provides as follows:

The defendants contended that the sheet pile wall was, before the grouting works were carried out, overstressed in bending by reason of three matters: (1) That the H-piles were close behind the sheet pile wall; (2) that the H-piles had been driven after the sheet pile wall had been driven; and (3) that the relieving platform was not 100 per cent. effective. They contended that such overstress was not reasonably foreseeable by an experienced contractor.

As to point (1), it is not suggested that the defendants were unaware of the closeness at the time of tendering. The tender documents included a drawing which showed the relevant dimensions. Mr. O’Brien gave evidence (day 26, pages 157 to 162) that the closeness of the H-piles to the sheet pile wall would give rise to extra stresses in the sheet pile wall that were, at the time of tendering and indeed now, incalculable. The combination of closeness and very strong ground was unusual. He accepted that a contractor would know that extra stresses would be imposed.

As to point (2), URS Corporation Limited (“URS”) was instructed by HSS to provide an opinion on the cause of cracking of the sheet pile wall. In its report, URS stated that the only records describing the construction of the berth were brief descriptions contained in two articles published in the July 1971 and November 1972 editions of the Dock and Harbour Authority. URS appear to have deduced from the information contained in those articles an “envisaged” construction sequence that the H-piles were driven after the sheet pile wall was driven.

The archive of ABP contains drawings 6573/P2E and P11A of the contractor Keir which show details of the piling frame for the relieving platform bearing piles (i.e., the H-piles) resting on the sheet piles and the landward rail beam for berth 201. Thus it is clear from that drawing that in the case of berth 201 the sheet piles must have been driven before the H-piles were driven. A similar arrangement was in fact used for the construction of berths 202 to 205. If Haecon and HSS had not already been on enquiry, that archive drawing, if they had seen it, would have put them on enquiry whether berth 205 had been constructed in the same way.

If HSS or Haecon had had any reason to suppose, or had supposed, at the time of design or tendering that the order of driving the sheet piles and the H-piles was relevant, they could perfectly well have asked ABP. There is no evidence that they did so.

The combined effect of points (1) and (2) appears to have been included in Mr. O’Brien’s figure of 10 to 25 per cent. as his estimation of the increase in bending moment in the sheet pile wall caused by his estimated increase in the value of (K0) from 1.5 to 2.5. I do not regard such an increase as beyond the foresight of an experienced contractor.

Point (3) is that the relieving platform was unforeseeably not 100 per cent. effective. It is implicit in my finding in paragraph 126 above that I accept that the relieving platform was not 100 per cent. effective, that is, that it did not relieve the soil at levels above the toe of the sheet pile wall of the whole load on the platform. Mr Reid said this about the relieving platform in his report, paragraphs 32 and 36:

I conclude from that that there is nothing unforeseeable about a relieving platform not being 100 per cent effective. Nor was there anything unknown to the defendants about the dimensions of the system. The tender documents included a drawing which was put to Mr. O’Brien (day 26, page 168):

I am not satisfied that there was anything unusual or unforeseeable about the relieving platform or about the degree of its effectiveness.

In paragraph 13 of Haecon’s defence to the part 20 claim, it is pleaded that Haecon undertook a structural analysis of the existing quay wall based on voluminous information which was listed in paragraph 14. Mr. Bowdery put to Mr. O’Brien paragraph 15, which said:

Mr. O’Brien gave this evidence (day 26, page 171):

In section 8.6 of his report, Mr O’ Brien said that in order to carry out sophisticated analysis of the interaction between the soil and the structure, consistent with the complexity of the structure, it would be necessary to have appropriate input parameters for the analysis, such as (in the case of the soil) the degree of non-linearity between small strains (0.01 %) and large strains (1.0%), the degree of anisotropy of the stiffness and of the strength, and the relationship between strength and mean effective stress. For the Bracklesham Beds, which are not well researched relative to London Clays, it would be particularly important to have such data if geotechnical analysis was to be reliable for design purposes. The relevance and importance of that kind of information is now well known, and Mr. O’Brien cited two papers, one published in 1989 and the other in 1992. I conclude that the relevance and importance of that kind of information was well known at the time of issue of the tender.

I am satisfied that Haecon was right in saying that it had sufficient information on the properties of the soil for the purpose of designing the strengthening works. But it went on to plead (ib., paragraphs 20, 21 and 22):

Mr. Gary Brown was the Engineering Manager at the Port of Southampton for ABP. He gave evidence (day 2, page 120) which I accept that at a meeting he had with the pre-tenderers in July 2001 he said that ABP had significant archive information.

I am satisfied on the evidence of Mr. Van der Eecken that HSS was not refused any information that it asked for from documents in the archive. I accept his evidence that there was no restraint on what HSS could investigate as part of the site investigation. But none of the investigations that HSS performed, he said, gave any indication of a need to carry out any more detailed investigation. The following exchange took place during the course of his cross-examination by Mr. Bowdery:

The designer, Mr. Joep Wijffels, gave evidence before me. He explained that when the tender was submitted the design was a concept design, not a detailed design. He said that the design period (by which he must have meant the period during which the concept design was prepared) stopped at the moment when the contract was signed. During that period, he said, he was not refused access to any information. Haecon obtained the information it needed from HSS and from the tender documents.

In his witness statement, Mr. Wijffels said

Mr. Reid said in his report at paragraph 46:

Section 101-07A of the specification of the contract contained the following provision:

Mr. Williamson submitted that a survey which would have been sufficient to show the condition of berth 205 at the time of tendering would not ordinarily have been undertaken by an experienced contractor acting reasonably without some indication or evidence that the wall was suffering from any distress. The cost of determining the profiles of the wall as has now been done would have been far more than was envisaged by ABP itself. ABP had allowed a figure of only £5000 for such a survey.

HSS’s method statement for survey of the quay wall provided for a photographic survey at the surface and for ultrasonic measurements to establish thicknesses of steel, and hence to establish the existence and extent of any corrosion.

I accept evidence of Mr. Brown that at a pre-tender meeting of tenderers he told the tenderers that they would need to conduct their own sheet pile wall survey and that there would be a contribution by ABP of £5000 for that survey. That figure was determined by ABP.

Mr. Wijffels said in his witness statement (paragraph 34) that he understood the expression “Condition survey” to mean a corrosion survey. That understanding was, he believed, consistent with the amount of money allocated for surveys, the time available and the commercial reality that contractors are most unlikely to incur the cost of surveys prior to acceptance of their tender.

In my judgment, it was for the tenderers to decide whether and in what way to survey the wall. If they considered such a survey prohibitively expensive or that the prospect of the work was too risky, they would not tender. Otherwise, they would allow an appropriate amount in their tender.

As I have found, most of the wall was not suffering distress; and I am not satisfied that any of it was.

My conclusion as to the initial state of the quay wall of berth 205 (see paragraphs 158 and 159) means that the sheet pile wall was not in the condition complained of. Nor were the matters upon which the defendants rely as causing the alleged condition unforeseeable by an experienced contractor. Thus this claim fails.

The relevant terms of the contract are:

Mr. Jones submitted that the fitness for purpose obligation must be read together with and/or subservient to clause 12 (Footnote: ¹²) of the contract conditions. Thus the interpretation to be given to “fitness for purpose” in order to give business efficacy to the agreement was that it must take account of the relevant information that was not in the possession or knowledge of Haecon. I treat this submission as also a submission of HSS, where the lack of information in the possession of HSS would be relevant. All that was necessary, submitted Mr. Jones, was for Haecon to raise the assertion that evidence of the physical condition of the quay wall was not foreseen by it or by HSS. Thereafter, ABP must prove that such information was either irrelevant, made available or that it was foreseeable. Mr. Jones submitted that to give any other interpretation to the contractual term would fail to give it business efficacy since it would be analogous to a warranty in respect of latent defects, and in particular the works undertaken in the 1970s, and there would otherwise be an untenable tension between the fitness for purpose obligation and the clause 12 condition.

I reject Mr. Jones’s submission. The contract term is perfectly clear that the works have to be fit for the purpose. Clause 12 enables the contractor to claim extra payment in certain circumstances but does not relieve him from the obligation, or modify the obligation, as to the fitness for purpose of the works. The burden remains on the contractor to prove his clause 12 claim, not on ABP. The burden does, of course, lie upon ABP to prove that the works are not fit for the purpose.

Mr. Reid carried out a number of calculations relevant to the question whether the grout wall was fit for its purpose. He wrote in his report that the calculations had been carried out by WMR Consultancy to demonstrate that the wall system as designed by Haecon at Berth 205 was adequate to carry the loading that would have acted on it provided the relieving slab and the supporting piles [i.e. the H-piles] were capable of shielding the wall from overburden pressure. The calculations were based on the assumption that the relieving slab and the support H-piles transmitted the vertical load of the overburden and superimposed load above it to a level beneath where the sheet pile wall would be subject to lateral forces from that vertical loading.

The first set of calculations considered the stability of the sheet pile wall under the hydrostatic pressure of liquid grout. Of that set, the first subset related to the case where the whole grout column was executed in one pass. The second represented the upper stage of the two-pass system, and the third, the lower stage of that system. In each such subset of calculations it was assumed that the grout pressure acted over the whole relevant height of the wall. In the first subset, the relevant height of the wall was from 0mCD to -14.8mCD; in the second subset, from 0mCD to -7mCD; and in the third subset, from -6mCD to -14.8mCD. It was assumed that the sheet pile wall was fixed at and below -14.8mCD. It was also assumed that the grout pressure acted on the sheet pile wall over the diameter (0.9m) of one grout column but was resisted by a width of wall equivalent to three such diameters. The maximum bending moments in the sheet pile wall in all cases appeared at relieving platform level, and came to 1800kNm, 1100kNm and 1480kNm respectively. Those figures are equivalent to 667kNm/m, 404kNm/m and 548kNm/m respectively. The calculations also showed shear forces and deflections. The deflections were only a few millimetres. Mr. Reid concluded that on those assumptions the sheet pile wall would have had a significant reserve of strength amounting to about 100 per cent., i.e. a factor of safety of 2.

The second and third sets of calculations, which related to serviceability loading of the sheet pile wall alone, I can pass over. The fourth set related to ultimate loading of the sheet pile wall alone, dredged to -16.5mCD. That set contained two cases, corresponding to subclauses (a) and (b) of BS 8002:1994 clause 3.2.5. That clause provided that a retaining wall should be designed to be in equilibrium mobilizing a soil strength of the lesser of two quantities, the one defined in subclause (a), and the other defined in subclause (b). Those calculations showed that although the sheet piles could in both cases carry the applied loading, the embedment in front of the wall was deficient by 270kN in case (a), and by 315kN in case (b). Those calculations related to a width of wall of 1.7m. The conclusion was drawn that each grout column would be required to deliver an ultimate shear resistance of 315kN. That was based on the design spacing of 1.7m. The worst case was a spacing of 2.7m, requiring a shear resistance of 500kN.

The conclusion expressed in the report included the following:

A final calculation showed that the H-piles could themselves each provide a shear resistance of 506kN. Mr. Reid explained in the text of his report (section D, paragraph 37):

The conclusions expressed by Mr. Reid at the end of his calculations included these:

It is implicit in Mr. Reid’s calculations and conclusions that, if the sheet pile wall were in the condition which the defendants were contending was the condition to have been expected when the strengthening works were designed, the strengthening works were unnecessary. The H-piles would have done the job. There is no doubt that he regarded the grout wall and the H-piles as acting mutually independently. He said in evidence in chief (day 17, page 52):

In my judgment, the H-piles ought not to be relied on to afford lateral support. The cranes run on rails and cannot operate if the rails become out of gauge. Mr. Bracegirdle made the point (day 8, page 28):

Mr. Bracegirdle, using the ICFEP finite element analysis, calculated that the grout wall would be required to withstand a shearing force of 200kN/m at -23mCD. (The relevant graph actually showed a figure of 225kN/m). That calculation was criticized. Mr. Reid calculated that each grout column would be required to deliver an ultimate shear resistance of 315kN. That conclusion appeared at page 13 of the appendix to his expert’s report. It was based on the design spacing between the soilcrete columns of 1.7m, and amounted to 185kN/m. Professor Pavlovic said (day 16, page 26), and I accept, that that was very close to the figure calculated by Haecon. Mr. Reid stated that the figure of 315kN had to be increased in line with actual spacing for the worst case of an average spacing of 2.7m. That led to a required shear resistance of 500kN per column. Mr. Reid expressed the conclusion that collapse as a result of shear rupture was not possible in view of the very high dowel action available from the reinforcing tubes. A pair of reinforcing tubes had a shear resistance of 525kN, according to his calculation and his oral evidence in chief (day 17, pages 54 to 57). As he explained, for that shear resistance to be mobilized, the soilcrete itself would have had to suffer a brittle failure, with lateral displacement of the parts of the soilcrete column above and below the failure surface. It is clear from page 14 of the same appendix that the amount of that displacement would be about 157mm. Mr. Reid did not say at that point whether or not the grout column had sufficient shear capacity apart from the dowel effect mentioned above.

In my judgment, the wall cannot have been fit for its purpose if in the ultimate limit state the wall can be saved from collapse only by the rupture of the soilcrete columns. BS8110-1:1997 provides by section 2.2.2.1 that a design should satisfy the requirement that that no ultimate limit state is reached by rupture of any section under the worst combination of ultimate loads. That in my judgment is also a requirement of fitness for purpose of the grout wall.

In response to a pre-trial question for clarification of his report, Mr. Reid wrote

Using equation 6a, the calculated value of the shear capacity increases linearly with the axial compression. For zero axial compression, the shear strength calculated in the above manner using equation 6a would amount to 396kN.

Mr. Reid produced a diagram in which he asserted that downward frictional forces, mobilized by the cracking of a column, would act on the column. I reject that evidence. It was put to Professor Pavlovic in cross-examination (day 6, page 168) that he had not taken account of those forces in his calculations. He said:

Further, Mr. Reid’s own diagram suggested that those forces would balance out across the width of the column.

It was implicit in Mr. Reid’s calculation that the whole weight of the grout pile above the level of -20.7mCD was borne by the soil below that level. That is also implicit in Professor Pavlovic’s perhaps reluctant acceptance of Mr. Bracegirdle’s opinion that upward frictional forces would not be mobilized.

It follows from my rejection of the proposition that downward frictional forces would be mobilized that I reject Mr. Reid’s addition of 149kN to the axial load represented by the weight of the pile. If the axial load is simply 198kN, then Mr. Reid’s calculation of the shear strength of the grout pile would come to 483kN. A more precise calculation using Mr. Reid’s dimensional data would, I think, lead to a figure of 506kN.

A fallacy underlying Mr. Reid’s calculation is that he has ignored the fact that the British Standard relates to concrete. The British Standard in question, BS 8110-1:1997 provides by clause 3.1.7.2 for a minimum strength of concrete. The actual figure does not appear. In Mr. Reid’s calculation, table 3.8 in the standard is used which suggests that a characteristic concrete strength of not less than 25N/mm² is envisaged. Mr. Reid accepted (day 18, page 31) that structural concrete is not expected to have a lower value than that. If the characteristic strength of the soilcrete were taken as 12N/mm², the effect would be to reduce a constant term of 0.45 N/mm² in Mr. Reid’s calculation by a factor of 0.783 to 0.35N/mm². That would cause a reduction of the values of the shear strengths of the column calculated by Mr. Reid’s method by about 86kN.

Mr. Reid said in cross-examination (day 18, page 33):

Professor Pavlovic said that he used equation 6b, but he did not think much of the standards and he did his own calculation. What he said in cross-examination was this (day 6, pages 160 to 165):

In equation 6a, a factor Vh/M appears. Note 2 states that the value of Vh/M should be taken as not greater than 1. Mr. Reid has, without explanation or challenge, taken Vh/M equal to 1. If one takes that value, it is manifest that equation 6b must give a lesser value than equation 6a if, as assumed here, the column is not in tension.

Professor Pavlovic set out two graphs in his report showing shear strength per column as a function of axial load. Each graph showed three curves representing different safety factors; one graph related to EC2 and the other related to BS 8110. The latter showed lower shear strengths than the graph for EC2, and was thus, to use his language, more “punitive”. He also set out his calculations. They were all apparently based on the formulae in the standards. The other calculations that he said he performed were not put before the court, nor were their results. As to BS 8110, Professor Pavlovic purported to use equation 6b, but I have been unable to see how he arrived at his result. In EC2, the equations corresponding to equations 6a and 6b are labelled 6.2a and 6.2b respectively. It was 6.2a, not 6.2b, that Professor Pavlovic used. He set out his calculation based on that equation. On the basis of the dimensional data that he used, which were not controversial, I accept that his graph of shear strength as a function of axial load, with no safety factors, is a correct representation of equation 6.2a. With no safety factor, the shear capacity of a column at zero axial load comes out at 474kN. At an axial load of 200kN, it comes out at 504kN. That is close to the figure of 506kN mentioned in paragraph 205 above.

Haecon, in their design, used a figure of 560kN as the axial load on a single column. For the purpose of designing columns strong enough to sustain that compression, that was a proper figure to take. But it cannot be used to calculate the shear strength of a column. Professor Pavlovic said this in cross-examination (day 7, page 52):

On self-weight of the columns as axial load, Mr. Reid said this (day 17, pages 66, 68):

The point made by Mr. Reid was this. While the soilcrete was fluid, there would be a hydrostatic pressure at any given level by reason of the weight of fluid above that level. I accept that. When the soilcrete solidified, that pressure would remain because he could see nothing that would relieve it. I am not satisfied that he is wrong about that. When the soilcrete was fluid, the whole weight of a column of fluid vertically overlying the fluid at a particular level would be borne by the fluid at that level. The weight of any of the overlying fluid in the interstices of the soil at the side surface of the column (where the surface “fluctuates in and out”) would be borne, at least in part, by the soil underlying the fluid in the interstice. Whether the whole, or only a part, of the weight of the fluid in the interstice would be borne by the soil depends upon the shape of the interstice. When the fluid solidifies it thereby becomes capable of bearing shear stresses. So the weight of the soilcrete right across any given level could be borne by the soil that has been supporting the fluid in the interstice. That would relieve pressure on the soilcrete below that level. The elastic forces thereby mobilized would involve some movement of the soilcrete relative to the surrounding soil, notwithstanding that, on the evidence of Mr. Reid (day 17, page 67), there is no change in volume of the soilcrete on solidifying. There is no evidence of such movement. Professor Pavlovic, reluctantly I think, accepted that proposition. Consequently, I accept Mr. Reid’s evidence that the self-weight of the column will be borne as a compressive stress in the column.

I conclude that an axial compression of 198kN in each soilcrete column is to be taken into account for the purpose of calculating the shear strength of the columns.

Mr. Williamson submitted that in determining the shearing capacity of the grout columns a number of factors must be taken into account which Professor Pavlovic had ignored. He invited my attention to a paper by Professor H.P.J.Taylor published in 1970. The conclusion expressed in that paper was that there were three factors which contributed to the shear forces in a reinforced concrete beam, unreinforced in shear. Such a beam, or, in this case, column, is, I accept, analogous to the columns of the grout wall in that the latter have tensile reinforcement but are not reinforced in shear. The three factors were the shear forces carried by the compression zone, those carried by aggregate interlock and those carried by dowel action of the main tensile reinforcement. The compression zone is that part of the cross-section that is in compression as a result of bending. Aggregate interlock comes into play only when the beam is cracked. Aggregate interlock refers to the frictional force between lumps of aggregate in the concrete as they rub together when the parts of the beam separate as the crack widens.

Whether the phenomenon of aggregate interlock occurs in soilcrete was spoken to by Professor Pavlovic. He had this to say about it (day 7, pages 59, 60):

Mr. Reid gave evidence (day 19, pages 66 to 68) that there are two kinds of dowel action: small strain dowel action, such as was referred to in Taylor’s paper, and dowel action leading to plastic hinges, as referred to by Mr. Reid in his calculation in the appendix to his report. In my judgment, the difference is simply one of degree. The argument is of little importance unless it be said that the two effects are additive in any given case. I do not accept that they are additive.

The question whether the VHP wall was unfit for its purpose is distinct from the question whether the design of the wall complied with standards. Where, as here, the wall is still standing but not being used as intended, the criterion of fitness for purpose in my judgment is that there should not be an unacceptable risk of failure or unserviceability in the conditions contemplated, viz. the berth having been dredged to -16mCD with an allowance for a further 0.5m. Unfitness is simply non-compliance with that criterion. For the claimant and the part 20 claimant to establish unfitness I have to be satisfied on a balance of probabilities not that the wall will (or would) collapse or be unserviceable, but that the criterion I have mentioned is not satisfied. That inevitably involves applying probabilities to probabilities. The codes can properly be taken into account in deciding whether the criterion is satisfied.

It is common ground that there are two kinds of design check. One check is for serviceability, which broadly speaking is a check that the artefact will not suffer undue displacement or deformation in service. The other check is for the ultimate limit state. That is broadly speaking a check that the artefact can sustain foreseeable loads without collapsing. Those propositions are encapsulated in paragraph 2.3.1 of BS 8110-1:1997, which reads:

In his report at section 5.1.3, Professor Pavlovic said this:

Professor Pavlovic’s and Dr. Greenwood’s opinions differed as to the compressive strength of the soilcrete. The characteristic strength is the strength below which only 5 per cent. of the soilcrete falls. I give a more precise definition below. In situ samples of soilcrete were taken and their compressive strength was measured. The data were set out in two tables that have been put before me, table 3 in Dr. Greenwood’s report and a table in appendix 6 to Professor Pavlovic’s report. The latter shows results for 20 samples; the former shows results for 21 samples, including the 20 samples shown by Professor Pavlovic. The figures have been processed, and the corresponding figures in the two tables do not agree. I shall consider first Professor Pavlovic’s evidence.

There is no standard for soilcrete. Professor Pavlovic referred to British and European standards for concrete. The British standards have not been put before me. The European standards are EN 1992-1-1:2003 (EC2) and a draft standard, prEN 13791:2003. Characteristic strength was defined in Professor Pavlovic’s appendix as the strength below which no more than 5% of all the strength test results (whether from standard test or in-situ test) are expected to fall. The test should be carried out on the 28^th day after the concrete is cast. Design characteristic compressive strength, (fc), is the characteristic compressive strength measured by standard tests of cylinder specimens made according to the standard method which forms the specimen by placing, into the standard mould, concrete from the same batch as the one used for the actual structure as cast. In-situ characteristic compressive strength, (fc,is) is the characteristic strength of a finished concrete structural element (often obtained by coring a cylinder specimen from the member). The concrete sample may be tested as a standard cylinder or as a standard cube specimen. The cube strength is usually 20 per cent. higher than the cylinder strength. The samples tested were not of standard shapes. In each case, measurements of peak stress were made. The in-situ cube strengths of the samples were calculated from the peak stress by use of a formula provided by BS 1881-120:1983. Concrete (and soilcrete) strengthen with age for a period after casting. The in-situ cube strengths were converted to the 28-day in-situ cylinder strengths by use of the following procedure. First, convert the cube strength to cylinder strength by multiplying the former by a shape factor of 0.8 [sic]. Then, following EC2, paragraph 3.1.2(6), divide the result by the factor (βcc)(t) where

The twenty values of 28-day cylinder strengths varied between 5.5N/mm² and 29.1N/mm². One then works out the sample standard deviation, s, of those values. That is the estimated standard deviation of a much larger population of putative measurements. What is sought is the value of the strength below which only 5 per cent. of that putative population would fall. The design characteristic strength (fc) is evaluated according to the following expression given in EN13791 (which has not been put before me):

Professor Pavlovic observed that the distribution of the values of 28-day in-situ cylinder strengths was far from Gaussian and that in consequence it would be unfair to apply the standard statistical calculations which depended upon such a distribution to obtain the characteristic strength since the result would yield lower strength values. In his opinion, the actual characteristic strength should be found from a cumulative frequency plot. By an unexplained method, he thereby arrived at a design characteristic strength (fc) of 7.35N/mm².

Dr. Greenwood adopted two approaches, both different from those of Professor Pavlovic. In his first approach, Dr. Greenwood estimated the water/cement ratio of the grout. I have extracted the following description from his report, paragraphs C125 et seq. The compressive strength of the mixed product (sc., the soilcrete) is determined by the water/cement ratio almost irrespective of the type, sizes and proportion of the other inert mineral content, i.e. the soil, provided that the mineral is stronger than the grout. That is broadly true of all cement based grouts. The relationship between the water/cement ratio and compressive strength has been established empirically. One such correlation was shown graphically in his figure 11. Figure 11 was a family of curves of compressive strength of cement grouts against water/cement ratio for different ages of the grout from 1 day to 1 year. Its source was Farmer, I.W., (1970) “Design of cement grouts”, an unpublished internal report of the Cementation Company Limited research department. Dr. Greenwood set out in a table four values of unconfined crushing strength of 17, 14, 10 and 10 N/mm². He had obtained those values from the water/cement ratio and the use of Farmer’s curves. He said that the calculated results were at best a guide to expectations based on the chosen parameters.

Dr. Greenwood’s second method was to use the test results referred to above. He said that the strengths of cemented soils and concretes increased rapidly during the first few days after casting. The strength roughly doubles from 3 to 7 days and almost a third again to about 28 days. Thereafter the rate slows: increases to 90 days or more are relatively small. He said that it was usual to compare strengths at the arbitrarily chosen time of 28 days. A more appropriate comparison for mixes of 0.8 to 1.0 water/cement ratios should be made by allowing a one-third increase of strengths recorded for tests at 8 and 13 days and a one-third reduction of test results for samples over 100 days old. It was by that means that he arrived at his 21 results mentioned above, which differ from those calculated by Professor Pavlovic. The mean of Dr. Greenwood’s values of “compressive strength 28 day equivalent” was 15.24N/mm². He stated that the characteristic strength was 12.60N/mm². He did not explain how he arrived at that figure. If one applies to Dr. Greenwood’s figures the European code method adopted by Professor Pavlovic the characteristic strength comes out at 5.44N/mm². That figure is to be compared with Professor Pavlovic’s figure of 5.88N/mm², not with his figure of 7.35N/mm².

I arrive at the figure of 5.44N/mm² as follows. To calculate the sample standard deviation, one sums the squares of each individual value, and multiplies that sum by the number of values (in this case, 21). One subtracts from that product the square of the sum of the individual values. One divides that difference by the product of the number of entries and that number minus one (i.e. by 21 X 20). The sample standard deviation is the square root of the answer. It comes to 7.17. The rest of the calculation is as explained above.

I cannot accept Dr. Greenwood’s figure of 12.60N/mm² as the characteristic strength.

Dr. Greenwood went on to consider twelve crushing tests made on samples cast from effluents. His final conclusion on strengths, taking into consideration all test results, was that the evidence suggested strengths somewhat higher than anticipated from calculation based on design assumptions. Since, like concretes, the soil cement comprises natural, inherently variable materials, variations in its strength from place to place or sample to sample are expected. A publication of A.M.Neville, Properties of Concrete (1995), showed a linear relationship based on numerous test results between mean test cube strengths for concrete and their standard deviation. For strengths in the range 10 to 15Mpa (N/mm²) the standard deviation was about 0.6Mpa. Dr. Greenwood said (quite rightly, as indicated above) that there would be a larger standard deviation for soilcrete than was typical for concretes. He concluded that the characteristic strength of the soil cement would be expected to lie between 8 and 12Mpa.

Dr. Greenwood did not explicitly contradict Professor Pavlovic’s approach. Professor Pavlovic said this in the course of his cross-examination on Dr. Greenwood’s conclusion (day 7, pages 71, 72):

Mr. O’Brien considered the question of compressive strength without any explicit reference to characteristic strength. In his report at section 11.2.2 he referred to a population of 34 relevant core samples that had been obtained. The mean ultimate compressive strength of those samples was 15.6Mpa and the standard deviation was 6.6Mpa. Whether that standard deviation was ordinary standard deviation or sample standard deviation does not appear, since the data for the individual samples are not given. Mr. O’Brien said that there were three approaches to selecting a design strength and the associated material factor (to change from a characteristic to a design strength). First, one could assess a “mean” strength, and use a high partial factor on material strength, say 3.0. Second, one could assess a plausible worst credible strength and use a low partial factor on material strength, say 1.0. Third, one could directly consider the statistical variability of the material strength, and derive the design strength from that. He did not evaluate design strengths by either of the first two of those approaches. But since the mean was 15.6, on the first basis the design strength would be 5.2Mpa. On the second basis, a least strength greater than the lowest value of the sample would scarcely be plausible. That lowest value is not stated, but one might take a figure of 5.5Mpa, being the lowest value in the sample used by Professor Pavlovic (Dr. Greenwood’s lowest value was 6.1Mpa). On the third basis, referring to research in Japan (Futaki and Tamura, 2002), Mr. O’Brien used a formula that gave a figure of 7.02Mpa. He said that allowing for a modest strength increase due to age effects would probably lead to a 10 per cent. increase in that design strength. He applied no factor of safety to it. He said nothing else about the ages of the samples in the population. If one applied that third method to Dr. Greenwood’s statistics, the answer would be 5.9 or 6.1MPa, depending on which definition of standard deviation is intended in the formula. The corresponding figures for Professor Pavlovic’s sample would be 6.2 and 6.5. The strength of each sample in Professor Pavlovic’s and in Dr. Greenwood’s populations was adjusted to 28-day strength.

Mr. O’Brien also stated the following:

Finally, Mr. O’Brien said that there were no available compressive strength tests for the soilcrete in the laminated silts/clays and sands below -12mCD. The strength of the soilcrete was likely to reduce below that level by reason of the increase in silt/clay content. For the laminated silts he believed it to be reasonable to assume that the design soilcrete strength was about 60% to 70% of that in the overlying sands, i.e. between about 4.2 and 4.9Mpa.

I prefer the approach by way of characteristic strength, supported as it is by published recognized standards (albeit in relation to concrete), but in truth it makes little difference. I conclude that the characteristic strength of the soilcrete was about 7. If one takes Professor Pavlovic’s figure of 7.35, then the shear strength of a column under a compressive stress of 198kN falls from the figure of 504kN (see paragraph 210 above) to 433kN.

As to the material factor of safety, Professor Pavlovic has, in my judgment, undoubtedly been generous in arriving at a strength of 4MPa after the application of a safety factor. I reach that conclusion disregarding Professor Pavlovic’c points “secondly” to “finally”. I take the characteristic strength as 7.35, so the factor of safety that he has applied is effectively 1.84. Since the strength of the material comes into the calculation of shear strength as a cube root, the effect of the factor of safety is to apply a divisor of the cube root of 1.84, namely 1.225, to that element of the shear strength which is independent of the axial compression, namely 403kN, leading to a figure of 329kN. But EC2 recommends the application of a material factor of safety of 1.5 even in the case of concrete (see section 2.4.2.4 of EC2), which is not to be reduced by taking its cube root. I am entirely satisfied that a factor of safety for soilcrete should be taken which is greater than that recommended for concrete, especially in view of Professor Pavlovic’s second point quoted above (that the strength of the soilcrete is likely to be less at the depth which is relevant for present purposes). Application of a factor of safety of 1.5 to a shear strength of 403kN yields a shear strength of 268 kN. To that figure 30kN has to be added to allow for the axial compression, making a figure of 298kN.

One of Professor Pavlovic’s graphs based on the EC2 formula was drawn on the basis of the following three safety factors: first, a material factor of 1.5 appropriate to concrete; second, a material factor of 1.15 for steel reinforcement; and third, a load factor of 1.35. His graph based on those factors showed a shear strength at zero axial stress of 230kN. I do not know how he arrived at that figure. He may have performed a more sophisticated calculation than is provided by the EC2 design code. Whether one divides 474kN by the product of those three factors, or by the product of the last two and the cube root of the first, one does not arrive at 230kN. In the former case, one arrives at a figure of 204kN. In the latter case, one arrives at a figure of 267kN. At this point I am not concerned with load factors, so I shall for the time being ignore the load factor. The strength of the steel reinforcement does not come into the formula since the reinforcement is tensile reinforcement and the formula is given under the heading “Members not requiring design shear reinforcement” in EC2. Thus I shall ignore the steel material factor. On the totality of the evidence I conclude that for the purpose of determining whether the wall is fit for the purpose so far as the shear strength of the columns is concerned, each column must be taken to have a shear strength of 298kN.

Mr. Reid’s evidence was that on the basis of the design spacing of 1.7m, each column is required to deliver an ultimate shear resistance of 315kN. Mr. Bracegirdle’s ICFEP figure was 200kN/m, i.e. 340kN per grout column.

Thus I am satisfied that the design of the wall was unfit for its purpose by reason of shear strength.

The position as built is this. Mr. Reid referred to a “worst case of an average spacing of 2.7m” (his report, appendix, sheet 13). In such a locality, the required shear strength of a column is 329kN X 2.7/1.7, i.e. 522kN. The experts had agreed to the following proposition:

I have mentioned the design average spacing of 1.7m. That is between centres of adjacent columns. In fact, the design was such that the columns (i.e., column pairs) were in triplets spaced at 1.68m, 1.29m and 2.11m. Thus considering any given column, the sum of the half distances between it and its two immediate neighbours was 1.485m, 1.700m or 1.895m, depending upon where in the triplet the given column was situated.

Mr. Bracegirdle gave evidence (his report, figure 55) which I accept about actual spacings between adjacent columns as built. Taking the “spacing” of a column as the sum of the half distances between it and its immediate neighbours, Mr. Bracegirdle gave evidence, set out in figure 55 of his report, that 37 columns had spacings of 2m or more. There were eight separate groups of five columns where the average spacing, calculated in accordance with Mr Reid’s revised method, amounted to 2.3m or more (up to 2.9m). Thus as built the wall was in this respect unfit for its purpose to a greater degree than it was as originally designed (but see paragraph 265 below).

Mr. Reid countered the argument that the grout wall would fail in shear by propounding the opinion that the wall would act as an arch against transverse forces. One has to imagine a vertical arch resisting transverse forces in the same way as a horizontal arch resists weights (and hence supports objects) placed or moving upon it. The grout wall is not, of course, in the shape of an arch, but the proposition was that the forces within the material would react to the horizontal forces in the same way as an arch would react to vertical forces. An arch, in order to develop its strength, needs to be firmly fixed at its two ends by forces pointing inwards. In the case of an ordinary arch, those forces would be horizontal. But in the case of the arch postulated in the grout wall, they must be vertical, downwards at the top of the arch and upwards at the bottom of the arch.

One objection to this theory is that the curvature of the grout wall changes direction twice (at least) along its length (an opinion held both by Professor Pavlovic and by Mr. Reid). Mr. Reid said (day 17, page 70; day 18, page 63) that in that case three arches would have to be postulated. The arch in the middle would have its notional concavity facing in the opposite direction to those of the other two arches. It is clear that such an arch would not be sufficiently supported at its ends. Mr. Reid put forward another argument in which there was a single arch, the vertical support being provided by the reaction of the seaward soil to a horizontal point load imposed by the landward soil at the level of the bottom of the sheet piles. Any tendency of the wall to move vertically would be resisted by friction arising from that reaction. That would provide the necessary firm fixing, at any rate at one end of the arch. However, the reaction in question would operate principally at the point where the lateral load was imposed. That is analogous to the load on a horizontal arch being imposed at the end of the arch. The effect of the arch as such is lost. Mr. Reid said this during the course of his cross-examination by Mr. Bowdery (day 23, page 52):

In the re-amended defence and counterclaim the following plea appears in paragraph 100(2)(f):

In his report, Mr. Reid considered two imaginable modes of failure of the sheet pile/grout wall system. The second such mode was by failure of the soil below the dredge level on the seaward side of the sheet pile wall to resist the pressures imposed on the landward side of the wall. In that scenario, he said, the passive wedge in front of the wall yields and allows the sheet piles to move forward. When that occurs, the grout columns are mobilized to provide additional shear resistance and prevent the wedge of soil behind the sheet piles from translating forward. He considered that the grout columns were indeed sufficient for that purpose, because of the dowel effect (Footnote: ¹³) produced by the tubular reinforcement in the grout wall. He expressed the opinion that there was no credible mechanism that would generate a shear failure at the base of the sheet piles, provided that the sheet pile wall was not highly stressed prior to the implementation of the strengthening works.

Mr. Bracegirdle’s ICFEP analysis produced the result that the grout wall would be subjected to a shear force of 200kN/m, which corresponds to a shear force of 340kN for each grout column assumed to be spaced at the average design interval of 1.7m. One might suppose that that was evidence that the system would fail if the grout wall could not sustain that shear force. Indeed, Mr. Bracegirdle gave evidence (day 8, page 27) that the VHP wall did not have enough strength to sustain the loads imposed upon it by the proposed dredge to -16.5mCD. I would interpret Mr. Reid’s opinion as merely asserting that the grout wall would sustain the shear loading that he calculated would be imposed upon it. There is no suggestion there that there would be no collapse if the grout wall were subjected to a shear force greater than it could sustain.

Mr. Bracegirdle introduced an exhibit illustrating a global mode of landslip with the shear surface passing underneath the sheet pile wall. He gave evidence (day 15, page 4) that he had done so in response to a criticism by Mr. Reid, which I take to be the opinion expressed above, that ABP’s finite element analysis did not predict a credible failure mechanism. I find Mr. Bracegirdle’s further evidence on this point difficult to follow. He appears to have thought (rightly or wrongly) that what was being said was that even if the grout wall failed there would not be a collapse. He said this (cross-examination, day 15, pages 5 to 8), with reference to a diagram in his report showing movements of the soil:

My first comment is that 180 or 190kN/m is not very small in relation to 200kN/m, though it is certainly less, which is surely sufficient for the argument. Second, if it were the case that there would be no land slip even if the grout wall were not there, I cannot accept that placing the grout wall there would cause it to be subjected to stresses that it could not sustain. If the soil is itself strong enough to resist the tendency to slip, it would not impose any shear strain on the grout wall, apart perhaps from a small effect caused by introducing the wall itself, since the soil would already be stable. Hence there would be no significant shear stress on the wall. Putting the matter another way, if there were no risk of the soil slipping, there would be no point in introducing the grout wall.

The key to the confusion is the distinction between a local landslip and a global landslip. Mr. Bracegirdle was considering the latter as a failure mechanism. Mr. O’Brien produced some diagrams in exhibits X38, X41 and X88 showing the results of some stability calculations that he had carried out in response to Mr. Bracegirdle’s evidence as to landslip. The calculations related to the potential for a mechanism for failure of the quay below -20.7mCD, i.e. below the bottom of the sheet pile wall and through the grout wall. The mechanism was a land slip along a failure surface which had a cross-section in the form either of the arc of a circle or of arcs of less regular curves roughly circular.

In those exhibits Mr. O’Brien carefully considered the soil characteristics at two different locations along the wall, and displayed diagrams with legends showing his calculated factors of safety, all sufficient to comply with standards. He gave no indication of the forces to which the grout wall would be subjected. Indeed, he did not state whether he followed Mr. Bracegirdle in ignoring the effect of the grout wall. However, it seems that he did. The following is an extract from the conclusions he expressed:

Mr. O’Brien gave this oral evidence (day 26, pages 77 and 78):

In what follows, I refer to the friction angle (angle of shearing resistance) φ (φ´ means effective friction angle) and cohesion c (c´ means effective cohesion). Whilst different kinds of soil have to be treated in different ways in calculations, for the purposes of this part of this judgment, the following can be assumed: (1) The greater the angle φ, other things being equal, the greater the strength of the soil; (2) the greater the value of c, other things being equal, the greater the strength of the soil. In particular, the greater the angle φ, the less the active pressure coefficient and the greater the passive pressure coefficient. Putting the matter simply, the stiffer or stronger the soil, the less likely it is to push a wall over, and the more effective it will be in supporting the wall against being pushed over.

In his exhibit X38, Mr. O’Brien plotted a putative profile of a global failure surface through the soil passing through a level of -21mCD at its lowest point. That was drawn to illustrate his proposition that there would be a factor of safety greater than 1.0 if the stratum below -18mCD had a friction angle φ greater than 31°. It showed a dredge level of -16.0mCD. Mr. Jones put to Mr. Bracegirdle that if you dredge down another half metre, that might reduce the safety factor by about 3 per cent. He replied:

Mr. O’Brien accepted that a φ figure of 28° below -18mCD was possible. The following exchange took place (day 27, page 25):

One such analysis described as worst credible has been put before me. It shows a profile where, at the relevant level, the value of φ is 34°. The factor of safety is shown as 1.168. On Mr. O’Brien’s own evidence, I cannot accept that a φ value of 34° at that level can be the worst credible.

In the course of his cross-examination by Mr. Bowdery, Mr. O’Brien gave the following evidence (day 27, page 32):

I cannot accept that the risk even of a global failure is negligible, having regard to the factor of safety of less than one even for a friction angle of 31°. Mr. Reid explained in his report (appendix, sheet 7) that he was considering a different mechanism. Although it involved consideration of the global surroundings, it was a study of the balance of forces and identified the extent to which the passive resistance in front of the sheet pile wall was deficient in carrying the ultimate loading applied. A diagram of the mode of failure in question appears as the third illustration (bottom left) in figure 3 at page 22 of the British Standard to which Mr. Reid referred (BS 8002:1994). It would certainly involve local movement of the land, but is to be distinguished, as Mr. O’Brien did, from overall global slip. I do not understand Mr. O’Brien’s remarks about the demands on the wall being “much lower” as being intended as a comment on, or as disagreement with, Mr. Reid’s evidence. If I am wrong, then I prefer the evidence of Mr. Reid.

Mr. Reid’s calculation was based on sections 3.2.5, 3.3.3.2, 3.3.3.3 and 3.4.2.3 of BS 8002:1994. Mr. Bracegirdle used the ICFEP finite element analysis, run 210. The respective values of φ´ and c´ that they adopted appear in the following table. The letters B and R refer to Mr. Bracegirdle and to Mr. Reid respectively.

The effect of the differences between the values of those parameters as assessed by Mr. Bracegirdle and by Mr. Reid has not been explored in detail. It will be seen that Mr. Bracegirdle and Mr. Reid were in agreement as to the values of φ´=28° and c´=5 at the level of maximum shear force in the grout wall. Those same parameters were used by Haecon in their design to represent the soil below -18mCD. I do not accept that they are “worst credible” values, but they are cautious estimates. Professor Pavlovic gave evidence, and I accept, that an action factor of safety of 1.35 should be applied, that is a factor of safety in relation to the imposed forces. No mention has been made of the application of any action factor of safety in the ICFEP analysis. It seems from a remark of Mr. Bracegirdle (day 13, page 12) with reference to a run of the program which has not been put before me that a factor of safety had to be applied to the result, that no factor of safety was included in the program. Nor was any action factor of safety mentioned by Mr. Reid in relation to his calculation.

In order to give a reasonable standard of confidence that the wall will withstand shear stresses imposed on it, and thus be fit for its purpose, each column must in my judgment be capable of withstanding a maximum shear force of 340kN. I am satisfied on a balance of probabilities that as designed it was not capable of withstanding such forces and was thus not fit for its purpose. The position was even worse as built.

Bond in this connection means the adhesion between the steel reinforcement and the surrounding soilcrete. Professor Pavlovic gave evidence that the bond was insufficient, thereby rendering the grout columns unfit for their purpose. He calculated that the bond stress, by which he meant the stress tending to break the bond, was 1Mpa (i.e. 1N/mm²), whereas Haecon had determined that the columns should have a designed bond strength of only 0.6Mpa, having calculated that the bond stress would not exceed 0.43Mpa. Thus the design was defective. Mr. Williamson submitted, and I accept, that the relevant comparator to the bond stress was not the design bond strength but the actual bond strength.

Dr. Greenwood gave evidence that the ultimate compressive strength of the grout was about 20Mpa. I accept that evidence. He applied a rule of thumb (day 25, page 51) that the bond strength was one-tenth of the ultimate compressive strength to arrive at a figure of 2Mpa as the bond strength of the steel reinforcement to the grout. Professor Pavlovic (day 6, page 138) agreed with the figure of one-tenth. In cross-examination he referred to a grout strength of 11.3Mpa which had been determined in experiments carried out by Professor England and Dr. Tsang of Imperial College. Those experiments had been instigated by Professor Pavlovic and were reported as Appendix 6 to his expert’s report. The purpose of the experiments was not to determine the bond, but the report in appendix 6 contained a reference to the grout/steel bond being impaired by channels created by flow of water as the grout “bled”. Professor Pavlovic’s primary point, however, was that the bond between the grout and the soilcrete had not been determined. He said that he was aware of no data to estimate how good a bond results between grout and soilcrete.

Mr. Williamson submitted that Dr. Greenwood in his report had shown that the bond strength between grout and soilcrete was 3Mpa, even higher than the steel/grout bond strength. I reject that submission. What Dr. Greenwood was considering in the relevant section of his report was the shear between the overlapping soil-cement (i.e. soilcrete) columns. The two columns constituting a combined double column were grouted at different times. Dr. Greenwood was concerned with the adhesion of the newly grouted column to the column where the grout had gone some way, or all the way, to solidifying. He concluded that for a compressive strength of 15Mpa there would be a bond strength for a “cold” soil cement (i.e. soilcrete) joint of about 3Mpa. If there were any doubt about what Dr. Greenwood was saying, it would be resolved by his oral evidence (day 25, page 69):

It was said that it was no longer the practice to consider local bond stress, because of anchorage. If a reinforcing bar in concrete is bent round at the end it is thereby anchored into the concrete. It takes a large force to pull it out of the concrete against the hook. The force required to pull it out is the anchorage force. The anchorage length or equivalent anchorage length, though not defined in any document that has been put before me, appears to be the length of a straight reinforcing bar that would require the same force to pull it out. Be that as it may, the standards no longer require the calculation of local bond stress because of the general experience of anchorage. Eurocode 2 (EC2; BS EN 1992-1-1:2004), which relates to concrete structures, provides in section 8.4.2, headed Ultimate bond stress, a requirement of principle that the ultimate bond strength shall be sufficient to prevent bond failure. The design value of the ultimate bond stress, (fbd), for ribbed bars may be taken as:

The formula given in section 8.4.2 of EC2 applies only to ribbed bars. The reinforcement in the grout columns is smooth. That in itself would tend to reduce the bond (evidence of Professor Pavlovic, day 7, page 44). On the other hand there is this evidence of Mr. Reid (day 17, page 81):

I conclude that Professor Pavlovic was right when he said that nobody knows the bond strength or the anchorage length of the reinforcement tubes. It is unnecessary for me to consider an argument of Mr. Williamson that Professor Pavlovic’s calculated figure of 1MPa for the bond stress was too high. I am not satisfied that the bond strength of the grout wall as designed was such as to render the grout wall unfit for its purpose.

The VHP columns were installed in pairs with the design intention that they would act as a composite pair, i.e., combined into a single column. Composite action greatly increases the bending moment capacity of the VHP columns. Professor Pavlovic said that loss of composite action results in a clear bending failure. Loss of composite action can occur if the diameters of the columns are substantially less than the design diameter, or if the axes of the two columns forming the pair are not parallel. Any error in positioning of the initial point of the drilling can also affect the outcome. It appears from Professor Pavlovic’s report, though the matter has not been the subject of comment, that loss of composite action actually increases the shear strength of the pair of columns (assuming any given diameter). The question of failure of the wall in bending, as opposed to shear, has not been pursued before me in the evidence and I shall not consider it.

As designed, the maximum overlap between the two members of a pair was 200mm, since the design diameter of each column was 900mm and the design separation of their central axes was 700mm. There is no direct evidence of the diameters achieved. The parties’ experts were at one in considering that the diameters would be likely to be less than 900mm at depth. Their views were not widely divergent. Mr. Bracegirdle considered it highly unlikely that the diameters would exceed 800mm, and the likely diameter was probably about 600 to 800mm. Dr Greenwood considered that the diameter could fall to 800mm at the toe of the columns, but might be in excess of 900mm at a depth of -10mCD. Mr. O’Brien suggested 750mm to 850mm.

It is clear that the bending and shear capacities of the columns will diminish as the diameters of the members diminish, but no calculations have been put before me of the quantitative effect of such diminution below the design diameter of 900mm.

I am satisfied that the diameters fell to about 800mm near the toe of the wall. Assuming composite action, that would have increased the margin by which the grout wall had insufficient shear capacity. The question arises whether that is a matter of design, for which Haecon would be responsible to HSS, or a matter of workmanship outside the design, for which HSS would have no remedy against Haecon.

Mr. Williamson submitted:

On the evidence of Dr. Greenwood, I am satisfied that the diameters of jet-grouted soil-cement columns formed in natural soils depend on the jet momentum and soil properties and structural characteristics which affect its resistance to erosion.

The design contract between HSS and Haecon does not specify or make mention of the input parameters for the jet grouting. It does not oblige the designer to verify that his design has been satisfactorily constructed.

The design is contained in the execution design report USR 2334/130 revision 2 dated 3^rd June 2002. It states:

It is true that Haecon gave the certificate mentioned by Mr. Williamson.

There is no suggestion that a diameter of 900mm at depth was not achievable. I am satisfied that the failure to achieve diameters of 900mm at depth is a question of workmanship and does not involve a breach of the design contract on the part of Haecon.

Mr. Williamson continued:

In my judgment, Haecon were under no duty to HSS to check the workmanship of HSS. No case has been argued that HSS has a remedy against Haecon on some other ground.

Loss of composite action can also arise from divergence of the two members of a pair of columns. Both are intended to be vertical, but it can be that one or both are not vertical. There is no direct evidence as to the verticality of the columns. Mr. Bracegirdle calculated, and I accept, that deviations of only 0.42 per cent. from the vertical in opposite directions can completely remove the interlock between the columns at -18mCD, assuming correct initial positioning of the drills.

Mr. Claessens supervised the following elements of the work: the installation of the guidance tubes; the grouting of the VHP columns; drilling works through the VHP columns; and installation of steel reinforcement tubes into the VHP columns. He exhibited to his witness statement a table showing standard procedure under which he was working. That showed that the maximum permissible tolerance in verticality of the columns was 1 per cent. Mr. Claessens wrote:

Professor Pavlovic said that the out-of-vertical tolerances were unlikely to be less than 1 per cent. and might be as high as 2 per cent. The European standard on jet grouting EN 12716:2001 at clause 8.4.4 provides:

Mr. Bracegirdle said that the normally accepted tolerance on verticality was “about double” 0.42 per cent.

A spirit level had been used to assist in making the columns vertical. Dr. Greenwood considered that a spirit level could have been used which would, in practice, give an accuracy of 1 in 250. He concluded that the deviation was likely generally to be much less than 100mm for “first phase” treatments and less than 150mm for “second phase treatments”. The latter would apply, in particular, at a depth of about -20mCD for the work done after the system of making each column of a pair in two passes had been adopted. The figures of 100mm and 150mm allowed for error in the initial placing of the drill. Clause 8.4.3 of EN12716:2001 allows a tolerance of 50mm in the drilling starting point. Dr. Greenwood considered that that was reasonably attainable in normal working circumstances, but not generous. He said that he would expect that positioning the drill rig to the point marked out by surveyors would be within that standard tolerance but not necessarily by much margin.

If a spirit level, when used in the circumstances contemplated, can give such accuracy as 1 in 250, it is not clear to me why the standard allows 2 per cent. However that may be, on the totality of the evidence on this point I accept that of Dr. Greenwood set out in the paragraph above.

If one column is vertical, the maximum displacement between its axis and that of its pair at -20mCD is likely to be 150mm. If that occurs, and if the diameters of the two columns are 800mm at that depth, then the separation between the columns will be 50mm. If the deviation of the inclined column were in a direction making an angle within ±127.4° of the line joining the intended centres at the top, the axes would be not more than 800mm apart at that depth, and there would not be complete loss of composite action. If the deviation of the inclined column were at a random direction relative to the line joining the intended centres at the top, the probability that there would be complete loss of composite action at a depth of -20mCD would be about 29 per cent. (2 X 52.6/360). What is perhaps more likely in any given case is that each column deviates from its intended position by 75mm at the depth mentioned. That leads to a corresponding probability of about 12 per cent. The calculation is set out in appendix 4 to this judgment. It is a rough calculation, precision not being justified by the imprecision of the data. Other combinations of pairs of deviations from intended axes from 0mm, 0mm to 150mm, 150mm are also possible. The calculation depends on two assumptions: that all angles of deviation from the intended position are equally likely, and that the probabilities of the directions of the deviations of the two members of a pair are mutually independent. In my judgment, those are reasonable assumptions. There is no reason to suppose that any errors in positioning or verticality of the two members of a pair of columns are other than random or are interdependent, nor has the contrary been suggested. There is, in my judgment, in view of the figures given by Dr. Greenwood, a substantial probability that the sum of the deviations (whatever their directions) of the members of a pair of columns will equal or exceed 100mm. I conclude that it is highly likely that several columns have lost their composite action at-20mCD.

Mr. Williamson submitted that this statistical approach was inappropriate. There was no direct evidence of any deviations. I would accept that if there were only a small number of columns, the statistical evidence, whatever answer it produced, might well be insufficient to enable the court to make a finding. But here there are 412 columns and I am well satisfied on a balance of probabilities that there must be several instances of loss of composite action. Whether the effect of that in the event of further dredging is likely to be a complete collapse of the wall, or local deformation to an unacceptable or an acceptable extent does not appear. But in my judgment, a wall such as the present where several of the columns have lost their composite action at the relevant level and others have diminished shear resistance and bending moment capacity cannot be regarded as fit for its purpose.

In about 8 per cent. of the columns as made, the reinforcement was outside the grout at the lower levels. That was dealt with by further grouting to cover the reinforcement. It was said that the eccentricity of the reinforcement would weaken the columns. Professor Pavlovic produced calculations to that effect. However, details as to the extent of the regrouting were not in evidence, and in those circumstances it is not possible for me to reach any conclusion as to the effect of the eccentricity of the reinforcement on the strength of the columns.

I am satisfied that the grouting works were not fit for their purpose in that they caused cracking and bulging or further bulging of the sheet pile wall. The question arises whether that was a result of the design of the works. If the consequence was confined to instances when the annulus was blocked and grout or soilcrete pressure on the back of the sheet pile wall rose substantially above hydrostatic pressure plus the pressure required to drive the effluent up the annulus, then I would not attribute the consequence to design. When blockages occurred and there was continuous fluid between the vicinity of the monitor and a substantial area of the sheet pile wall, it was necessary almost instantaneously to lift the monitor. If there was not continuous fluid, there would be some period of grace. The length of that period is not in evidence. Whilst there may well have been such occasions when the monitor was not lifted in time, I am not satisfied on a balance of probabilities that there were. I am satisfied that the consequence could occur in the absence of such blockages and in such a case the consequence is attributable to the design. However, the evidence is insufficient to enable me to decide which is the more probable: (a) that there were occasions when the monitor was not lifted in time, or (b) that there were no such occasions. Thus I am not satisfied that this aspect of the unfitness of the works for their purpose is attributable to the design of the works.

Mr. Williamson submitted that ABP’s fitness for purpose case was not concerned with the condition of the sheet pile wall, or with the pressure imposed on the sheet pile wall by the liquid grout, but only with Professor Pavlovic’s criticisms of the completed grout columns. That submission was made for the first time on the antepenultimate (29^th) day of the trial. I reject it, for the following reasons.

The following is an extract from ABP’s amended particulars of claim:

In his opening, under the heading “ABP’s case on the defects and deficiencies in the installation of this defective design”, Mr. Bowdery referred to the loading on the sheet pile wall caused by pressure of grout. In his opening, Mr. Williamson said that ABP advanced a series of criticisms of the design and workmanship, detailing a positive case that the works were not fit for purpose. He stated one of the broad issues as the impact of the carrying out the works on the sheet pile wall (assuming that it was not already deflected past its elastic limit) and referred to the amended particulars of claim.

Thus as regards the statements of case and the submissions of counsel, other than the submission under consideration, the matter was indisputably in issue. But Mr. Williamson relied on the experts’ joint statement and Mr. Bracegirdle’s evidence. The first item of the experts’ joint statement upon which Mr. Williamson relied was part of the preamble, headed Remedial Works. The first paragraph of that item states:

The existing loading must mean the loading in the context of the existing dredge level of -12.8mCD. The second paragraph reads:

Mr. Williamson’s point here is, I think, that ABP do not state that remedial work prior to further dredging is rendered necessary by the distortion of the sheet piles. But it must be implicit in the view expressed by Mr. Reid and Mr. O’Brien (though this item as such was not put to them) that whatever the cause, it was the pre-existing condition of the sheet pile wall that necessitated the remedial works. Thus its condition after the grouting, being, to say the least, no better than it was before, must also on this view have necessitated the remedial works. I do not regard this agreement as removing from controversy in relation to the fitness for purpose issue the question of the cause of the configuration that the sheet pile wall bore after grouting. The remainder of this item is not relevant to the present argument.

The second item of the experts’ joint statement upon which Mr. Williamson relied was item number 79. That item had not previously been mentioned once during the whole proceedings. It reads:

The next, and last, item to which Mr. Williamson referred in this context was item number 80. That item reads as follows:

Finally in this context, Mr. Williamson referred to evidence of Mr. Bracegirdle. Mr. Bracegirdle said this:

The coping beam claim is a claim by HSS pursuant to clause 12 and clause 52 of the Conditions of Contract that it is entitled to additional sums as a result of having to drill or core through an additional unforeseen thickness of concrete under the coping beam in order to instal the VHP columns. The claim was for an additional payment of £303,424.97 exclusive of VAT. The amount has now been agreed at £69,653.12, subject to proof of liability. The basis of the claim is that the extra thickness of concrete constituted a physical condition or artificial obstruction which was not reasonably foreseeable by an experienced contractor. It is common ground that HSS had to drill through thicknesses of between 250mm and 2400mm of concrete in addition to the 600mm thickness of the coping beam for which HSS made allowance in its price.

The effect of clause 12 of the contract is given in paragraph 12 above. Clause 52 relates to valuation of the claim.

The only drawing relevant to this claim to which I was referred during the course of the hearing is drawing number 2015/133C, dated 22^nd June 1970. That drawing shows the coping beam with two inches of blinding concrete below it. There is no relevant legend at that point. The two inches is an approximate scale measurement. There is, however, a reference to two-inch blinding below the relieving platform. No other concrete is shown below it. The drawing does show massive concrete in that position at points along the length of the wall where the bollards are founded. It was suggested that absence of concrete in that position except where the bollards were founded was to be inferred from the drawing.

Mr. Bowdery relied on the terms of the specification in the contract between ABP and HSS, which, he said, expressly provided that ABP did not warrant the accuracy of drawing 2015/133C. He also submitted that HSS in any event assumed the risk in respect of the thickness of the coping beam and the material around it by virtue of clause 11 (Footnote: ¹⁴).

As to the former submission of Mr. Bowdery, the specification provides by clause 101 05:

Mr. Bowdery also relied on clause 101 06A, which provided:

Regarding the terms of clause 11, as to “information available in connection therewith”, such information was in the archive, since the original contractor made a claim for extra payment in respect of the concrete in question. The archive included documents showing the existence and nature of that claim and its rejection. By letter of 19^th May 1971 the then contractor claimed to be unable to obtain the required profile to the backfill material between the culvert and the sheet piling, stated that it had commenced backfilling with concrete to the required profile, and asked that the additional costs be included in its final account. The engineer replied by letter of 25^th May 1971 stating his view that any difficulty encountered in backfilling was the result of the sequence of work adopted by the contractor, and rejected the claim for additional costs.

Mr Bracegirdle (day 10, pages 160, 161) considered that HSS should have foreseen that there was a significant risk that the original contractor would have used mass concrete below the coping beam, though not that it should have foreseen that that contractor would probably have used it.

In his report, section 15.2, Mr. Bracegirdle said this:

Mr. Bracegirdle gave this evidence in the course of his cross-examination by Mr. Williamson (day 10, pages 162 to 164):

On the question of the ease of investigating the thickness of any concrete below the coping beam Mr. Bracegirdle said in his report, paragraph 15.3.1, that it would have been a simple matter for HSS to have used a small diameter hand-held percussive drill.

Mr. van der Eecken said this in the course of his cross-examination by Mr. Bowdery:

I conclude from that evidence a point which may not have been in issue, namely that HSS’s site investigation was carried out before, and not after, the decision to bore through the coping beam to create the VHP columns. If so, HSS would have had no occasion at that time to consider the possibility of the existence of mass concrete under the coping beam.

Mr. Reid said in his report (section G, paragraphs 3 and 5):

However, the question I have to decide is whether the claim falls within clause 12(1). The concrete in question was undoubtedly a physical condition or artificial obstruction. By common consent, and indeed on the pleadings, although the drawing was not labelled “As built”, it was apparent that it was issued for the purpose of constructing the works, and that both HSS and ABP understood it generally to represent the structure of the quay wall. In my judgment, once it had been decided that the VHP columns would require drilling through the coping beam, (and that would have required 412 drillings instead of 24 as originally envisaged), an experienced contractor would carefully consider the implications of drawing 2015/133C. He would ask himself how much drilling would be likely to be required, especially given the size of the job. I am impressed by the evidence of Mr. Bracegirdle. An experienced contractor would be able to deduce, as he did, that the original contractor might well have found it convenient and economic to use mass concrete beneath the coping beam. Notwithstanding the drawing, an experienced contractor in my judgment would at least make enquiries of the employer whether the employer had information on the subject, or would have carried out an investigation using a percussive drill or some other method. This contract is framed in terms onerous to the contractor, but I must give full effect to clause 12. The question is could the mass concrete reasonably have been foreseen by an experienced contractor. I have stated above what in my opinion an experienced contractor would do. But even if I am wrong, an experienced contractor could, I am satisfied, have adopted the reasoning of Mr. Bracegirdle or otherwise have appreciated the risk of the existence of the mass contract. Thus he could reasonably have foreseen it. This claim fails. I am far from saying that an experienced contractor would necessarily do that of which I have found him capable. Under pressure of competition, or in the hope of making a successful claim, he might be willing to take the risk. But in my judgment the drawing would not so set the mind of an experienced contractor at rest as to give him a false sense of security so that he did not appreciate the risk.

HSS claims an additional payment of £1,028,908.28 plus VAT for the disposal of surplus grout returned to the surface from the jetting process. The volume of grout so returned was 22,566m³. The original tender proposal of HSS was that the surplus grout should be dumped on the river bed, with a view to its being removed when the river bed was dredged. That proposal was not acceptable to ABP.

On 28^th November 2001, ABP wrote to HSS in relation to HSS’s tender submission of 8^th November 2001 requesting further clarification/confirmation of nine points, including as point e):

HSS replied on 3^rd December 2001, in relation to point e):

A meeting between the parties was held on 3^rd December 2001. Mr. Brown wanted a lump sum price for disposal of grout as an extra to the contract. The minutes, which were drafted by Mr. Brown, note “HSS to provide extra cost for disposal of surplus grout mixture offsite”. According to evidence of Mr. van der Eecken, which I accept, Mr. Brown made it clear that the reference to “offsite” was to a landfill site. Mr. Brown gave evidence about that meeting. He said in his witness statement that he told Mr. van der Eecken that ABP required HSS to provide a lump sum price to cover all the costs of the disposal of the surplus grout. That evidence was challenged, Mr. Brown having made no mention of a lump sum price in the minutes of the meeting. Nevertheless, I accept that that is what Mr. Brown said he wanted, having regard to evidence given by Mr. van der Eecken (day 3, page 137) with reference to the costs of grout disposal:

Mr. Brown went on in his witness statement:

A correct calculation would be as follows:

On 6^th December 2001 Mr. Brown sent his minutes of the meeting of 3^rd December to Mr. van der Eecken referring to an attached document, which must have been HSS’s clarification/confirmation submission of 3^rd December. Mr. van der Eecken replied the following day:

Annex 1 was headed “Disposal of surplus grout mixture offsite” and sub-headed “Budget outline”. It gave a total under the heading “Load in on site” of £77,616.00. A breakdown of that figure was given in terms of manpower (plant operator and labourer) and pumping equipment, both in terms of hourly rates for 12 weeks. Annex 1 also gave a total under the heading “Disposal offsite” of £110,000.00 broken down as local quotation, £100,000.00 and 10% GO and risk £10,000.00.

Mr. Williamson submitted that having regard to Mr. van der Eecken’s letter of 3^rd December it was implicit in annex 1 that the area on site for loading in must be within 500m of the injection locations. I consider that submission below.

On 19^th December 2001 the parties met. Mr. Brown’s manuscript note of that meeting mentions disposal cost of inert grout of £8 a tonne, density 1.75 tonnes/m³, cost £14/m³; not inert £13 to £20/tonne. It also states “5000m³ calculated amount”. The disposal costs mentioned in the note are equivalent to £22.75 to £35/m³. Mr. Brown gave the following evidence, which I accept. At the meeting, Mr. van der Eecken again explained his method of calculating the figure of 4400m³ before rounding the figure up to 5000m³ to allow for contingencies. Mr. Brown felt that Mr. van der Eecken had reassured him as to the maximum volume of grout for disposal, but he recognized that the price given by Mr. van der Eecken had been based on a verbal quotation and he wanted HSS to confirm its price in writing. The two men discussed the question whether HSS could confirm the surplus grout disposal rate at £20/m³. Mr. Brown gave to the court examples from his own experience of £9/m³ at the relevant time and £11 and £12/m³ currently. At the meeting, Mr. van der Eecken explained that the reason the quote was seemingly expensive was that the landfill operators had not been convinced that the surplus grout would be inert. Mr. van der Eecken stated that for inert material disposal costs would be £8 per tonne, and that for material not classed as inert the cost of disposal could range from £13 to £20 per tonne. It was agreed at the meeting that the disposal costs should be based on the surplus grout being considered to be contaminated. As no sample could be provided to the landfill operator until the start of the works there could be no prospect of confirming a lower rate prior to contract. Mr. Brown was confident that the surplus grout would not be classified as contaminated as there was little or no possibility of there being a contaminant in the ground to be grouted.

Mr. van der Eecken gave this evidence, which I also accept, about the meeting of 19^th December:

On 21^st December 2001 Mr. Brown sent to HSS his minutes of the meeting of 19^th December, which noted that HSS undertook to confirm that the rate used in the disposal item for the waste grout mixture, provided with their facsimile dated 7^th December 2001, was £20/m³. On the same day, 21^st December, Mr. van der Eecken sent to Mr. Brown a fax message which made reference to HSS’s fax message of 7^th December, and stated:

On 19^th March 2002 ABP wrote to HSS accepting HSS’s tender for the sum of £3,856,465 “inclusive of a provisional sum of £110,000 for the cost of disposal of surplus grout arisings”. The letter listed the documents comprising the contract, including HSS’s facsimiles of 7^th and 21^st December 2001 (but not that of 3^rd December). It was signed on behalf of HSS on 26^th March 2002 as having been received. It is evident that the price was also inclusive of the figure of £77,616 as quoted for “Load in on site”, since ABP’s liability to pay that sum, and the fact that it has been certified and paid, are not in dispute. The dispute over that matter is simply as to the location of the site provided and the question whether ABP is liable to pay more than £77,616. How the figure of £3,865,465 was arrived at from the original tender figure of £3,834,563 is not in evidence.

The words “Provisional Sum” are defined as follows in clause 1(1)(l) of the contract:

“Provisional Sum” means a sum included and so designated in the Contract as a specific contingency for the execution of work or the supply of goods materials or services which may be used in whole or in part or not at all at the direction and discretion of the Engineer.

Mr. Williamson submitted that the sum of £110,000 could not be a provisional sum as defined in the contract, since (1) the sum was not a contingency for works which might not be carried out: the surplus grout was necessarily going to require disposal; and (2) the disposal of grout was not to be at the direction and discretion of the Engineer. Moreover, he submitted, the parties could not have intended the sum to be a maximum since they contemplated that the cost of £20/m³ could well have been exceeded and the volume of 5000m³ could have been exceeded. I reject the latter submission. Having regard to Mr. van der Eecken’s calculation, it must at the least have been in the contemplation of reasonable parties that 5000m³ was a likely upper limit to the volume of grout. And the figure of £20/m³ was generous for inert grout, which Mr. Brown considered was the likely description of the grout. It is clear on the evidence of Mr. van der Eecken (day 4, pages 6 and 7) that in the event, the cost paid to the landfill contractor was only £8.67/m³. Mr. van der Eecken said:

It is clear that Mr. van der Eecken understood the meaning of the expression “Provisional Sum” in the contract. He was asked (day 3, pages 137, 139, 140) about his use of the expression in his fax of 21^st December:

Mr. Brown gave this evidence in his witness statement:

The first post-contract meeting between the parties took place on 4^th April 2002. The purpose of the meeting was to resolve as many of the outstanding practical issues as possible so that the strengthening works could begin on site as soon as possible. At that meeting a large number of logistical issues were discussed, including details of where the areas required by HSS during the works would be situated. But there was no discussion of an area for surplus grout handling by HSS. There was no discussion at the meeting that an additional area would be required or that it should be sited within 500m of the grouting operations. HSS did not mention that double grout handling would be required. The only discussion regarding grout disposal was raised by Mr. Brown at the end of the meeting. HSS had not included it as an item on the agenda and the issue was recorded in the minutes under the heading Miscellaneous/Any other business. The minute stated:

A further meeting between the parties was held on 15^th April 2002, when Mr. van der Eecken was absent. HSS’s requirements regarding areas on the quay and within the terminal area were again discussed and agreed. But there was still no discussion regarding the disposal of surplus grout or any requirements on the part of HSS for space for the purpose. As recorded in the minutes of the meeting:

On 24^th April 2002 HSS sent a fax to ABP asking for an urgent meeting to discuss a temporary method of grout disposal involving the use of skips. A meeting was held the same day. At that meeting David Holt of Southampton Container Terminals (“SCT”) suggested a location for the skips at the back of SCT’s offices. Mr. Voorhuis, the project manager, said “It sounds perfect”. Ms. Bellardinelli’s note of the meeting records that Mr. Voorhuis, Mr. Holt and she visited the potential storage site. It says “The site looks perfect for our plans”. The whole of Ms. Bellardinelli’s notes appear to be conscientiously and efficiently prepared. I am satisfied that the expression “The site looks perfect for our plans” records something that was said by one of those persons, presumably Mr. Voorhuis. The evidence given by Mr. van der Eecken in his witness statement was that the area was suitable as a temporary storage area, but it was more than 500m away from the VHP locations and it was therefore not possible to pump the grout effluent. I am satisfied from Mr. Brown’s plan GWB1 that the area in question was located largely, about four-fifths, within the 500m radius. In examination in chief (day 3, page 87), Mr. van der Eecken gave a different reason why that area was unsuitable for pumping. He said:

By the same day, 24^th April 2002, there had already been allocated to HSS an area (“area (2)”) averaging about 80m behind the quay wall as a grout batching plant area. HSS had requested of ABP an area of 400m² for the grout batching plant. Mr. Brown gave evidence that the extent of that area was 2500m² which, he said, would have been adequate for a far larger number of skips than were actually used by HSS. And, he said, there was an unused underground conduit from the quay wall to that area which could have been used to service the area. ABP had pleaded that the area was 2500m² and had been available since 22^nd April 2001. HSS pleaded in response that area (2) could not be used and was impractical as a stock pile area. Mr. van der Eecken in his witness statement said simply that area (2) was already occupied by HSS’s equipment. He was questioned about that. He said in cross-examination (day 4, page 8):

Minutes of a meeting between the parties held on 15^th April 2002 record, under the heading “Contractor’s requirements”:

Ms. Bellardinelli’s notes of the meeting of 24^th April (when Mr. Brown was not present) record Mr. Voorhuis as saying:

I am unable to make a finding precisely when the full 2500m² was made available to HSS. But I am satisfied that it could have been made available by 22^nd April if HSS had asked for it. It appears to have been made available when HSS did ask for it (as I infer that they did) for the additional plant and equipment some time later. HSS did not formulate their method statement for disposing of the surplus grout until 10^th May 2002.

On 26^th April it transpired that HSS were obtaining quotations for the supply of the skips which would have involved a total cost of some £250,000 for on-site handling, not including the costs of loading trucks for disposal off-site. By 7^th May, it was clear to Mr. Brown that the temporary solution of using skips was far from ideal. No permanent solution appeared to him to be forthcoming.

Mr. Brown was active in helping to solve the problem of disposal of surplus grout. On 7^th May 2002 he called a meeting on site, where Mr Voorhuis said that a possible solution was to build a retaining structure where the surplus grout could be stored. Water would be added to the surplus grout to allow it to be pumped to that retaining structure. But additional space would be required and it would have to be within pumping distance of the works. Mr. Voorhuis was estimating that there would be up to 1500m³ of surplus grout per week and that a structure with at least 3000m³ capacity was required. Mr. Brown found a solution to the problem later the same day. He made available an area behind berth 109 where there were a number of large underground tanks where surplus grout could be stored. The area was about 2km by road from berth 205. The tanks were then being used to stockpile glass cullet. HSS cleared the area of glass cullet, the tanks were excavated and the skips were then decommissioned. The volume of the tanks, however, was less than 3000m³. On 29^th July 2002, Mr. Brown made available to HSS an additional storage tank in the same area.

On 10^th May 2002 HSS for the first time sent to ABP a method statement for the disposal of surplus grout.

HSS obtained two silos from Belgium which arrived on site on 25^th May 2002. They were kept in the jetting area and the surplus grout was pumped into them. The advantages of the silos in Mr. van der Eecken’s opinion was that they occupied a relatively small amount of space on the quay wall and they could be moved by crane within the working area. In addition, HSS hired tankers and concrete mixer trucks to transport the surplus grout from the silos to the storage tanks located behind berth 109. Large quantities of water were added to the grout in the silos to ensure that it remained in liquid form prior to its transfer to the tanks. That led to additional costs.

I shall adopt Mr. Brown’s description of the process of disposal:

On 22^nd May a meeting was held to discuss who would pay for the increased costs of the grout disposal. There is a dispute about the accuracy of ABP’s minutes of that meeting. Those minutes, sent to HSS about a month after the meeting, include the following:

Whilst I do not regard this particular dispute about the minutes as important, I shall quote Ms. Bellardinelli’s minutes of the meeting. As resident site engineer for the contract she was charged with the task, among other things, of maintaining records of all site events, including keeping minutes of meetings. Although I was not referred to her minutes of this particular meeting, it is worth quoting from them at length:

In my judgment the total sum of £187,616 does cover internal transport on site.

As to the question whether ABP provided a suitable area on site within 500m of the quay for the dumping of surplus grout, ABP, in its reply and defence to counterclaim, pleaded that it provided five areas to HSS. The area numbered (4) was identified as “Adjacent SCT office”. It is shown on a plan GWB 1 exhibited to Mr. Brown’s witness statement. It is there marked in green and labelled “Grout disposal and skip area”. Mr. Brown gave evidence, which I accept, that the extent of the area was 3250m².

In my judgment, it was not a term of the contract that ABP would provide a dumping area within 500m of the injection locations. My reasons are these. First, the method of grout disposal proposed by HSS on 3^rd December 2001 gave one option as a dumping area within 500m of the injection locations, and another option where no such location was provided. Moreover, the document indicated that those solutions were “amongst others”. Second, the proposal of 3^rd December was not one of the documents listed in the acceptance dated 19^th March 2002 of the tender. Third, grout disposal was included in the contract sum, including the provisional sum. The whole contract was a package. The schedule of prices was a contractual document, being included in the form of tender. The schedule of prices provided in section A: preamble (page 114 of the contract documents) as follows:

Mr. Williamson’s argument was that the quoted sum of £77,616 was agreed on the basis that ABP would provide a suitable storage area within 500m of the grouting works for temporary storage and hardening of the grout before off-site disposal. If and to the extent that that was not provided, the additional expense would “therefore” be borne by ABP. In the circumstances, when considering quantum, it was necessary to do so on the basis that the £77,616 was subject to adjustment by reason of the non-provision of a suitable disposal area within the vicinity of the works.

The legal basis of Mr. Williamson’s argument was not stated. The claim as pleaded is under clauses 12 and 52 of the contract and “Further and in the alternative, the Engineer has wrongfully refused to certify the aforesaid sums and/or ABP has wrongfully refused to pay such sums claimed and HSS and/or DIUK claim the same by way of damages”. Clause 12 is the clause relating to unforeseeable physical conditions or artificial obstructions. It clearly has no place in the present argument. Clause 52 relates to variations. But the disposal of surplus grout, though additional to the tender, was part of the original agreement. Therefore it did not constitute a variation. There remains to be considered the further and alternative argument. It must entail a duty on the part of ABP, in the events which happened, to pay to HSS the costs incurred by HSS in disposing of surplus grout, whatever those costs might be.

The basis of the claim must, I think, be one or both of the following. First, that it was a term of the contract, sounding in damages, that ABP would provide a suitable area for grout dumping. Second, that the price of £77,616 included in the contract sum was conditional upon such provision, and that in the absence of such provision the price for grout disposal would be at large. It has not been suggested that there was any collateral contract.

The alternative method of disposal of surplus grout given by HSS on 3^rd December 2001 provided for a dumping area within the 500m limit where HSS could “discharge” the surplus grout, or a dumping area outside that limit whither a barge or trucks could be used to transport the surplus grout. At that point it is clear that a 500m limit was not a necessary feature of the method. Pumping was not mentioned, nor in my judgment was it implicit by reason of the use of the word “discharge”.

The quotation of 7^th December 2001 for £77,616 included an amount of £38,304 for pumping equipment. Whilst no mention is made of barge or trucks, in my judgment the reference to pumping equipment does not imply the 500m limit. If I am wrong about that, then I would accept that the offer was made on the basis submitted by Mr. Williamson, namely that a suitable area within that limit would be provided. I do not accept that it follows from that that if such area were not provided, the sum payable would be at large. That is inconsistent with the terms actually agreed by 19^th March 2001, the sum of £77,616 being part of the contract sum and payable on the terms of the contract. It was not suggested that the purported acceptance on those terms was a mistake.

Thus HSS’s grout disposal claim fails in point of law. But I find the 2500m² grout batching plant area to have been an area located within 500m of the jetting operation. The balance of the area, amounting to about 2100m², outside the grout batching plant I find to have been suitable for pumping surplus grout there for the purpose of dumping, though not for prolonged storage of surplus grout. HSS could have asked, but did not ask, for that area to be made available for the purpose of dumping surplus grout. That area was made available when asked for, and there is nothing to suggest that it would not have been made available when asked for if it had been asked for earlier. I am not satisfied that it would not have been. Thus if I am wrong in holding that there was no term or condition that ABP make available a suitable area within 500m of the jetting area, I am not satisfied that there was any breach of such term or non-fulfilment of that condition.

ABP seeks specific performance of an obligation on HSS to obtain a parent company guarantee of its obligations under the contract. In the alternative, it claims damages for breach of contract in not obtaining such a guarantee or an entitlement to withhold the sum of £383,456.30 being 10 per cent. of the tender total.

Clause 10 of the contract provides, so far as material:

On 27^th May 2002 HSS sent to ABP a parent company (Dredging International NV) guarantee headed “Corporate Guarantee” expressed to expire on 31^st December 2002. On 29^th July 2002 ABP (by Mr. Steele, the Engineer) wrote to HSS a letter which included the following:

Mr. Williamson’s arguments were these. First, no particular form of guarantee was annexed to the contract. It was thus for the parties to agree between them an appropriate and acceptable form of guarantee. That argument is based on a mistaken premise. A particular form of guarantee was annexed to the contract.

Second, the parent company provided a guarantee expiring on 31st December 2002 which was accepted by ABP, who paid £4,000 for it in July 2002. That satisfied HSS’s contractual obligations. If some greater obligation were imposed by clause 10 of the contract, it was expressly varied and/or waived by accepting the guarantee that was in fact provided.

Mr. Williamson did not suggest that the £4,000 was paid in settlement of any dispute. It does not appear why ABP paid that sum. Mr. Williamson did not rely on any evidence to support the propositions contained in HSS’s letter of 8^th January 2003, namely that the original guarantee was “as agreed” and that its terms were “accepted”, which are plainly contradicted by the rest of the correspondence.

There is clearly no variation or waiver. HSS is in breach of contract in failing to procure the parent company guarantee beyond the end of December 2002.

I am not prepared to order specific performance of an obligation to procure a third party to enter into a guarantee. The parent company has no obligation to enter into the guarantee. The damages claimed for breach of contract are not specified, but if and in so far as they arise out of any incapacity of HSS to satisfy any judgment for damages against it, the claimed relief seems futile. I am, however, prepared to hear counsel further on any relief that should be granted.

In the contract between HSS and Haecon (“the design contract”) HSS was referred to as the Main Contractor and Haecon was referred to as the Designer. Article 1 of the design contract provided, so far as material:

In the particulars of the part 20 claim against Haecon the claim includes the costs of any remedial work for which HSS is held responsible. HSS allege that in the event that the works are found unfit for their purpose, such unfitness was caused or contributed to by the breach of contract of Haecon under the design contract. There is an alternative allegation of negligence which has not been pursued before me.

HSS seeks from Haecon an indemnity in respect of any loss or damage that ABP may be entitled to recover from HSS or, in the alternative, a contribution or damages. The design contract does not provide for an indemnity in the event of its breach. A contribution as such cannot, in my judgment, be ordered. Any remedy must, in my judgment, be by way of damages. The measure of damages will be any loss suffered by HSS which is attributable to breach of the design contract and which would not otherwise have been suffered by HSS.

The respects in which I have found the works to have been unfit for their purpose are the following:

As to point (2), the original design spacing could not everywhere be adhered to. Revised setting out was designed by Haecon, as appears from the following evidence of Mr. Wijffels given in the course of his cross-examination by Mr. Williamson (day 6, pages 77, 79). Mr. Williamson was putting to Mr. Wijffels questions relating to the setting out of the grout columns:

As to 3(a), the workmanship was within acceptable tolerances. The design should have allowed for such tolerances. Accordingly, (a) was a matter of design. I have found that (b) was not a matter of design for which Haecon is responsible: see paragraph 276 above.

Thus I find that Haecon was in breach of the design contract in respect of items (1), (2) and (3)(a).

On the question of damages, I make the following finding as to the effect on loss of composite action of 3(b), given 3(a). I have carried out an exercise similar to that mentioned in paragraph 285 above on the hypothesis that the diameters at depth of the columns of a pair were 900mm as designed. I have taken the example of two deviations each of 130mm: the results are set out in Appendix 5. The probability of loss of composite action in that case comes out at approximately 9 per cent. It must be borne in mind that for any possibility of loss of composite action the sum of the individual deviations of the columns must be at least 200mm in this case. I find that the probability of that is rather remote. I am not satisfied that 3(a) would have caused any loss of composite action but for 3(b).