hosted by Vicnet
[Main Index.] [Wall Structures Index] [People Index.] [Abbreviations.] [Units & Currency.] [Glossary.]
Further details of issues mentioned on the main Preston No.2 page.
The first reference to the Reservoir in the RCMPC files is in a letter from Kussmaul to Monash dated 30 January 1908, giving its basic dimensions. Monash spent two days making rough calculations. He considered a design for a wall with counterforts in the shape of tall triangles, the wall plate being inclined back towards the bank. Another had a vertical face, but no toe. He priced the first at £7-10-0 per foot run and the second at £6-0-6.
Having decided that a toe was necessary, and settled on what we would now consider a conventional form, Monash then searched for the most economic spacing of counterforts and width of base plate. On 1 February 1908 he wrote to tell Kussmaul that a great deal of money could be saved by using reinforced concrete. He would like to talk to him before submitting the design and would lend his copy of Emperger's Beton und Eisen, Volume 3 to back up the proposal. JM referred to the sections on 'Mauerwerkbau' and 'Wasserbehalter', especially p.376.
Throughout February, Monash and his assistants checked his basic design and made more detailed calculations. He brought his assumptions about the weight of the earth fill into line with Kussmaul's to ensure a fair comparison of the two schemes. He also studied a description of a retaining wall in Sambor, Galicia [Sambir, Ukraine] described in Beton und Eisen Vol.3, p.130, and further modified his own design.
JM's notes include:
"Neglect friction between filling and wall.
"The calculation given of this wall is not too satisfactory, but suggests the following amendments to my former calculations.
In estimating the horizontal pressure exerted on the wall, Monash used the standard formula for the time. This predicts that, if the fill is a granular material, the back of the wall vertical, and the surface of the fill level; the horizontal pressure on the wall is:
p = w D ( 1 - sin φ) / ( 1 + sin φ ).
This expression neglects friction between wall and fill. Assuming that the rest of the theory is reasonably accurate it therefore provides an overestimate of the pressure.
The meaning of the symbols is:
w = weight per unit volume of fill
D = depth at which the pressure is to be calculated
φ = 'angle of repose' of the material
In simple terms, the angle of repose is the steepest angle at which the material can be heaped up, before particles begin to slip down the sides of the heap. It can be thought of as a measure of the ability of a bank of the material to support itself, without the aid of a retaining wall. The greater the value of φ, the smaller is the pressure on the wall.
In his initial calculations, Monash assumed φ = 35 degrees, the value for "loosely tipped ordinary loamy soil". However he argued, in the document that accompanied the tender, that the fill to be used by the Board would be "of a selected clayey nature and specially well rammed in layers having an inclination of 1 in 50 away from the wall". He took this to mean that a higher value of φ could be adopted. Starting from the assumption that the Board's engineers had designed for a "factor of stability" of 2 for the mass concrete wall, JM had analysed its statics and decided that it was based on a φ of 45 degrees. Also, he noted that there would be a bench of undisturbed ground behind the wall [see diagram at top of this page] that would reduce the horizontal pressure. He therefore considered he was being quite conservative in adopting a figure of φ = 40 degrees.
My understanding of item (6) of JM's notes above is based on the fact that JM had chosen a face-plate thickness of 6 inches for what he described as "sentimental" reasons, and on the following probabilities:
Monash had a cross-section of the reinforced concrete version superimposed in red ink on a blueprint of the MMBW's original design to highlight the saving in concrete.
In this detail the earth is to the left and the water to the right of the wall. (J. Thomas Collection.)
Although Ritchie had strong reservations about the "experiment" he must have studied the texts or had the design assessed by others. Just before the contract was signed he questioned the adequacy of the reinforcement connecting the counterforts to the base plate. Monash produced calculations to prove its adequacy, and the signing went ahead.
JM sent his drawings to Gummow Forrest & Co in Sydney, asking them to carry out an independent check of the design. [I have not been able to find their answer in RCMPC's files.] It was only when work was due to start on the wall that P T Fairway prepared the final computations and drawings for the reinforcement. JM sent copies to Ritchie, asking him to kindly remember that "I asked you to treat these drawings strictly as for the use of the Board's officers only, and not to be accessible to the public".
Return to Overview of Preston No.2 Reservoir.
[Top.]
The explanatory document addressed possible doubts about the watertightness of the proposed 3" (76mm) floor, the strength of the reinforced concrete wall, and RCMPC's credentials as designer and builder. It mentioned the lining of the aqueduct of the Prospect Reservoir in Sydney with plates only 2 inches thick, and RCMPC's own service reservoir tanks. "In Victoria we have built numbers of circular reservoirs, subject to heads of water a to 60 ft. Under heads of up to 20 ft. these reservoirs with a thickness of 3", and in some cases 4", although subjected to considerable tensions, and wind stresses, are perfectly watertight. Monier pipes made by us in Victoria only 2" thick have frequently been proved under test to be perfectly watertight under heads up to 100 ft. Numerous reservoirs, swimming basins and tanks in Europe and America have been successfully made quite watertight with reinforced concrete linings from 2" to 3" thick. These qualities arise, firstly, from the richer, stronger and denser nature of the cement concrete employed, and secondly, from the special surface treatment employed, and thirdly, from protection against any movement through thermal stresses afforded by the embedded reinforcements."
Thus JM felt able to write, in reference to the sheet of puddle clay required by the MMBW design behind the mass wall: "the 6" reinforced concrete wall will be entirely watertight under a 21 ft head, and we prefer the entire omission of the clay, on account of its liability to swell or shrink and thereby introduce uncertain stresses".
Regarding strength, JM's document cited a large retaining wall in Seattle, 37 feet high, holding up a heavily trafficked main street; a reservoir at Orange, NJ; walls of similar design at several locations in the US and Europe; and a 22 foot wall under construction by RCMPC for the Metropolitan Gas Co's works at South Melbourne. It noted that Gummow Forrest & Co were working on about £25,000 worth of walling to support the Darling Harbor foreshores in Sydney together with wharf frontage roads. The document also drew attention to RCMPC's great experience in reinforced concrete design in connection with buildings.
Index of JM's Building Projects.
In the dispute between RCMPC and the MMBW this document was cited as proof that Monash had misled the Board by assuring them that a sound wall could be built in reinforced concrete and that the reservoir would be watertight.
Return to Overview of Preston No.2 Reservoir.
[Top.]
According to The Argus the next lowest tender after RCMPC's was that of James Moore & Sons at £30,910. Taken with the statement in The Age that the lowest mass concrete price was £31,361, this implies that Moore & Sons also put in an alternative proposal in reinforced concrete. However, Monash stated during the Arbitration proceedings that only one other reinforced concrete tender had been submitted, by Reid Bros & Russell. The MMBW records in PROV hold the answer to this conundrum. RB&R appear to have been iron and steel merchants, so perhaps they were to supply the reinforcement to Moore & Sons as general contractors.
McClelland's comment during the design stage suggests that a competitor prepared a design using expanded metal reinforcement. Unfortunately, there is nothing in the RCMPC records to indicate which engineer responsible for this.
Return to Overview of Preston No.2 Reservoir.
On 25th September 1908, Monash postponed a trip to Adelaide, explaining: "An unfortunate mishap at the Preston Reservoir works is the cause of this. Viewed in a proper light it is really of trifling significance, but in view of the critical attitude of the lay members of the Board towards our work, it assumes important proportions, and the Chairman of the Board has asked me particularly not to leave Melbourne for the next few days. The trouble arose from the pouring of liquid mud to a head of nearly 15 feet behind a wall less than 14 days old, with the inevitable result that in two days the faceplate in the lower part of the wall tore away from the counterfort and allowed the liquid slurry to disgorge itself under the toe of the wall." Carre Riddell also called Calder Oliver back from his annual leave.
As we have seen, Monash based his design on the formula
p = w D ( 1 - sin φ) / ( 1 + sin φ ).
The trigonometrical expression is a measure of the 'rigidity' of the soil, sometimes written K. It is quite sensitive to the value adopted for φ, being 0.17 for φ = 45°, 0.22 for 40°, and 0.27 for 35°. (The accuracy of the theory does not justify the use of more than two significant figures.) If the earth placed immediately behind the wall prior to the breakaway was indeed "liquid mud" or "slurry", the value of φ would be much lower than 35°, placing a much greater pressure on the wall.
Within a few days of the 'breakaway', JM had prepared material to convince Ritchie that his design was basically sound. In an unofficial letter he provided detailed calculations proving the strength of the "face plate" (i.e. a typical panel between counterforts). However, he still used 40 degrees for his value of φ. He calculated the maximum "shear" [or diagonal tension] stress in the panel as 32 psi and compared this with the then current, but optimistic, permissible value of 65 psi. Regarding the connection of the face plate to the counterfort, he depended on two 3/8-inch reinforcing bars (referred to as "tie backs") per foot of depth. These he calculated to be working near the permissible stress of 17,000 psi, but he argued that the tensile strength of the concrete would provide further safety.
Two days later, JM informed Ritchie that the connection of the face plate to the base plate would be strengthened by adding four bars 3/8" to ½" diameter and about 12" long in each bay to act as dowels between the face plate and the base plate. Also, a 6" agricultural drain would be installed behind the wall to ensure that hydrostatic pressure could not build up.
The damage had occurred over a 60-foot length of the North wall. The debris was cleared, and the face plate rebuilt. This length was then back-filled with "rubble stone filling hand packed instead of earth" for the first 12 feet of height. This would exert less outward pressure on the wall and would not retain moisture, thus ensuring that full hydrostatic pressure could not develop. For the same reason it was decided to include agricultural drains behind the length of wall still to be built. These drains would gather any water that seeped through the wall, or downward from the surface, and conduct it to inspection chambers.
Return to Overview of Preston No.2 Reservoir.
[Top.]
The first reference to problems with the walls comes in a sheet of cryptic pencil notes dated 19 October 1908. It refers to cracks and shows that some panels in the walls were out of plumb by 1¼". The experience of the 'breakaway' had shown Monash that if it became very wet, the clayey material extracted from the bottom of the reservoir and piled behind the walls could exert a pressure much greater than he (and, it seems, the Board's engineers) had estimated. News that the walls were either flexing or tilting sent him back once again to his calculations. He also set his assistant engineers, J A Laing and H G Jenkinson to carry out completely independent and thoughtful investigations of the pressure on the wall, the resulting stresses, and the factor of safety against overturning.
Jenkinson assumed the weight of filling to be 130 pounds per cubic foot and φ = 20°. He found the factor of safety "much reduced". As explained above, Monash had adopted w = 120 pcf and φ = 40° for the tender design, resulting in a pressure at depth D of p = 26 D pounds per square foot. HGJ's figures give p = 64 D psf, or 2.4 times JM's estimate. (The expression for pressure of water is simply p = w D. Taking w as 62.4 pcf for fresh water gives p = 62.4 D psf.) The 'factor of safety' for the main reinforcing bars was reduced from 4.2 to 1.7, and that for overturning from 2.7 to 1.2.
Laing attempted an analysis taking account of the fact that where the wall projected furthest above the natural surface, it was still founded at least 3 feet below it. He allowed for the fact that Rankine recommended adoption of φ = 17° for thoroughly wet clay, and that if the wall were porous, the clay could remain wetted to a high level. He arrived at a factor of safety of only 1.17, with the thrust-line lying outside the 'middle third', implying that the rear of the base slab would lose contact with the foundation.
Actually, the wall seems to have stopped moving by the end of the maintenance period. Smith (1990, p.247) states: "If the face of the wall is to be exposed then general practice is to provide it with a small backward batter of about 1 in 50 in order to compensate for any slight forward tilting of the wall". Reports on the worst forward movement at Preston No.2 vary from 3" in 20 ft (The Argus) to 4" (Nolan), and at one point during testing 5". These represent tilts of 1 in 80, 1 in 60, and 1 in 48. The fact that in places, as the toe of the wall pressed into the clay, the edge of the adjacent floor slab rose suggests there might have been some 'failure', or plastic flow of the clay, and this is what worried Coane. Monash argued, however, that some compaction of the clay would occur under the toe, with consequent decrease in moisture content, and that this would increase its strength.
Return to Overview of Preston No.2 Reservoir.
[Top.]
The idea of omitting the sheet of puddle clay behind the wall appears to have come from Monash. As we have seen, the explanatory document accompanying the tender contained an assurance that the reinforced wall would be watertight. In addition to saving the MMBW £400, omission of the puddle would have had the effect of simplifying and speeding McClelland's work, and hence suiting both him and RCMPC. Ritchie informed JM on 7 September 1908 that the Board had approved the idea.
However, the 'breakaway' incident seems to have caused Ritchie to doubt all JM's assurances and the latter wrote to him on 3 October 1908: "Dear Mr Ritchie. I sent up to you this morning copy of Handbook giving particulars of Service Reservoirs & Standpipes. Several have been built since Handbook was published. Herewith is photo of the Standpipe at Corowa, which kindly return as soon as done with. Enclosed sheet gives particulars of the Shell thicknesses. All these Structures are absolutely watertight; note the small thickness near the top. In only two cases is the thickness of Preston at its extreme depth exceeded; but in these cases the actual shell tension was considerable. At Preston there is no 'shell tension'. Note also that all our Goulburn Valley Subways (inverted siphons) carried out under Mr Dethridge are only 5 inches to 6 inches thick. They work under heads up to 18'-6". There is no justification for any attack on the watertightness of a 6 inch reinforced plate. In Emperger's Handbook (which Mr Kussmaul has), there are numerous European examples e.g. p.399, 445, 447, 458, 401 [?] of Vol. III. See also our Preston type p.376 ibid. Yours in haste, John Monash."
The "enclosed sheet" gave the depth of water, and wall thicknesses at the top and bottom, for tanks at Gisborne, Wunghnu, Pyramid Hill, Bairnsdale, Magill (SA), Richmond Match Factory, and Caldermeade (by RCMPC) and at Corowa and Kiama in NSW (by Gummow Forest & Co).
Return to Overview of Preston No.2 Reservoir.
[Top.]
A pattern of joints was specified for the concrete floor of the reservoir to divide it into sections about 200 feet square, allowing for the shrinkage that occurs after concrete is cast - and for expansion and contraction due to subsequent fluctuations in temperature. To render them watertight the joints were to be filled with asphalt. Since 1905, JM had employed Neuchatel Asphalte on several building projects. However, he gained the impression that they were not interested in the Preston job, and opted for a small supplier, industrial chemist Augustus Wolskel. This resulted in a pained protest from Neuchatel when they discovered the circumstances. JM told Wolskel that some 1000 yards of joint would be required, 1 to 1.5 inches wide and through the full 3" thickness of the slab.
After the decision to leave the wall without backfill for 6 weeks, JM persuaded Ritchie that the additional exposure to the sun might cause movement and lead to cracking, so that the wall also should be provided with a number of expansion joints. In addition, complex joints were needed between the reinforced walls and the mass concrete structures for the inlet; outlet, and 'scour'.
Although the contract made provision for joints in the floor, it was not until October 1908 that Oliver, Kussmaul, and Ritchie, after a meeting with JM, finally agreed to their use and promised to pay for them! Kussmaul had specified the composition of the filler in his original scheme, but had later given contractors the right to submit alternatives for approval. JM put forward his proposal in December 1908. On 2 February 1909, Monash reminded Ritchie that he had still not given his definite approval. On 3 February the latter replied that he did want to do so until the joints had been in use for some time!
However, Wolskel was intent on making up his own mind on the correct amount of bitumen flux to add to the asphalt, and was steadily researching it. On 10 February, JM told him: "The delay is very seriously embarrassing us not only with the Officers of the MMBW, but with the progress of the earthwork". Daily promises of a start had been made, but not kept. Wolskel replied with a lecture: "I wish to advise you that no unnecessary delay is taking place … The portion now in hand … is of the utmost consequence, and upon its success depends in large measure the stability of the subsequent work. It is therefore, essential that it be carried out in the most thorough and careful manner, and the man now employed is doing this as rapidly as circumstances permit. I appreciate your desire of not wishing to make unnecessary complaints, and can assure you I am fully seized with the urgency of the work … I am sorry if anyone has made daily promises in my name as to the commencement of the filling-in work - no one had authority to do so. [It] cannot be proceeded with until the refining of the different materials is completed: this occupies somewhere about the time that one man would take to perform the preliminary work." He estimated that joint-filling would take about a fortnight, and explained: "The trouble in this, as in so many like matters, is that it has to be dealt with as part of, and at the end of, a large contract when sufficient time is not available; however I know quite well these matters are often beyond control. While I will do all I can to expedite your work, I must express my regret at having taken it up if it is to be an annoyance to you, or the cause of any embarrassment."
Return to Overview of Preston No.2 Reservoir.
[Top.]
Fig.6. Inlet structure. University of Melbourne Archives, Reinforced Concrete & Monier Pipe Construction Co. Collection, GPNB/1167.
Fig.7. Scour (rapid emptying) structure. University of Melbourne Archives, Reinforced Concrete & Monier Pipe Construction Co. Collection, GPNB/1166.
Fig.8 As water is let into the reservoir, it swirls around the circular outlet in the floor. University of Melbourne Archives, Reinforced Concrete & Monier Pipe Construction Co. Collection, GPNB/1180.
Preston No.2 Reservoir links:
[Preston No.2 Overview.] [Further details of controversies.]
JM website links:
[Main JM Index.] [Brief Overview of Projects.] [People Index.]
[Abbreviations.] [Units & Currency.] [Glossary.]
