Origin Energy LPG Ltd v BestCare Foods Ltd

Judgment 1MACFARLAN JA: I agree with Hoeben JA. 2HOEBEN JA: Nature of Appeal At about 7.55pm on 25 January 2003 a pet food factory at Gunnedah was destroyed by fire and a massive explosion. The factory was owned and operated by the respondents, BestCare Foods Ltd (BestCare) and BestCare Food (Sales) Pty Ltd (Sales). It was common ground that the explosion and fire resulted from the leakage of liquefied petroleum gas (LPG) into the factory which then ignited. 3The respondents brought proceedings against the appellants on the basis that they had supplied and constructed the gas installation with related equipment which was used to supply LPG for use in the factory. 4It was the respondents' case that the liability of the appellants arose out of their breach of common law, statutory and contractual duties of care in respect of the construction, maintenance and inspection of the gas installation. They also claimed relief under ss 71(1) and 74(1) Trade Practices Act 1995 (Cth) for breaches respectively of implied terms that the goods supplied were not of merchantable quality and that the services rendered were not reasonably fit for the purpose of operating a gas storage facility and for the provision of gas to the manufacturing business. 5In a trial which lasted for 46 days between 18 October 2010 and 17 March 2011, the respondents were successful in establishing liability. The appellants have appealed. The appellants accept the existence of a duty of care and further accept that they breached that duty of care. The finding which they challenge is that any breach on their part caused the explosion and fire. 6This appeal is restricted to the issue of liability, with particular focus on causation. Factual Background 7Between 1994 and June 2001, Bayer Australia Ltd (Bayer) carried on business as a pet food manufacturer at Gunnedah. Under a contract made 12 August 1994 with Bayer, Boral Gas (NSW) Pty Limited (Boral) supplied and constructed a gas installation with related equipment for the factory. The gas installation and related equipment remained the property of Boral and was used by Boral to supply LPG to Bayer for the operation of the factory. 8On 29 June 2001 BestCare bought the factory from Bayer as an ongoing concern and commenced operations. At about that time, Origin Energy LPG Limited took over Boral's operations and became its successor in title to the installations. On or about 15 August 2001 Origin Energy Retail Ltd (Origin) entered into an agreement with BestCare for the supply of LPG, an arrangement which continued until the explosion on 25 January 2003. Boral and Origin are the appellants. To the extent that it was necessary to separate the appellants, Origin Energy LPG Ltd was referred to as "Boral" in the proceedings. 9It was the respondents' case at trial that the gas leakage and subsequent explosion and fire were caused by the failure of the first stage regulator (FSR) by reason of the deterioration of internal components, namely the disc holder and sealing disc. The consequence was that the FSR failed to provide a proper seal against the high inward pressure of LPG flowing from the higher pressure storage tank and allowed gas, at high pressure, to flow downstream. In these circumstances the downstream line became over-pressurised, the filter-boxes on the dryer gas line were subjected to pressure beyond their capacity and fractured, thereby allowing gas to leak into the factory, where it ignited. 10The respondents claimed that at the time of the installation of the gas system, the appellants wrongly failed to provide over-pressure protection to the FSR by means of an over-pressure shutoff device (OPSO), which would have operated to prevent over-pressurisation of the downstream line. An OPSO when activated shuts off the flow of gas at the inlet of the regulator. It was common ground that over-pressure protection must be provided where, as in this case, the inlet pressure to the FSR exceeded the pressure rating of downstream equipment. Its purpose was to ensure that the rated working pressure of this equipment was not exceeded. In this case the storage tank/inward pressure was 850 kPa and the downstream equipment was set to operate at a pressure of 140 kPa. 11Further allegations made by the respondents were that the appellants wrongly failed to ensure that the gas installation was suitable and safe for use and failed to carry out regular inspections and maintenance which would have led to the discovery of the absence of an OPSO at the FSR and its installation. 12At trial, the appellants denied liability on the basis that the respondents had failed to prove their case that the explosion and fire resulted from over-pressurisation and leakage through the dryer filters. The breaches of common law and statutory duties and contractual obligations were denied. The appellants' substantive case (which was called the "pressure wave theory") was that the FSR failed after and not before the explosion and that the explosion generated a pressure wave, which struck the FSR with such force as to damage its disc and disc holder, components necessary for its operation. 13The appellants claimed that the respondents were guilty of contributory negligence in that they failed to isolate the LPG storage tank when the factory was closed down and unattended, and that they themselves wrongly failed to install over-pressure protection at the FSR. 14At trial the primary judge rejected the "pressure wave theory". In the appeal, the appellants did not rely upon the theory. The appeal was focused on what the appellants said was the respondents' failure to prove the causation component of their case. The only aspect of contributory negligence pursued by the appellants was that the respondents had failed to isolate the LPG storage tank when the factory was closed down and unattended. 15The LPG was stored on site in a 43,000 litre pressurised tank. The pressure in the tank was in the range of 850 kPa. The appliances used in the factory were not designed to withstand such a high pressure, nor were the gas lines leading from the tank to those appliances. Consequently, as is normally the case in installations using LPG stored under high pressure, a first stage pressure reducing regulator was fitted immediately after the tank. This was the FSR previously referred to. Its function was to regulate the pressure of gas flowing into the downstream (in the direction away from the tank) gas lines by reducing the pressure from a high upstream (towards the tank) pressure to a lower downstream pressure. 16The Gunnedah factory manufactured dry pet food by two different methods: baking and extrusion. The extrusion operations and baking operations took place in different parts of the factory. The extrusion operations took place in a building referred to as the "extrusion plant", which included a boiler room. The baking operations were housed in a room to the north of the extrusion plant marked as the "biscuit baking room" in the diagram used at trial. 17The explosion occurred at approximately 7.55pm on 25 January 2003 which was the Saturday of the Australia Day long weekend. The factory had operated on 24 January and baking operations had been shut down at around 9.30pm. Extrusion operations had been shut down earlier in the day. Although the factory did not operate on Saturday 25 January 2003, some employees worked at the factory that day packing finished product or doing paperwork. The last of those employees left the factory at around 1.30pm. 18It was common ground that the initial explosion that occurred at the factory on 25 January 2003 was an LPG explosion. The size of the explosion, or sequence of explosions was not agreed. 19LPG forms an explosive mixture in air at concentrations between 2.1 percent and 9.5 percent (by volume). A mixture containing a concentration below 2.1 percent or above 9.5 percent will not explode when exposed to an ignition source. An LPG/air mixture within that range will explode when exposed to an ignition source. LPG vapour is heavier than air and therefore any LPG escape will flow downwards and may accumulate in low-lying areas. Hydrocarbon gases in LPG are odourless so a stenching agent is added as a safety measure to give LPG its characteristic gas smell to aid in the quick detection of leaks. 20The metric unit for the measurement of pressure is kilopascals (kPa). Ambient air pressure is normally around 100 kPa. The pressure of gas in a tank or pipeline is expressed in kilopascals above ambient air pressure. For example, the statement that tank pressure was 850kPa means that the pressure of gas inside the tank was 850 kPa above ambient air pressure (or around 950 kPa in absolute pressure). 21At trial and in the appeal, considerable assistance was provided by a model and a schematic diagram which showed the layout of the gas installation at the Gunnedah factory. There were two other documents to which reference needs to be made: (a) Exhibit U was a photo montage that showed the components of the 43,000 litre tank. (b) Exhibit V was a photo montage that showed a number of pieces of equipment that were relevant to the competing theories about the cause of the explosion. The photographs in Exhibit V were positioned around the schematic diagram to which reference has already been made. Regrettably, Exhibits U and V cannot be reproduced in this judgment. A copy of the schematic diagram, however, has been attached. 22Starting at the 43,000 litre tank, the gas installation can be described as follows: (i) Immediately under the tank was a valve or tap (item 1 on the diagram) referred to there as the "LP gas vapour outlet". If the tap were turned off, it stopped the flow of LPG out of the tank. (ii) After the vapour outlet was the FSR. (iii) Immediately downstream of the FSR was a Bourdon tube pressure gauge. This was the first of two Bourdon tube pressure gauges considered, referred to as the FSR Bourdon tube pressure gauge. The function of this gauge was to display the pressure of gas in the line. It is item 3 on the schematic diagram. (iv) Downstream from the pressure valve was a filter box. (v) Downstream from the filter box was a further manually operated isolation valve. This isolation valve is not shown in the schematic diagram. This valve was closed after the explosion. (vi) Downstream from that isolation valve was a gas meter (to measure the volume of gas consumed). (vii) After the gas meter there was a fork in the gas line, one branch of which led to the extrusion plant and the other to the baking operations in the biscuit baking room. There was a further isolation valve on the branch leading to the baking plant. The isolation valve is numbered six on the schematic diagram. (viii) After leaving the enclosure around the LPG tank, both gas lines ran underground a distance of about 70 metres to the factory buildings. (ix) Beginning with the line that ran to the extrusion plant, it served both the boiler and the extrusion dryer. The boiler was on ground level and the extrusion dryer was elevated on a mezzanine level, partly above the boiler. The dryer was a long box-shaped oven with doors down each side. It had two heated zones. The western end was designated zone 1 and the eastern end, zone 2. There was also a Bourdon tube pressure gauge fitted to the gas line serving the extrusion dryer, not shown in the schematic diagram. This was the second of the Bourdon tube pressure gauges referred to as the "dryer Bourdon tube pressure gauge". (x) The line that ran to the baking plant served the baking oven, which was a long oven through which biscuits were conveyed and heated. It ran east-west inside and had three burners, each of which was served by a separate gas train. 23I incorporate into this judgment pars [21] - [53] of the primary judgment (BestCare Foods Ltd & Anor v Origin Energy LPG Ltd (formerly Boral Gas (NSW) Pty Ltd) & Anor [2011] NSWSC 908) which were not in dispute in the appeal. 24In order to understand the issues raised by the appeal, it is necessary to set out in some detail how the FSR operated. This was described by the primary judge as follows: "54 The plaintiffs claimed that the fire and explosion were caused by the failure of the FSR. An understanding of the evidence requires an explanation of the operation of relevant components of the FSR namely, the disc holder, the sealing disc, and the seat ring. 55 The function of the FSR is to control the pressure at which LPG from the storage tank flows downstream. The LPG flows at high pressure into the FSR through an orifice within the seat ring. The disc forms part of the sealing surface within the FSR and is held in place against the seat ring by the disc holder. Any loss of its ability to seal against the seat ring could allow LPG to flow downstream through the second stage pipe work at an unregulated pressure. 56 When there is no demand for LPG for the downstream appliances, and the appliances are shut off, the FSR diaphragm causes the disc holder to push the disc against the seat ring to act as a seal to shut off the inward flow. When there is demand downstream the disc holder and disc are moved off the seat ring to allow LPG to flow in through the orifice to the outlet side of the regulator, and thence downstream at the set pressure. 57 The plaintiffs claim that the failure of the FSR was caused by the fracturing of the disc holder which then forced the disc onto the seat ring where it was damaged with the consequence that there was no complete seal against the seat ring. The FSR was then unable to regulate the pressure and, there being no OPSO, LPG flowed downstream at high pressure. In these circumstances, the dryer filters (the filters) were unable to withstand the over-pressure, and allowed gas to leak into the factory where it ignited. 58 The FSR is fitted with an orifice which is termed the seat ring. Its edge is rounded. The surface area of the end of the orifice forms an effective seal against the disc. 59 It is also necessary to understand the meaning of the term "wadding". Mr Arthur Donnelley (Ex AAG) explained: "9.15 This is a term applied to the non-metallic sealing disc when a circular incision is made in the sealing face of the sealing disc by the disc having been forced onto the gas orifice inside the regulator. Such circular marks are common on regulator sealing discs which have been in service for a long time. 9.16 The term derives from a 'wad punch' which is a tubular punch with a sharp bevelled edge on its end and is commonly used in leatherwork or for punching holes in gasket material. ... 9.23 Wadding of the regulator sealing disc is the most common cause of downstream excess pressure problem." And (Ex AAH): "6. ... Wadding is an indentation on the sealing disc caused by the sealing disc repeatedly contacting the sealing ring. It has the effect of compromising the sealing ability of the disc ... Once wadding occurs it makes the disc more prone to accumulate dust and dirt particles and other foreign bodies thus further compromising the ability of the disc to seal ... 7. A wadded seal will cause, and in this case did cause, an overpressure situation downstream of the regulator ... when the gas demand ceases or the appliances are shut off the downstream pressure will continue to build up in the closed pipeline system ..." " 25It is also necessary to say something more about some of the individual items on the schematic diagram. 26The manual isolation valve (the LPG gas outlet on the schematic diagram), if activated, prevented downstream flow of gas. It was common ground that if this valve had been turned off at the time of the incident, a failure of the FSR would not have caused the explosion. 27It was common ground that the FSR was not fitted with over-pressure protection. This is a safety device to prevent over-pressurisation of the downstream gas lines, in the event of a failure of the FSR. The manufacturer's instructions required that it be installed and it was not contentious that an appropriate form of over-pressure protection for this gas system was an OPSO. It was common ground that good practice and relevant Australian standards, as well as the manufacturer's instructions, required the installation of over-pressure protection on this gas system. It was also common ground that over-pressure protection was not installed. 28The OPSO should have been installed downstream of the FSR. The OPSO would have operated such that if the FSR failed to regulate pressure so that the pressure of the gas downstream of the FSR exceeded the set pressure, the OPSO would be triggered and a barrier would be mechanically forced into the gas line ceasing all flow of gas downstream from that point. 29Reference is made to a number of gas filters or "filter boxes". The function of a filter box is to remove impurities from gas flowing through it. An example of an impurity is a fragment of metal that comes away from the inside of the piping. 30Working downstream from the tank, the first filter box was located immediately downstream from the FSR. When that filter box was opened and inspected after the explosion, it was found to contain part of the disc holder and one of the annular fragments of the sealing disc from the FSR. 31The FSR Bourdon tube pressure gauge was immediately downstream from the FSR. The workings of the Bourdon tube pressure gauge contain a tube known as a "Bourdon tube", curved into a circular shape. Gas is able to enter the tube and distends the tube. The amount by which the tube is distended depends on the pressure of the gas, allowing the pressure to be measured. The Bourdon tube is designed to function elastically, i.e., after being distended by a higher pressure, it is designed to revert back to its original shape. 32After the explosion, the needle on the FSR Bourdon tube pressure gauge was found to have gone full circle and was resting against the reverse side of the zero peg. When the gauge was opened, it was found that the Bourdon tube inside was bulged and/or bent out of its original shape. 33The extrusion dryer was located on a mezzanine level within the extrusion building. It was a long, box-shaped oven. It had two burners that consumed LPG: a zone 1 burner and a zone 2 burner. Each burner was fed by a gas line that included (from upstream to downstream) a ball valve, a filter box and a regulator. Each of the two filter boxes was rated to withstand a pressure of 100 kPa. 34The filter boxes were fitted immediately before the regulators. Each regulator had a built in OPSO. At the hearing the respondents argued that the filter boxes had been the source of the leak of gas into the factory before the explosion. The filter boxes were damaged in the explosion and thus it was not possible to test them. Pressure testing was carried out on a test filter. The results of that testing and the inferences to be drawn therefrom formed an important part of the appeal. 35There was a Bourdon tube pressure gauge fitted to the gas line feeding the extrusion dryer known as the "dryer Bourdon tube pressure gauge". At trial and in the appeal, there was a dispute as to the state (bulged, deformed, bent or other) of the dryer Bourdon tube. The needle of the gauge was on the correct side of the zero peg but was displaced from the peg by 5 degrees. 36As indicated, the baking plant contained a long oven through which biscuits were conveyed. The oven ran east-west and had three burners. Each of the burners in the baking oven was served by a gas train which included (from upstream to downstream) a valve, a filter box and a regulator. Each of the three regulators was fitted with a built-in OPSO. Two of the three filter boxes were rated to 240 kPa and the third was rated to 450 kPa. The gas trains for the baking oven are shown in the schematic diagram. After the explosion, the two lower rated OPSOs on the baking line were found to have tripped and there is a dispute as to the significance of this. Evidence and Findings at Trial 37At trial the expertise of almost all of the respondents' experts was challenged. That challenge was not maintained in the appeal. Because the evidence of these experts will be more closely examined in the context of the issues raised by the appeal, it is useful to identify them and their areas of expertise and to briefly summarise the effect of their evidence. 38Mr Kerruish was an important witness for the respondents. He retired in March 2009 as a Senior Safety Inspector for WorkCover, NSW. In 1968 he was awarded a Certificate of Mechanical Engineering from Granville Technical College and in 1978 he obtained an unrestricted Gas Installer's Licence. In 1987 he was employed by the Department of Industrial Relations (later known as "WorkCover") as a Boiler Inspector. Subsequently, he was employed in advising and training on the design, operation and inspection of boilers, pressure vessels and pressure equipment. He gained experience in the investigation of accidents involving LPG and possible design failures in the design and installation of LPG plants and facilities in numerous locations. He said that his speciality was in gas boilers and pressure vessels. He did not claim any expertise in metallurgy or in the investigation of the origin of fires and explosions. 39Mr Kerruish in his capacity as a WorkCover Inspector visited the site of the explosion on 3 February 2003 for about two weeks and then again on 8 July 2003. In his visits he inspected the damage in all parts of the factory and the site. On completion of his investigations, he formed the opinion that the source of the gas leak which caused the explosion was in the area of the extruder plant. He located the pipe work connecting the gas trains servicing the dryer and noticed that the second stage regulators were burnt and each gas train was substantially damaged, the train for zone 2 more so than for zone 1. These trains and the appliances were removed for testing. Since he considered that a possible cause of the over-pressurisation was the failure of the FSR, this also was removed for testing. 40On 11 and 12 June 2003 Mr Kerruish tested the FSR at TestSafe and in his report of 25 September 2009 said: "12.7 ... It then became obvious why the regulating of the outlet pressure of the FSR was not taking place. The rubber seat of the disc holder had been "wadded" and the central plug from the seat trapped in the orifice .... 12.8 A portion of the outer section of the rubber seat was found lodged in the outlet side of the regulator and similarly the aluminium disc holder had been shattered and only a portion of it remained in the regulator ... 12.9 The remaining piece of the disc holder, rubber seat and orifice containing the wadded piece of rubber seating were removed and secured ... 12.11 Visual examination of the fracture surfaces of the disc holder appeared to me to be a fatigue type failure due to repeated pressurisation and release. When the disc holder fractured, this allowed an increased force to be applied to the nitrile rubber disc which was already partially indented from repeated seating onto the orifice. When the extra force was applied, the central section wadded out and the seat, frozen through the throttling of gas past the fractured disc holder, failed in a tensile manner and fragmented. ... 12.13 This allowed me to positively establish that over-pressurisation of the downstream reticulation pipe work downstream of the FSR was caused by a failure of the disc holder in the FSR. 12.14 My primary opinion of the cause of the over-pressurisation was wadding (or cookie cutting) of the seat disc and associated stressing of the seat disc holder to the point where it eventually failed." 41Mr Kerruish was of the opinion that while it was conceivable that a sudden over-pressure event could trigger the defining failure of the disc and seat disc, he noted that a sudden event was not necessary. He said in that same report: "19.15 I do not agree that this explosion was as a result of a single over-pressure event. It is my opinion that this occurred as the result of gradual deterioration of the seat disc. That the final failure came about when the seated disc failed to regulate the downstream pressure and the resulting over-pressurisation led to the ultimate failure of the disc and disc housing". 42Mr Kerruish was confirmed in that opinion as follows: "19.9 The other reason that confirms my view that it was not a single over-pressure event, is that with 2.3 - 3.4 tonnes of LPG unaccounted for and the gas supply simply being cut from the reticulation system within 15 minutes of the explosion and fire, there is in my view no other explanation available to account for the tonnes of gas that were missing other than the FSR fatiguing to a point where it was no longer able to carry the load and fractured during the afternoon of 25 January 2003." 43Mr Kerruish subsequently recovered missing pieces of the disc holder and disc from the filter located immediately downstream from the FSR which confirmed his view that the over-pressure event had come from the storage tank. Mr Kerruish carried out a number of tests on salvaged pieces of equipment, such as the Bourdon pressure gauges. He also tested a filter box similar to the zone 1 filter box. 44The results of this testing by Mr Kerruish formed the basis of much of the appellants' challenge to causation in this appeal. The details of that testing and the results obtained are set out at [112] hereof and the paragraphs following. 45As was conceded in the appeal, the conclusion arrived at by Mr Kerruish from the testing of the filter box was incorrect. The lid on the zone 1 filter box had not suffered the same damage as the test unit. On the contrary, the lid on the zone 1 filter box had not deformed whereas that on the test unit had. 46On 8 June 2004 Mr Kerruish examined the FSR Bourdon tube pressure gauge and the dryer Bourdon tube pressure gauge. The state of the FSR Bourdon tube pressure gauge has already been described. In relation to the dryer Bourdon tube pressure gauge, Mr Kerruish noted that as well as the needle being found to be offset by five degree from the zero position, the Bourdon tube was found to be out of true roundness. Mr Kerruish concluded that excessive pressure within the tube had caused it to yield and not return to its true circular shape. In the conclave of experts he said that he would have expected the second gauge to have suffered similar damage to the first, namely bulging of the Bourdon tube, but not necessarily. 47Mr Kerruish examined the three second stage regulators from the baking oven. Apart from minor impact damage, the three gas trains for this oven did not show visual signs of fire damage. He observed that the nearest two regulators were rated to operate to a maximum of 240 kPa and had tripped, indicating that they had been subject to over-pressure. The third regulator, located at the furthest point along the gas train, was rated to operate to a maximum of 450 kPa. It had not tripped which indicated to him that once the escape of gas commenced at the point of least resistance, it was unlikely that the gas line pressure in this regulator reached 450 kPa. 48His Honour summarised the opinion of Mr Kerruish as to the sequence of events as follows: "80 ... (1) LPG leaked through the FSR past the disc which caused pressure downstream to increase. This activated the OPSOs for the second stage regulators and stopped all flow past those regulators; (2) the increased pressure caused the distortion of the disc holder which allowed pressure to increase at a greater rate downstream, which eventually caused the disc and disc holder to fail; (3) the dryer filters, not being protected by OPSOs, suffered an increase in pressure beyond their rated pressure which caused their lids to bow allowing leakage of LPG through their O-ring seals into the plant below; (4) upon failure of the FSR sealing assembly, full tank pressure, estimated at 850 kPa, was applied to a filter rated for only 100 kPa, and caused it to fail, allowing an uninterrupted flow of LPG from the storage tank into the factory." (Red 101B-K) 49Another expert relied upon by the respondents was Mr Donnelley. Mr Donnelley was a licensed plumber, gas fitter and drainer and was granted an unrestricted LPG licence in 1965. He subsequently obtained an advanced town gas licence which qualified him to work on large industrial gas installations. He was experienced in the design and building of industrial gas fired appliances and in the installation of LPG storage systems for shire councils and service stations. He had many years' experience as an investigator into gas related accidents for police, loss adjusters and insurance companies, and for many years was a member of relevant professional associations. He died in August 2009 aged 72 and his evidence was contained in his reports of 9 April 2008 and 27 July 2009. 50Mr Donnelley first visited the site on 29 January 2003 and over the next few weeks seven more times. He inspected several installations and appliances and tested the underground gas lines. He was present on 11 June 2003 with Mr Kerruish at the premises of TestSafe when the FSR was tested, dismantled and examined. 51Mr Donnelley concluded that the LPG escape had emanated from either the zone 1 or zone 2 filters, or from both of them, in that gas would readily cascade down over the dryer to the boiler and in time could permeate throughout the site resulting in an explosion. The primary judge summarised Mr Donnelley's opinion as to the sequence of failure of the FSR as follows: "94 ... (1) the effect of wadding compromised the ability of the disc to properly seal against the inlet pressure, thus allowing slow leakage of LPG past the disc; (2) the leakage caused over pressuring of the downstream pipe and fittings, which caused LPG to leak from the filter(s), rated at 100 kPa, into the factory; (3) the explosion occurred when the leaked gas ignited; (4) the filters fractured due to mechanical damage from the explosion. This created two open ends in the gas lines which caused a large flow of gas through, and a sudden excessive load on, the FSR; (5) the massive flow of LPG towards the open ends of the gas line caused the fracture of the disc holder." (Red 104S-Q) 52Another expert, retained by the respondents, was Dr Corderoy. He was a qualified mechanical and metallurgical engineer and fulltime consultant to Unisearch Pty Ltd at the University of New South Wales. He was unavailable to give oral evidence and his findings and opinions were not tested under cross-examination. He was retained to report to the coroner on the condition of equipment retrieved from the site, including the FSR, the filters and the Bourdon tube pressure gauges. 53Dr Corderoy examined the components of the FSR and the filters with the assistance of scanning electron micro grafts and x-ray analyses. He concluded that the fire and explosion was a consequence of the failure of the FSR by reason of the fracture of the disc holder which was of insufficient toughness to withstand the repeated force of impact against the seat during the operation of the FSR. This resulted in the initial leakage of the LPG past the disc. 54Mr Pearson was an expert called by the respondents. He was a senior fire and explosion investigator employed by TestSafe, Australia. He had been so employed for over 20 years, during which he gained extensive experience into the explosability of dust and gas explosions. He had investigated numerous incidents involving the causes of gas and dust fires and explosions. He did not claim any expertise in metallurgy, material science, the operation of LPG regulators or the mechanism of Bourdon tube pressure gauges. 55He provided two reports and gave evidence at trial. He concluded that a failure of the FSR was the only credible way to produce high gas pressures in the LPG pipes within the plant. 56Mr Pearson visited the site between 29 January 2003 and 20 March 2003 on several occasions. He found that the overall pattern of damage indicated the centre of the explosion to be somewhere inside the extrusion plant. This was accepted by the parties on the appeal. He said that there was evidence that the pipe work, including the dryer and bakery gas trains, had been subjected to high internal pressure. He concluded that the downstream line was subjected to high pressure and that the FSR might have failed progressively between the shutting down of the baking area at about 9pm on 24 January 2003 and the explosion 23 hours later. 57He also concluded that the likely location of the LPG leak was from the zone 1 and zone 2 filters caused by over-pressurisation and their rating to withstand a pressure of only 100 kPa. He said that the likely failure of the FSR which caused the over-pressurisation, was that the sealing/regulating surface of the FSR became compromised and unable to seal completely which allowed excessive pressure to be applied to the filters, resulting in their failure. Under cross-examination he agreed that he relied upon information provided to him by Mr Donnelley and by Mr Kerruish in describing matters relating to the FSR, the gauges, the filters and OPSOs and to explain faults associated with them. 58Professor Hoffman was called on behalf of the respondents. 59Professor Hoffman obtained a doctorate in material science and engineering in 1994, and was the Head of the School of Material Science and Engineering at the University of New South Wales. His speciality was in the field of materials fracture and failure and the structural integrity of materials. As part of his scientific research he had worked with high pressure gases, but did not claim expertise in gas dynamics or combustion. There was no issue that Professor Hoffman was a world authority in his field. 60The effect of the evidence of Professor Hoffman was summarised by the primary judge as follows: "117 Professor Hoffman found that there had been intergranular failure of the disc holder which indicated that the metal from which it was made was brittle, and of a low fracture toughness. His view was that the disc holder was manufactured with inappropriate material, being a brittle alloy of low fracture toughness, and/or that in the manufacturing process the disc holder was subjected to inappropriate heat treatment which resulted in low fracture toughness. He said that the failure of the disc holder was caused by fatigue crack growth followed by fast fracture, a process which resulted from its defective manufacture. 118 He observed that the disc had been punched out by the sharp edge on the seat ring, or wadded, which caused it to split; and that the first Bourdon tube pressure gauge had been overloaded and its needle pushed to the stop position; and that the filter boxes had failed as a result of brittle fracture of the aluminium boxes. 119 In his opinion, the operation of the FSR propagated fatigue cracking of the disc holder, and a subsequent fast fracture occurred which caused it to fail. As a result the FSR was unable to operate properly with the result that it pushed the disc with high pressure into the seat ring, leading to the failure of the disc. This would have led to an incomplete seal on the seat ring, meaning that the FSR would then seek to increase pressure via the disc holder. The incomplete seal allowed gas at high pressure to flow downstream. The components downstream were unable to structurally withstand the unregulated high pressure which led to their failure and the leakage of LPG which, assumedly, caused the fire." (Red 110V-111P) 61Because of his unchallenged and acknowledged expertise in the field of materials fracture and failure, the evidence of Professor Hoffman in relation to the zone 1and zone 2 filter boxes was of considerable importance. An issue in the appeal was an apparent conflict between the opinion expressed in his reports and what he said at the conclave of experts. This is a matter which will be explored further when the specific issues raised by the appeal are dealt with. 62Certainly in his report of 21 June 2010 Professor Hoffman adhered to the proposition that the zone 2 filter box could have failed through over- pressure and possible explosion. Apparently relying upon the test results obtained by Mr Kerruish, Professor Hoffman expressed the opinion that there was a high probability that the zone 2 filter box failed, from which flammable gas leaked and that leakage prior to its failure was the most likely scenario. One of the difficulties in the appeal is that Professor Hoffman was not cross-examined with respect to his evidence concerning the filter boxes. 63Professor Masri provided reports and attended the conclave of experts. He was a Professor of Mechanical Engineering in the School of Aerospace Mechanical and Mechatronic Engineering, as well as Associate Dean for Research in the Faculty of Engineering and Information Technologies at the University of Sydney. One of his specialities was thermal engineering, including combustion. He had published extensively in the field of turbulent combustion, including research in deflagrations. 64His reports were relevant to the amount of gas needed to produce an explosion of the size observed by witnesses. He was also expert in the flow of LPG. 65Mr Cox was a chartered engineer who practised as a forensic consultant with expertise in metallurgy. He was the only expert relied upon by the appellants at trial. He did not commence his investigations until 2009. He did not visit the factory. He did, however, visit the TestSafe laboratory in April 2009 with Messrs Donnelley and Kerruish and examined equipment recovered from the site, including the components of the FSR. 66He disagreed with the conclusions of Professor Hoffman and Mr Kerruish. It was he who put forward the pressure wave theory as an alternative to the respondents' explanation of how the explosion occurred. He accepted in cross-examination that on the available evidence if the pressure wave theory were rejected, the only other theory to explain the failure of the FSR was that put forward by the respondents, i.e., the FSR had failed causing over-pressurisation downstream, leakage of LPG and ultimately the explosion and fire. 67Mr Cox agreed that he had no experience in relation to the operation of FSRs. He agreed that he had no experience in fluid dynamics, aerospace shock tube technology, the design of Bourdon tube pressure gauges, the design and installation of gas systems and in the design and operation of boilers. He accepted that his opinions were based on a review of evidence many years after the event and that it would have been far preferable to have examined the accident scene firsthand rather than to rely upon photographs and statements. 68His Honour's conclusions with respect to Mr Cox were: "168 My overall impression of Mr Cox's evidence was that he ventured opinions on numerous issues without the necessary factual basis, qualifications, and experience to do so, and demonstrated an absence of scientific rigour in propounding the pressure wave theory as the probable cause of failure of the FSR. As this case showed, Mr Cox propounded a theory which was doomed to dissolve when subjected to objective analysis and scrutiny. It is difficult not to accept that had the theory been tested before it was tendered, its weaknesses would soon have been exposed. I was left entirely unconvinced of its fundamental soundness. 169 In my assessment, Mr Cox's readiness to advance opinions which lacked support served to undermine any standing he might have had as a credible and reliable witness on matters about which there was a contest, including, in particular, his challenges to the plaintiffs' witnesses on the failure of the FSR, over-pressurisation, and the cause of the fire and explosion." (Red 124J-S) 69His Honour summarised the respondents' case at trial (which largely remained the same on appeal) as follows: "182 The plaintiffs' case is that, when considered as a whole, the evidence supports the finding that the FSR failed prior to the explosion. On examination it was found that it was unable to effectively seal off the inflow of gas from the storage tank at 850 kPa, and regulate the pressure at which it flowed downstream. The common view of their witnesses was that the dryer filters, which were located before and, hence, unprotected by, the second stage regulator OPSOs and with a pressure rating of 100 kPa, yielded to high pressure and allowed leakage into the factory, and subsequent ignition. Consistent with this was the evidence that those OPSOs with an inlet pressure rating of 240 kPa had tripped, indicating that LPG at high pressure had reached the second stage regulator prior to the explosion. Consistent also was the condition of the first Bourdon tube pressure gauge as found by Mr Kerruish, which indicated to him that it had been subjected to a pressure greatly in excess of 250 kPa for which it was designed. The condition of the second gauge provided a similar indication." (Red 127M-V) 70The crucial findings made by his Honour and which were criticised in a number of respects in the appeal were: "Determination 194 In my assessment the defendants failed to negate the evidence as to the filters, second stage regulator OPSOs, and the Bourdon tube pressure gauges which supported the views of the plaintiffs' experts that over-pressurisation resulted from the failure of the FSR prior to explosion. 195 It is unnecessary to resolve the dispute between Mr Kerruish and Mr Cox as to the right conclusion after testing the zone 1 filter. Assuming, without deciding, that Mr Cox's evidence should be preferred, without more there remains open the inference that there was leaking through the zone 2 filter. Indeed, the probability of leakage from either filter before the explosion is not inconsistent with the fracture of the filter boxes by the explosion. Relevant also is the rating of the filters at 100 kPa, and the fact they were unprotected by the second stage regulator OPSOs. 196 In any event, Prof Hoffman's unchallenged opinion that it was highly probable that the zone 2 filter box failed through over-pressurisation and possible explosion is sufficient for the finding that this happened. I am satisfied that, taken with the totality of his evidence, his answer in the conclave that he had not considered the cause of failure of the zone 2 filter box was mistaken, and should not be understood as an abandonment of the views expressed in his reports of 21 September 2009 (Ex AAK), and 21 June 2010 (Ex AAL). 197 The defendants demonstrated the possibility that OPSOs may be tripped by mechanical shocks. Nevertheless, in this case, Mr Kerruish adhered to his opinion that it was over-pressurisation, not the violence of the explosion, which caused the relevant OPSOs to trip. His was an explanation he was well qualified to give, and which I accept. It is strengthened by the evidence that the zone 3 regulator OPSO, with a pressure rating of 450 kPa had not tripped. I am satisfied on the probabilities that over-pressurisation tripped the second stage regulator OPSOs prior to the explosion. 198 The evidence concerning the Bourdon tube pressure gauges, in my opinion, founds the inference that they were subject to over-pressurisation before explosion. The defendants established no alternative explanation for their condition, and left the question open. Mr Kerruish's conclusions in respect of both gauges were based on his examination of them, coupled with his experience of LPG systems and installations. I accept them. I prefer his evidence to that of Mr Cox, who agreed that he lacked similar experience. 199 The common thread through the theories of the plaintiffs' experts is that the immediate cause of the failure of the FSR was the failure of the disc to effect a seal against the inflow of LPG. The evidence established that this is what happened. To reach this conclusion it is not necessary to identify the precise cause of the failure of the disc, and the point at which the disc holder fractured. Nevertheless Prof Hoffman's evidence supports the finding, which I make, that it was highly probable that the fracture of the disc holder led to the failure of the disc, thence the unregulated flow of gas downstream. I also find that the fracture of the disc holder was caused by the defective manufacturing process as explained by Prof Hoffman. 200 It was admitted that the failure of the FSR caused over-pressurisation downstream. The evidence, taken in combination, just as a cable is made up of strands, firmly established, in my opinion, that over-pressurisation caused LPG to leak from the filters to cause the fire and explosion. Contrary to the defendants' contention, the possibility of alternative explanations for the condition in which the filters, OPSOs, gauges, and the FSR was found after the event neither contradicts nor diminishes the evidence in support of this conclusion. The defendants' strategy of doubt-mongering was not enough to undermine the plaintiffs' case, or to overcome Mr Cox's concession that if the pressure wave theory failed, the only available explanation was that advanced by the plaintiffs. 201 For these reasons I find that it was the failure of the FSR for which no over-pressure protection had been provided which caused the fire and explosion. The defendants admitted that the over-pressurisation and leakage of gas from the failure of the FSR would have been prevented by the installation of an OPSO." (Red 131M-133N) 71There were issues raised at trial as to the nature and the content of any duty of care owed by the appellants to the respondents and as to the terms of the contracts between them. The meaning and application of exclusion clauses in those contracts was also a matter of contest. None of those issues was pursued in the appeal. 72Apart from the issue of causation, the only other issue pursued by the appellants in the appeal was contributory negligence. The single ground relied upon was: "The plaintiffs failed in their common law and/or contractual duty to isolate the storage tank when the factory was closed down and left unattended on 25 January 2003." 73At trial, the appellants submitted that the duty to isolate the tank when the factory was unattended arose from the inherently dangerous nature of the installation, the requirements of clause 10.2.10 AS 1596-2002 and the respondents' contractual obligations to comply with basic safety requirements. Clause 10.2.10 of the Australian Standard provided: "When a site is unattended by trained staff, the tank is not in use and not fitted with an automatic shut-down system, the tank isolation valve shall be kept closed." 74On the appeal, the requirement that the respondents comply with contractual obligations was not pursued. 75At trial, the primary judge resolved the contributory negligence issue as follows: "317 To succeed in the claim of contributory negligence, the defendants must prove that the plaintiffs had been contributorily negligent in failing to take precautions against the risk of such harm (s 5R(1)). In deciding what a reasonable person in the plaintiffs' position would have done, and whether he exercised reasonable care in the circumstances, it is necessary to take into account the evidence of the plaintiffs' knowledge, understanding, and actions, shortly prior to the accident (Council of City of Greater Taree (par 108)). In my opinion the relevant risk of harm against which precautions might have been taken was over-pressurisation downstream from the failure of the FSR, which would have been prevented by the installation of an OPSO. Accordingly I do not accept that failure to isolate the tank was a failure to take a precaution against the risk of harm which is capable of establishing negligent conduct on the part of the plaintiffs. It follows that the defendants' claim on this ground should be dismissed. 318 Nevertheless, it is appropriate to consider the issue that isolation of the tank was a precaution to be taken by the plaintiffs against the risk of the harm. 319 On the question of knowledge, I accept the evidence of Mr Heness and Mr Goldring as truthful. I find that neither knew of the requirement of cl 10.2.10 of AS 1596-2002, or of similar information. I also infer that it was improbable that anyone else employed by the plaintiffs had this knowledge. It follows that the defendants have failed to prove that the plaintiffs knew of any instruction or information that the tank should be isolated when the factory was closed down. ... 321 In my opinion, the defendants have utterly failed to prove that it was negligent of the plaintiffs, or either of them, to have closed down the factory on 25 January 2003 without isolating the tank. To the contrary, I find that in the circumstances it was reasonable for the plaintiffs to have closed down the factory without isolating the tank, doubtless on the assumption that there was no risk of harm and/or it was safe to do so. It follows that the defendants have established no basis for apportionment under s 5S. Accordingly, their claim on this ground is rejected." (Red 75U-177M) THE APPEAL The OPSOs on the Second Stage Regulators 76As indicated, there were two OPSOs fitted to each of the second stage regulators on the dryer. There was an OPSO on the boiler, which it was agreed had been tripped by mechanical means because the boiler was, at the relevant time, isolated. Finally, there were three OPSOs on the baking oven; one fitted to each second stage regulator. His Honour's findings at [197] and the submissions of the parties were directed to the three OPSOs on the baking oven. 77The findings made by his Honour in relation to those three OPSOs were: "197 The defendants demonstrated the possibility that OPSOs may be tripped by mechanical shocks. Nevertheless, in this case, Mr Kerruish adhered to his opinion that it was over-pressurisation, not the violence of the explosion, which caused the relevant OPSOs to trip. His was an explanation he was well qualified to give, and which I accept. It is strengthened by the evidence that the zone 3 regulator OPSO, with a pressure rating of 450 kPa had not tripped. I am satisfied on the probabilities that over-pressurisation tripped the second stage regulator OPSOs prior to the explosion." (Red 132G-K) 78The appellants submitted that the findings made by the primary judge at [197] were not open to him. They submitted that once it was established (as his Honour accepted) that the OPSOs on the second stage regulators might be tripped by mechanical shock, there was no logical basis for rejecting that possibility and preferring over-pressurisation. The issue was one of scientific fact and could not be resolved by a credit finding in favour of Mr Kerruish. 79In support of those submissions, the appellants referred to the evidence of Mr Kerruish to the effect that he had examined the OPSOs in situ to ensure that the removal process did not result in an accidental tripping and that even movement in the back of a vehicle during transport could have that effect. The mechanical shock relied upon by the appellants was the explosion which would have exerted forces significantly greater than those involved in removing the OPSOs and transporting them. 80In further support of this challenge, the appellants relied on the non-controversial fact that the OPSO on the boiler had tripped in circumstances where it was clear over-pressurisation could not have been involved. (That is because the manual isolation valve on the boiler gas train had been closed prior to the explosion.) They submitted that this made it clear that substantial forces were in play, sufficient to trip the OPSOs on the baking oven. 81The appellants submitted that had over-pressurisation accounted for the tripping of the first two OPSOs rated to 240 kPa, it did not explain why the third OPSO rated to 450 kPa did not trip. They submitted that if over-pressurisation to 850 kPa had occurred, this should have tripped the third OPSO and its failure to trip counted against over-pressurisation having occurred. 82The appellants submitted that there was an issue between Mr Donnelley and Mr Kerruish as to the rating of the second stage OPSOs on the baking oven. The ratings attributed to them by Mr Kerruish (set out above) were based on labels which were affixed to the regulators. The ratings attributed to them by Mr Donnelley (of 870 kPa each) were based on the ratings in the manufacturer's pamphlet. Although his Honour implicitly accepted the assessment of Mr Kerruish, the appellants complained that he did not explain why. 83This issue was potentially important in that if Mr Donnelley were correct in rating the OPSOs at 870 kPa, over-pressurisation to 850 kPa would not have caused them to trip and it was much more likely that mechanical shock had done so. 84The appellants submitted that contrary to his Honour's finding, the post explosion "non-tripped" state of the third OPSO could not strengthen the conclusion of Mr Kerruish that over-pressurisation had caused the first two OPSOs to trip. On the contrary, it was a difficulty which stood in the way of an acceptance of his conclusion and needed to be explained. They submitted that his Honour had not resolved that issue. 85The explanation given by Mr Kerruish under cross-examination was that the over-pressurisation had been in excess of 240 kPa, but not more than 450 kPa, with the consequence that the third OPSO did not trip. He explained that this occurred because the third OPSO was the furthest away on the gas line and that the gradual over-pressurisation of the gas line from the slow failure of the sealing disc in the FSR, caused the zone 1 and zone 2 filter boxes to leak before the third OPSO was exposed to a level of over-pressurisation above 450 kPa. By way of an alternative explanation, Mr Kerruish said that devices such as OPSOs could often withstand pressures greater than their rated maximum. 86Although the theory of Mr Kerruish as to gradual over-pressurisation was contrary to that expressed by Professor Hoffman, it was consistent with the opinions of Mr Donnelley and Mr Pearson. While his Honour preferred the opinion of Professor Hoffman on that question, he did not reject the other approach. The gradual increase in pressure theory was well and truly in play. 87It was open to the appellants to challenge the conclusion of his Honour on this issue on the basis that it did not explain why the third OPSO did not trip. However, the alternative theory put forward by the appellants that mechanical shock tripped the first two OPSOs did not explain why that mechanical shock did not also trip the third OPSO. This is particularly so since the evidence of Mr Kerruish relied upon by the appellants on this issue indicated that it would not take a great deal of mechanical shock to trip such devices. 88While the issue was essentially a scientific one, there was a part to be played by his Honour's assessment of Mr Kerruish. His evidence was based on a lifetime of actual experience in the design and operation of LPG installations, including regulators and OPSOs. He had held an unrestricted gas installer's licence for over 30 years. His Honour was, therefore, entitled to give his opinion considerable weight, particularly when the only opinion to the contrary was that of Mr Cox. 89That having been said, I have concluded that there is force in the appellants' criticism of this finding by his Honour. If for other reasons it was open to his Honour to find that the respondents had established their case on causation, then the tripping of the first two OPSOs on the second stage regulators would support that conclusion. As an independent piece of evidence, however, they do not take that matter forward and provide little assistance to the position of either the appellants or the respondents. The Bourdon Tube Pressure Gauges 90The appellants submitted that the finding by the primary judge at [198] was not open to him on the evidence: "198 The evidence concerning the Bourdon tube pressure gauges, in my opinion, founds the inference that they were subject to over-pressurisation before explosion. The [appellants] established no alternative explanation for their condition, and left the question open. Mr Kerruish's conclusions in respect of both gauges were based on his examination of them, coupled with his experience of LPG systems and installations. I accept them. I prefer his evidence to that of Mr Cox, who agreed that he lacked similar experience." (Red 132M-Q) 91The appellants' main criticism of his Honour's conclusion on this issue was that his Honour did not take into account all of the evidence, in particular that of Professor Hoffman, and that his Honour misunderstood the effect of the evidence as to the Bourdon tube pressure gauges. 92It was the respondents' case that both the FSR Bourdon tube pressure gauge and the dryer Bourdon tube pressure gauge had been subject to over-pressurisation and that this supported a failure of the FSR before the explosion, otherwise there was no explanation for that over-pressurisation. The appellants submitted that what his Honour did not take into account was that the condition of the two Bourdon tube pressure gauges after the explosion was quite different. 93It was common ground that the FSR Bourdon tube pressure gauge was damaged by over-pressurisation. That was the accepted explanation for the position of the needle below the zero peg and for the bulging of the Bourdon tube. The damage to the dryer Bourdon tube pressure gauge was different. There was no bulging of the Bourdon tube and the needle was displaced about five degrees from the zero peg. The dryer Bourdon tube was also slightly "out of true roundness". 94It was the evidence of Professor Hoffman at the conclave of experts that "if the Bourdon tube pressure gauge located adjacent to the dryer was subjected to the same degree of over-pressurisation as the Bourdon tube pressure gauge located immediately downstream of the first stage regulator it would have suffered similar damage namely, bulging of the Bourdon tube". (Conclave Q.15.4) (Blue 1494S-T) 95The appellants submitted that his Honour had not taken that evidence into account. They submitted that rather than assist the respondents, the condition of the Bourdon tube pressure gauges suggested that over-pressurisation had not occurred before the explosion. This was because it was common ground that if the FSR had failed and caused the second stage pipework to be over-pressurised, both Bourdon tube pressure gauges would have been subjected to the same maximum pressure. (Conclave Q15.3) (Blue 1494N-P) 96I agree with the appellants' submission on this issue. There was a real question to be decided in relation to the Bourdon tube pressure gauges and his Honour did not properly analyse the evidence in reaching his decision. That having been said, I do not accept that the evidence on this issue supports either the appellants or the respondents. 97It was common ground that after the explosion, the manual isolation valve (item 28 in Exhibit U) upstream from the dryer Bourdon tube pressure gauge and downstream from the FSR Bourdon tube pressure gauge was closed. This had the effect of exposing the FSR Bourdon tube pressure gauge to the full tank pressure from the LPG tank for a period of between 12 and 15 minutes until the isolation valve at the tank (item 9 in Exhibit U) was closed. It was also common ground that the FSR Bourdon tube pressure gauge had been damaged by over-pressurisation. That of itself was neutral in that such damage could have occurred after the explosion in the period before the isolation valve on the tank was closed. 98Accordingly, the fact that the FSR Bourdon tube pressure gauge was damaged by over-pressurisation does not assist either side. 99In answer to Question 15.4 of the conclave as to whether having been exposed to the same pressure, the Bourdon tube pressure gauges would have suffered similar damage, the reply of Mr Kerruish was recorded as: "Yes, he would have expected it but not necessarily." When asked to explain that qualification, Mr Kerruish initially suggested that the FSR Bourdon tube pressure gauge might have been of an inferior quality. Having been shown photographs of that gauge and of the dryer Bourdon tube pressure gauge, he withdrew that explanation. 100His second explanation was that the five degree displacement of the needle from the zero peg on the dial of the dryer Bourdon tube pressure gauge meant that as a result of over-pressurisation the needle's centrepin had broken loose. Even if that explanation were correct, it is difficult to see how it would account for the final position of the needle, slightly offset as it was. This is particularly so when the slight "out of true roundness" of the dryer Bourdon tube explains the offset of the needle from the zero position. 101If that were the only evidence on the issue, it could be said that the difference in damage to the two Bourdon tube pressure gauges did not support the respondents' case. There is, however, other evidence. While the damage to the dryer Bourdon tube pressure gauge was not nearly as extensive as that suffered by the FSR Bourdon tube pressure gauge, according to Mr Kerruish the slight "out of roundness" of the dryer Bourdon tube pressure gauge was due to over-pressurisation. There was no other evidence to explain the slight "out of roundness". 102The respondents submitted that there was an explanation for the different states of the two Bourdon tubes in that they were exposed to high pressures for different lengths of time. This was because the closing of valve number 28 had the effect that the FSR Bourdon tube pressure gauge was exposed to 850 kPa for 12-15 minutes longer than the dryer Bourdon tube pressure gauge. While there is a clear difference in the length of time of exposure, no expert gave evidence as to its significance. In the absence of such evidence, it would be speculation on the part of this Court to take that into account as an explanation for the difference in damage between the two Bourdon tubes. 103There is, however, another consideration which is relevant. It is implicit in the appellants' submissions that had the FSR failed, causing over-pressurisation downstream, the over-pressurisation would have been at the level of 850 kPa, i.e. the maxiumum pressure in the tank. That proposition is controversial. In the debate concerning the significance of the tripping of the OPSOs on the second stage regulators, there was an issue as to whether, and to what extent, the pressure in the gas lines ever reached 850 kPa. A similar issue has been raised in relation to the failure of the zone 1 and zone 2 filters boxes. A difference in the damage to the Bourdon tube pressure gauges is consistent with the gradual increase in pressure in the gas lines as a result of wadding of the FSR. If it be the case that the pressure in the gas lines did not reach 850 kPa before the explosion, that would provide an explanation for the difference between the damage to the FSR Bourdon tube pressure gauge and the dryer Bourdon tube pressure gauge. While both may have been exposed to the same maximum pressure before the explosion, that pressure may have been less than 850 kPa. Following the explosion, it is clear that the FSR Bourdon tube pressure gauge was exposed to an 850 kPa level of pressure. 104It follows from the above analysis that his Honour erred in taking account of the Bourdon tube pressure gauges as an important piece of evidence supporting the respondents' case of over-pressurisation before the explosion. The evidence was at best neutral. The Dryer Gas Filters 105This section of the judgment relates to the two filter boxes at either end of the extrusion dryer. The extrusion dryer was a long box-shaped oven with doors down each side. The western end was designated zone 1 and the eastern end, zone 2. The extrusion dryer was elevated on a mezzanine level, partly above the boiler. There was a filter box in zone 1 and a filter box in zone 2. Each filter box was rated to a maximum of 100 kPa. It was the respondents' case that as a result of the failure of the FSR, increased gas pressure was introduced into the gas line thereby exposing the zone 1 and zone 2 filter boxes to pressures well in excess of their rated amounts. This in turn caused the zone 1 and zone 2 filters to fail, leaking gas down into the dryer area of the factory. Both filter boxes were damaged in the explosion, the zone 2 filter box more severely than that in zone 1. 106The primary judge's findings in relation to these filter boxes were: "195 It is unnecessary to resolve the dispute between Mr Kerruish and Mr Cox as to the right conclusion after testing the zone 1 filter. Assuming, without deciding, that Mr Cox's evidence should be preferred, without more there remains open the inference that there was leaking through the zone 2 filter. Indeed, the probability of leakage from either filter before the explosion is not inconsistent with the fracture of the filter boxes by the explosion. Relevant also is the rating of the filters at 100 kPa, and the fact they were unprotected by the second stage regulator OPSOs. 196 In any event, Prof Hoffman's unchallenged opinion that it was highly probable that the zone 2 filter box failed through over-pressurisation and possible explosion is sufficient for the finding that this happened. I am satisfied that, taken with the totality of his evidence, his answer in the conclave that he had not considered the cause of failure of the zone 2 filter box was mistaken, and should not be understood as an abandonment of the views expressed in his reports of 21 September 2009 (Ex AAK), and 21 June 2010 (Ex AAL)." (Red131Q-132E) 107The appellants submitted that contrary to his Honour's finding, it was necessary for him to resolve the issue between Mr Kerruish and Mr Cox as to the correct conclusion to be drawn after the testing of the filter by Mr Kerruish. They submitted that had his Honour properly analysed the evidence, he would have concluded that the zone 1 and zone 2 filter boxes had not failed as a result of over-pressurisation and therefore could not have been the source of the gas leak. As part of that evidentiary analysis, the appellants submitted that his Honour needed to closely examine the evidence of Professor Hoffman concerning the filter boxes and that the apparent inconsistency between his reports and his responses at the conclave of experts on that issue could not be explained by saying that he was "mistaken" in those responses. 108The appellants accepted that the challenge to his Honour's findings concerning the gas filter boxes was fundamental to their success on the appeal. If it were not possible that the gas filters had failed as a result of over-pressurisation, they submitted that the respondents had failed to establish causation. This was because the zone 1 and zone 2 gas filters were the weakest part of the system, being rated to only 100 kPa each. If they had not failed due to over-pressurisation before the explosion, over-pressurisation of the gas lines could not have occurred. 109Because of the importance of this issue in the appeal, it is necessary to set out the relevant evidence. 110In his statement of 9 April 2008 (Exhibit AAG), Mr Donnelly said of the gas filters: "10.1 During my examination of the site, I did not see any evidence of there having been a gas escape at the gas tank area, along the underground pipelines or in any remaining intact gas lines within the buildings. 10.2 As a result of my on site examinations I had concluded that LPG had escaped inside the extruder building and at high level in an area above the elevated dryer. I then went in search of the device in the elevated dryer gas trains which had the lowest pressure rating. I had seen the gas trains from the appliances in the evidence store at TestSafe in company with David Pearson, Senior Fire and Explosion Specialist at TestSafe. 10.3 I noticed that the filters on the inlet to each of the elevated dryer burner gas trains had suffered fractures due to mechanical damage during the explosion. These filters were the same as some filters that I had in my own store at my Mortlake workshop, which had a pressure rate of 1 bar or 100 kPa. Filter Testing 10.4 When the first stage regulator failed the pressure in the gas tank would have been approximately 825 kPa. On 11 August 2003 I decided to carry out a pressure test on one of the two filters in my own stock (which was exactly the same as the filters on the LPG installation except for minor differences in the inlet pipe size). Those tests revealed that the filter had a pressure rate of 100 kPa. 10.5 To pressure test the filter I enclosed the filter inside a steel box with a pipe connection to outside. The box was lined with rubber in case the filter shattered at high pressure and the rubber would allow any flying metal to be caught undamaged by contact with the inside of the steel box. 10.6 Using a cylinder of nitrogen and a regulator I progressively increased the pressure on the filter in 100 kPa steps. After each pressure increase I opened the steel box and checked the filter. 10.7 As the filter had not shattered I was then able to safely admit pressure to the filter while it was in the open air and soap and water test for leaks. 10.8 At 300 kPa small leakages were noted around the rubber "O" ring seal under the cover plate. 10.9 At 800 kPa the cover plate was permanently distorted and serious leakage was evident. 10.10 Shortly after I carried out the above tests I met with David Pearson at TestSafe. We compared the filter I had pressure tested, and a second filter of mine, with those on the dryer gas trains and agreed that the filters appeared identical. 10.11 This comparison showed that the body of the filter that I had tested was identical to those originally fitted to the elevated dryer gas trains. The only difference was that the female threads at the inlet and outlet of the two filters from the elevated dryer gas trains were tapped 1-1/2" BSP while my filters were tapped 1-1/4" BSP. (British Standard Pipethread.) 10.12 David Pearson informed me that he had acquired two filters with 1-1/2" BSP connections, from a supplier. He told me that the filter had been tested by subjecting it to air pressure in an explosion proof chamber with the filter immersed in water. 10.13 Photographs of this testing of the filters may be seen as photos 8 - 12 within the report of Alan Kerruish dated 11 June 2004 entitled Examination of Ancillary Items in the LPG Pressure Train. 10.14 Photograph 12 shows the rubber "O" ring having been blown out. 10.15 Following my own testing of a filter and the subsequent formal filter testing by TestSafe I believe the means of gas escaping from the gas line had been found. 10.16 From my observations and my previous experience I concluded that the gas escape emanated from either or both of the gas filters fitted to the gas trains of the elevated dryer. 10.17 In my opinion, a gas leak from a filter, or even both filters on the elevated dryer, would readily cascade down over the dryer, down to the boiler and given time could permeate throughout the premises resulting in an explosion. 10.18 Without over pressure protection on the first stage regulator every component downstream of the first stage regulator, including the gas pipelines and filters would have to have a pressure rating in excess of the maximum pressure in the gas tank as per AG 601 - 1992 clause 4.2.1, otherwise each component would be at risk of failure." (Blue 171K-172S) 111In his report of 5 January 2004 (part of Exhibit AAE) Mr Kerruish said: "Remaining Possible Source of the Gas Escaping ... It is not possible to test the filter boxes on the dryer for leakage as the filters are not intact. The main body is in two parts for the unit fitted to the Zone 1 burner train (see photograph 10) and the unit fitted to the Zone 2 train has been shattered and the cover and some pieces are missing (see photograph 11). It cannot be determined, by the available examination methods, whether the damage to this filter was caused by a mechanical blow resulting from the explosion or by a rupture due to excessive internal pressure prior to the explosion. A possible way to establish the cause of this filter box's failure is to allow a metallurgist to examine the fracture surfaces. If it can be shown that an excess internal pressure rupture caused the damage to the filter, then this becomes the most probable source of a large-scale gas leakage into the factory building. Had this filter housing ruptured due to excess internal pressure it would allow gas at around 500 kPa to freely discharge through a 40mm pipe. From available data on the discharge through a Fisher 620 model regulator fitted with a 3/8" orifice this would result in the discharging of approx. 14.51 M3 /min of gas at atmospheric pressure from the pipeline and based on the upper explosive limit of 9% for LPG would give a volume of 161.26 M3 /min of explosive gas mixture being formed." (Blue 22S-23C) 112In his report of 11 June 2004 (part of Exhibit AAE) Mr Kerruish said: "Testing of Identical Filter Box to those Fitted to the Gas Trains to the Drying Ovens Two identical filter boxes to the units fitted to the gas trains on the drying ovens were obtained for comparison and pressure testing (photograph 8). Unlike all other filters on the site examined, these units were rated 1 bar (100 kPa) pressure, whereas all others examined were rated at 2 bar (200 kPa). One unit was set up in the explosion chamber at TestSafe in a plastic garbage can filled with water (photograph 10) the filter body could be pneumatically tested by raising the body pressure in 100 kPa increments and observing the effect of this pressure on the filter box body and cover. The supply of air was from a cylinder of air and the pressure was controlled through a pressure Regulator. The pressure was raised in 100 kPa increments and at 400 kPa pressure a significant leak was observed on the Left Hand Side of the inlet port of the filter box between the cover and the body (photograph 10). At 690 kPa a much more substantial leak developed on the Right Hand Side between the cover and the body (photograph 11). Then at 975 kPa the box lid had distorted to such a position under the pressure contained within the box that the sealing "O" ring between the cover and the body blew out completely allowing unrestrained discharge of the air from the filter box (photograph 12). Examination of the filter box, when it was removed from the test apparatus, showed both the Left Hand Side and the Right Hand Side of the lid had distorted to an extent that retention of pressure was not possible. Comparing this with the one box that had its lid intact from the drying oven gas train showed that the lid of this unit had suffered identical damage to the test unit. The conclusion that can be drawn from this is that this filter box and similarly the identical one that had been shattered, had been subjected to a pressure in excess of the 100 kPa that it was rated for and of a sufficient magnitude to visibly and permanently distort the box lid." (Blue 33B-N) 113A report relating to the same test and the results obtained is set out in the statement of Mr Kerruish of 25 September 2009 at [16.11] - [16.18] (Blue 68F-S). 114In his final report of July 2004 (Exhibit AW) Mr Pearson said in relation to the gas filters: "9 CONCLUSION ON THE CAUSE OF THE EXPLOSIONS AT BEST CARE It appears reasonable to conclude, based upon the available evidence, that the apparently simultaneous and powerful explosions in the Mega Mill and the Boiler had a common cause. This was the ignition of a cloud of Propane, which had leaked into the Extrusion Building. More than 2 tonnes of Propane seems to have been discharged into the factory. However, less than 0.5 T may have been involved in the initial Propane explosion event. This initial explosion appears not to be sufficiently powerful so as to cause damage to off site buildings. The likely location of the Propane leak was from the Filter Boxes on the Gain Trains above the Dryer. The cause of the leak was a combination of over pressurisation of the Propane supply, and the fact that the Filters were only rated to withstand a pressure of one bar. The cause of the over pressurisation was failure of the 1st Stage Reducing Regulator. The cause of the failure of the regulator was that the sealing/regulating surface of the regulator is likely to have become compromised and so could not seal completely. This had allowed a pressure of several bar to be applied to the Filters, and ultimately their failure." (Blue 402C-K) 115Professor Hoffman, in his report of 31 March 2008 (Exhibit AAJ) said: "OBSERVATIONS AND OPINION ON THE BESTCARE EQUIPMENT I OBSERVED 9 I base my statements here upon my own observations of the Bestcare equipment undertaken on 31 October, 2007 at the Test Safe laboratories at 919 Londonderry Road, Londonderry. The equipment was located in a secure area. I have also been provided with reports as outlined in Appendix A. (Blue 427J-L) ... 15 The failed filter boxes which had resided between the first and second stage regulators were also observed by me. The failure occurred as a result of brittle fracture of the walls of the boxes. One of the failed filter boxes is shown at Figure 20 of Dr Corderoy (Appendix A). (Blue 430B-C) ... 20 Once the disc holder failed the regulator was unable to operate properly. The result of this is that the remaining disc holder pushed the disc with high pressure into the seat ring, leading to failure of the disc as outlined above. The regulator was subsequently unable to regulate the gas flowing from the high pressure LPG tank, leading to high pressure LPG gas flowing downstream from the regulator. 21 The components downstream from the first stage regulator, e.g. the filter boxes, were then unable to structurally withstand the unregulated higher gas pressure leading to their failure and gas leakage, which assumedly caused the fire." (Blue 430P-431D) 116In his report of 21 September 2009 (Exhibit AAK) Professor Hoffman said: "14 The filter boxes which had resided between the first and second stage regulators were also observed by me. Both of these filter boxes had failed. One (Zone 1 in Cox report) had clearly failed via a lateral fracture. In the case of the other (Zone 2 in Cox report), a large part of the filter box and lid were missing meaning that it was not possible to ascertain the mode of failure from the remaining components alone. Observations of the available fracture surfaces showed that failure had been as a result of brittle fracture of the aluminium boxes. The filter material was melted within the boxes. (Blue 444P-T) ... 25 The components downstream from the first stage regulator were then unable to structurally withstand the unregulated higher gas pressure leading to failure and gas leakage, which assumedly caused the fire. 26 I believe that it was highly probable that the Zone 2 filter box failed through over pressure and possible explosion. The reasons for this observation are as follows. Pressure testing of a similar filter box by Kerruish (11 June 2004) revealed that overpressure resulted in a blow out of the o-ring seal between the lid and housing at the side of the filter box. The failed pressure gauge indicated that comparable pressures had been observed in the gas train. Following the explosion, pieces of the Zone 2 filter box, corresponding to the location of the o-ring blow out observed by Kerruish, were unable to be recovered. This is consistent with them breaking into many small pieces as the result of an explosive failure, originating from the same location as the o-ring blowout in Kerruish's test, which may have been caused solely by high gas pressure or accompanied by ignition. 117In his report of 21 June 2010 (Exhibit AAL) Professor Hoffman said: "12 My reasoned proposition that the zone 2 filter box could have failed through over pressure and possible explosion was in response to Mr Cox's assertion that the failure was not caused by over pressure and explosion. 13 It is not possible for Mr Cox to make this statement as pertinent pieces (i.e. the pieces that would demonstrate there had been an explosion) of the filter box are in fact missing. 14 It is also, I believe highly coincidental that the pieces that were missing also correspond to the location of observed leakage during over pressure testing by Mr Kerruish. Mr Cox notes a slight difference between my opinion of the site of leaking and his analysis of the images of Mr Kerruish and attributes differences to different levels of bolt tightness. I would posit that the same argument could be used to explain the slight differences in the location of the explosion. 15 It is my view that Mr Cox's suggestion that failure of the filter box was caused by movements of the pipes is indeed probable, however, given that there is clear evidence of failure of the first stage regulator and the observed failure locations of the filter box in Mr Kerruish's testing, there is a high probability that the filter box in fact failed, leading to leakage of flammable gas at this location. 16 Mr Kerruish did not observe fragmentation during his testing, however, the testing was done in a controlled manner under water and hence there was significantly less energy stored in the system. Fragmentation occurs due to unstable release of energy, as would occur in a gas environment, supplemented by rapidly expanding gases associated with ignition. 17 Mr Cox also asserts that, should the filter box have failed through an over pressure scenario, then it would have failed on a longitudinal direction. This would indeed have been the case if the filter box had not leaked prior to explosion. However Mr Kerruish's tests show that leakage prior to failure was the most likely scenario and hence Mr Cox's assumption is not appropriate." (Blue 471B-U) 118Mr Cox prepared a large number of reports on behalf of the appellants. His Honour rejected Mr Cox's theory of how the FSR was damaged and cast doubt on the reliability of his evidence. His report of 20 July 2010 (Exhibit 31) is, however, important. It was this report which established that the conclusions of Mr Kerruish, based on his testing of a similar filter box, were incorrect. Specifically, the measurements carried out by Mr Cox (the accuracy of which were never challenged in the proceedings) established that the lid of the zone 1 filter box had NOT deformed, as had the lid on the filter box tested by Mr Kerruish. 119In that report Mr Cox said: "In the course of my attendance at the TestSafe facility, the gas filter boxes which had been the subject of tests undertaken by Mr Kerruish and described in his report dated 11 June 2004 were made available to me for examination. This examination was undertaken in the presence of Mr David Pearson of TestSafe who also provided me with various tools for measuring the deformation of the lids of the filter boxes ... Examination of the gas filter boxes The two gas filter boxes which were provided to me for examination at TestSafe are shown in Figures 1 and 2. According to Mr Pearson, the filter box labelled SAM 5311/8 was that which had been pressure tested by Mr Kerruish while the filter box labelled SAM 5311/9 was an identical filter box which had not been tested but retained as a control sample. An examination of the filter box labelled SAM 5311/8 revealed that the O ring was protruding from the gap between the lid and the body of the filter box on one side of the filter box. This caused the lid to deform in order to accommodate the O ring, relative to the equivalent position on the opposite side of the filter box which appeared to be deformed to a lesser amount. The extent of the deformation on both sides was then determined by measuring the gap between the lid and the body of the filter box using a feeler gauge. This revealed a gap of 1.30-1.35mm at the position of the O ring and .75-.80mm at the equivalent position on the opposite side of the filter box. The gap between the lid and the body of the control filter box SAM 5311/9 at the equivalent position to where the O ring was protruding on the sample SAM 5311/8 was measured and observed to be .40-.45mm while the gap on the opposite side was measured and observed to be .60-.65mm. On this basis, it was concluded that the lid of the filter box labelled SAM 5311/8 was deformed relative to the control filter box, with the extent of the deformation being noticeably greater as a result of the protrusion of the O ring. The filter box for the Zone 1 dryer gas train from the BestCare site, as examined on June 17, 2010, is shown in figure 3 while the gap between the lid and the body of the filter box is shown in figure 4. Measurement using the same method as above revealed that the gap between the lid and the body of the filter box at the equivalent position to where the O ring was protruding on the sample SAM 5311/8 was .55-.60mm while the gap on the opposite side was .60-.65mm. By comparing these values to the measurements obtained for the control sample SAM 5311/9 I concluded that the Zone 1 dryer gas train filter box lid from the BestCare site had not deformed as a result of excessive internal pressure, as had been claimed by others. A similar measurement on the filter box for the Zone 2 dryer from the BestCare site could not be undertaken as the lid for this filter box had become detached from the body of the filter box and was not available for examination. ... Discussion At paragraph 16.16 of his statement dated September 25, 2010, Mr Kerruish concluded that the lid of the filter box from the Zone 1 dryer at the BestCare site had suffered identical damage to the filter box that had been subjected to pressure testing by him namely SAM 5311/8. On this basis, he concluded at paragraph 16.17 that the filter box from the BestCare site "had been subject to a pressure far in excess of 100 kPa, the pressure that it had been rated for, and a pressure of a significant magnitude to physically and permanently distort the filter box lids". The above measurements show that this conclusion is incorrect as the lid of the filter box from the Zone 1 dryer at the BestCare site was not deformed in the same manner as the lid of filter box SAM 5311/8." (Blue 924F-925P) 120The conclave of experts took place over three days on 17, 26 and 27 August 2010. The statements of opinion by the experts were recorded in Exhibit AAW. Relevant extracts are: "11 The source of the Gas Leak for the Primary Gas Explosion Mr Kerruish: agreed with Professor Hoffman Professor Hoffman: concluded that the first stage regulator failed and the most likely location for the leak, based on the available evidence, was the dryer filter boxes Professor Masri: agreed with Professor Hoffman Mr Cox: concluded that the cause of the gas leak was failure of the boiler main burner regulator OPSO" (Blue 1457B-F) "14. Dryer filter Boxes 14.1 Do you agree that the lid of the Zone 1 dryer filter box shown in Figures 3 and 4 of Mr Cox's report dated July 20, 2010 is not deformed relative to the control sample SAM 5311/9. Mr Kerruish: yes, at the time of the photographing they are not, but not at the time of the explosion and fire Professor Hoffman: did not comment as has not considered Mr Cox: yes 14.2 Do you agree that the lid of the Zone 1 dryer filter box shown in Figures 3 and 4 of Mr Cox's report dated July 20, 2010 is in its original position in that it is not protruding in the same manner as that of the sample SAM 5311/8 from the testing undertaken by Mr Kerruish. Mr Kerruish: yes Professor Hoffman: did not comment as has not considered Mr Cox: yes 14.3 Do you agree that the O ring between the lid and the body of the Zone 1 dryer filter box shown in Figures 3 and 4 of Mr Cox's report dated July 20, 2010 is in its original position in that it is not protruding in the same manner as that of sample SAM 5311/8 from the testing undertaken by Mr Kerruish Mr Kerruish: yes, in that the O ring is not protruding but he cannot say that it is still in its original position Professor Hoffman: did not comment as has not considered Mr Cox: yes 14.4 Do you agree that it is not possible to determine whether the lid of the Zone 2 dryer filter box was deformed prior to the explosion in the manner that the sample SAM 5311/8 from the testing undertaken by Mr Kerruish was deformed. Mr Kerruish: yes Professor Hoffman: did not comment as has not considered Mr Cox: yes 14.5 Do you agree that the Zone 1 dryer filter box failed as a result of mechanical stresses caused by the explosion Mr Kerruish: no, it was caused by the over-pressurization of the second stage pipework Professor Hoffman: did not comment as has not considered Mr Cox: yes 14.6 If the answer to the above is no, what do you say caused the Zone 1 dryer filter box to fail in the manner observed Mr Kerruish: no, it was caused by the over-pressurization of the second stage pipework Professor Hoffman: did not comment as has not considered Mr Cox: N/A 14.7 Do you agree that the Zone 2 dryer filter box failed as a result of the mechanical stresses caused by the explosion Mr Kerruish: no, it was caused by the over-pressurization of the second stage pipework Professor Hoffman: did not comment as has not considered Mr Cox: yes 14.8 If the answer to the above is no, what do you say caused the Zone 2 dryer filter box to fail in the manner observed Mr Kerruish: no, it was caused by the over-pressurization of the second stage pipework Professor Hoffman: did not comment as has not considered Mr Cox: N/A" (Blue 1491E-1493H) "Additional Questions from McCabe Terrill (and answers) 1. Assuming that the lid on the filter boxes [sic] box for the Zone 1 dryer had been deformed in the manner suggested by Mr Kerruish, what is the explanation for the measurements noted by Mr Cox in the report of July 20, 2010. Mr Kerruish: the lid had deformed but returned to its original position due to the stress relieving effect of the fire Professor Hoffman: did not comment as he has not considered Mr Cox: The Zone 1 dryer filter box lid examined by Mr Kerruish was not deformed relative to the control sample" (Blue 1473D-J) "24 Which of the following is the sequence of failure for the first stage regulator ... (i) Hoffman June 21, 2010 paragraphs 27-33 Failure of the disc holder Incomplete seal of the sealing disc on the seat ring Short term leakage past the sealing disc of the first stage regulator Failure of the centre portion of the sealing disc Failure of the annular portion of the sealing disc Failure of the dryer filter box lids Leakage of gas into the building Gas explosion" (Blue 1512W-1513G) 121The position in relation to the opinion of Professor Hoffman was complicated when in evidence in chief he stated that he still held the opinions set forth in his reports (Black 418M-419K). In cross-examination, however, he also adhered to the expressions of opinion which he had given in the conclave of experts: "Q. Professor, prior to giving evidence Monday and today have you become familiar with your reports '08, '09 and June 2010? A. Yes, I have. Q. And have you read the conclave report referable to your answers recently? A. Yes I have. Q. And so far as your reports are concerned in evidence here, I think you said you agreed with their content and that you adhere to what's in those documents? A. Yes. Q. So far as the conclave answers are concerned, firstly you were at the conclave where there was a moderator, I think? A. That's correct. ... Q. Questions were asked and answered and a draft was circulated? A. Correct. Q. And the opportunity was given to make alterations if required? A. Yes, that's correct. Q. And you read it through, I take it? A. Yes. Q. You saw that the answers accorded with what you'd said? A. Yes. Q. And having been satisfied with the questions and answers you signed as your opinion in relation to the answers set out in the conclave referable to you? A. That's correct, yes." (Black 447W-448S) 122The appellants submitted that taking into account that evidence, his reports and responses at the conclave of experts, it was simply not open to his Honour to explain an apparent contradiction in what Professor Hoffman was saying on the basis that he was mistaken when making the responses attributed to him at the conclave of experts. The appellants submitted that the evidence of Professor Hoffman at the conclave should be interpreted as follows: "What he's saying here is "I've expressed my views but I can't comment in point of fact because I haven't considered the debate between Mr Cox and Mr Kerruish about the position of the lid on the zone 1 filter box". In a sense when you look at the history of the questions and answering the professor is saying "do not comment because have not considered" is really saying "I've given opinions and I maintain those opinions up front as to the cause but that's obviously on the basis of available evidence, my assumptions. I do not express opinions about a dispute which exists between Mr Cox and Mr Kerruish which I have not considered". (AT 74.7-26) 123I have concluded that the interpretation given to the evidence of Professor Hoffman by the appellants should be accepted. It is clear that Professor Hoffman had not abandoned the opinion which he expressed in his reports, i.e. he still thought that the most likely source of the gas leak was the failure of the zone 1 and zone 2 filter boxes. That having been said, he had not considered the dispute between Mr Cox and Mr Kerruish arising from the tests carried out by Mr Kerruish. As a result, he was not prepared to express an opinion concerning that issue. Neither side at trial asked Professor Hoffman any questions directed to the results of the testing of filter boxes by Mr Kerruish nor the possible contradiction in his expression of opinion. 124The appellants and respondents criticised the failure by the others to ask any questions of Professor Hoffman directed to that subject. The Court was invited to infer that this failure reflected adversely on one side or the other. I am not prepared to draw such an inference. Quite clearly each side at trial made a considered decision not to explore this issue with Professor Hoffman. As a result, the parties and this Court must deal with the evidence, as it stands, unassisted by any further clarification. 125That part of Professor Hoffman's evidence which dealt with the possible fracturing of the filter boxes from internal pressure does not appear to have been greatly pressed at trial. The testing by Mr Donnelley and Mr Kerruish did not produce such a result, although Professor Hoffman explained that the test conditions under water may have prevented such a result. There was evidence that mechanical forces also affected the filter boxes. Nevertheless, there is the unchallenged evidence of Professor Hoffman that the pieces missing from the zone 2 filter box corresponded to the location of the leaking and deformation observed during pressure testing by Mr Kerruish. (Blue 471G) That observation by Professor Hoffman would favour, albeit only slightly, the proposition that the zone 2 filter box failed as a result of over-pressurisation. 126Before analysing the submissions of the appellants and respondents in detail, it is necessary to clarify an issue which arose in the course of submissions relating to the testing of a filter by Messrs Donnelley and Kerruish. In their oral and written submissions, the appellants asserted that there had been deformation of the filter lids at pressures of 300 kPa, 400 kPa and 690 kPa in the course of those tests. That is not the evidence. A proper reading of the evidence (set out at pars [110] and [112-113] hereof) makes it clear that deformation of the filter lids tested was only described at a pressure of 800 kPa in the case of Mr Donnelley's test and at 975 kPa in the case of the Kerruish test. 127Because of the care with which the test results were described by Messrs Donnelley and Kerruish, I infer that had deformation of the filter lids occurred when exposed to lower pressures, a statement to that effect would have been made. Accordingly, I have concluded and have proceeded on the basis that deformation of the filter lids did not occur during the testing process until they were exposed to pressures of 800 kPa and 975kPa. 128The appellants submitted that the finding by his Honour at [195] that the gas leak occurred in the zone 2 filter was not open to him. The appellants developed their submission as follows. The zone 1 and zone 2 filter boxes were the same. When tested to 975kPa, the lid on an identical filter box deformed so that the rubber "O" ring became unseated. When the filter box which had been tested to that pressure was compared with the zone 1filter box, they were different. Whereas the lid on the test filter box had significantly deformed, there was no such deformation of the lid on the zone 1 filter box. In fact, the condition of the lid on the zone 1 filter box was similar to the filter box which was used as the test sample, i.e. no deformation or damage. 129The submission continued. There was the evidence of Professor Hoffman that similar pieces of equipment exposed to the same pressure should react in the same way. The zone 1 filter box and the zone 2 filter box had been exposed to the same pressure. The lid on the zone 1 filter box had not deformed. It could be inferred, therefore, that the lid on the zone 2 filter box (if it had been capable of examination) would have been found not to have deformed. Accordingly it was not open to his Honour to conclude that the gas leak came from the zone 2 filter. 130In summary, the appellants submitted that the tests carried out by Mr Kerruish established that the zone 1 filter and therefore the zone 2 filter, could not have been over-pressurised otherwise the lid on the zone 1 filter would have been deformed as was that on the test filter. If the zone 1 and zone 2 filters had not been over-pressurised so as to cause them to fail, then over-pressurisation could not have occurred in the gas lines before the explosion. The appellants submitted that this was objective evidence which conclusively established the absence of over-pressurisation in the gas lines before the explosion. That being so, it defeated the respondents' theory as to causation so that the respondents had failed to prove their case. 131It also followed that the crucial finding made by his Honour at [195] was not open on the evidence. 132Implicit in that submission was a rejection of Mr Kerruish's theory (which was not supported by any other witness) to the effect that the lid on the zone 1 filter box may have deformed but that it had in some way been restored to its original condition, as a result of the effects of fire. I agree that if the other elements of the appellants' submissions are made out, this explanation by Mr Kerruish should not be accepted. It was effectively refuted by Mr Cox (Black 658T-659E). 133The respondents submitted that the test results obtained by Mr Kerruish, with respect to the filter boxes, were not decisive. At most, they submitted, they constituted evidence which might assist the appellants. The respondents submitted that the appellants' theory was based on a number of assumptions which were not made out and it left a number of important questions unanswered. For example, if the gas leak did not come from the filter boxes, where else could it have come from? 134On that last issue, the respondents submitted that there was no evidence of any other source of a gas leak other than the zone 1 and zone 2 filter boxes. The start point was the origin of the gas explosion. It was agreed between the experts that the origin was in the extrusion building, either in the electrical control room under the dryer, or in the boiler room and boiler (Blue 1454D-J). That, of course, was consistent with LPG being heavier than air and flowing down from the filters after having escaped. 135Only two sources of the gas leak for the primary gas explosion were identified at the hearing. Mr Kerruish, Professor Hoffman and Professor Masri concluded that the most likely location for the leak was the dryer filter boxes (Blue 1457C-E). Mr Cox concluded that the cause of the gas leak was failure of the boiler main burner regulator OPSO (Blue 1457F). The conclusion of Mr Cox had to be rejected when his Honour found as a fact that the boiler valve was closed on the gas train leading to the main burner when the factory closed down on 25 January 2003 ([175]; Red126H-J). 136Mr Cox did not propose any other location for the source of the gas leak. Importantly, Mr Cox conceded that if his pressure wave theory were not accepted, the only theory to explain the failure of the FSR was that put forward by the respondents. Implicit in that acceptance of the respondents' theory is an acceptance that the source of the gas leak was the zone 1 and zone 2 filter boxes. 137Although the appellants tried to qualify that concession by Mr Cox, it was clearly made on the evidence. "Q. Do you accept that if the pressure wave theory is not accepted, then the only other theory to explain the failure of the first stage regulator is that put forward by the plaintiff? A. On the available evidence, yes. ... Q. Your pressure wave theory is that the pressure wave came back up the pipe and damaged the first stage regulator, isn't it? A. Yes. Q. The plaintiff's theory, as you understand it, is that the first stage regulator failed causing over pressure downstream, isn't it? A. Yes. Q. And there is no other possible explanation for the failure of the first stage regulator apart from those two theories, is there? A. They are the only possibilities that I have addressed, yes, on the evidence that I have. Q. Do you accept that if the first stage regulator failed before the explosion, then firstly there would have been over pressure downstream and it would have leaked into the factory - "yes" or "no"? A. Yes, but with the clarification -- ... Q. You said "yes, but with the clarification" in answer to the last question that was put to you. Did you wish to add to your answer? A. Yes, I did. Q. What was it you wished to say? A. That would only apply if the over pressurisation downstream of the first stage regulator was sufficient to cause a failure of the pipe work within the factory." (Black 750E-751F) 138That concession of Mr Cox has to be looked at against the background of the nine reports prepared by him for use in proceedings comprising almost 400 pages. His analysis of the evidence was comprehensive and to the best of his ability, exhaustive. I take his qualification "on the available evidence" to mean that there was no other evidence that he had seen which would lead to a different conclusion. 139Because the origin of the gas explosion could be identified with some precision, the first persons to inspect the accident scene examined all of the pipe work available in order to identify the source of the gas leak. The unanimous conclusion of the persons involved in that exercise, i.e. Messrs Kerruish, Donnelley and Pearson, was that the most likely source of the gas leak was the zone 1 and zone 2 filter boxes. 140The extensive nature of the investigation of the site post-accident can be seen from the report of Mr Pearson of June 2004 (Exhibit AV) at Blue 310-333, with particular emphasis on Blue 323C-V, being the extrusion plant which he had identified as the centre of the explosion. He said: "The "Zone 2" gas train (eastern end) was removed and taken to TestSafe for examination and testing on 5.2.03. The "Zone 1" gas train was removed for examination and testing on 7.2.03. The dryer was examined in detail on 19.3.03." (Blue 323K) 141Reference has already been made to the investigations of Mr Donnelley shortly after the accident at [110] hereof, in particular pars 10.1-10.3. In his report of 25 September 2009 Mr Kerruish said: "7.7 Having investigated the entire closed pipe system at BestCare, it is my opinion that the most likely point of escape of a substantial quantity of gas into the factory area was the two gas filters fitted immediately prior to the second stage regulators on the elevated dryer located in the extrusion building." (Blue 50O-P) "10.11 All three gas trains from the baking facility were removed from the site and taken to TestSafe at Londonderry for further testing. The gas trains were removed by a fitter in the presence of both David Pearson and myself and under our direction. Photographs were taken of the equipment for identification and custody purposes by David Pearson on site. "10.12 We also inspected the whole of the gas train and the surrounding area of the other gas consuming appliance in the building which was the drying oven in the extruding room (known as the dryer). This appliance was elevated on a steel supporting structure at a level which could be classed as a mezzanine floor. It was located directly above the main electrical switch room for the factory and office." (Blue 59J-N) 142Accordingly, the evidence as to the source of the gas leak was all one way: an examination of the pipes and gas lines following the accident by those experts who were on the scene soon after, i.e. Messrs Kerruish and Donnelley, failed to identify any other source. Mr Cox, the appellants' expert, was unable to identify any other source if his theory were rejected (as it had to be). All other experts who gave evidence identified the zone 1 and zone 2 filter boxes as the source of the gas leak. 143The appellants submitted that gas could have leaked from solenoids in the extrusion dryer. This submission was based on paragraph 12 of a report of Mr Cox, dated 13 August 2010 (Blue 935K-O) in which he said: "Of course, this assumes that the solenoid valves operated as they were intended to however, this was never tested. Given the presence of gas in the boiler furnace, coupled with the comment by Mr Donnelley during the course of a joint inspection that such devices were "notorious for leaking", it is my view that they should have been tested." 144This was the only evidence on the subject. It related to solenoid valves on the boiler, not the extrusion dryer. Its basis was an untested piece of hearsay evidence. Implicit in the submission is that the Court should accept this evidence as accurate and it should conclude that the solenoids on the extrusion dryer which was a very different piece of equipment, were likely to leak. I am not prepared to make such a finding. 145It follows that there was no evidence before the Court of any source of the gas leak other than the zone 1 and zone 2 gas filter boxes. Put at its highest, the appellants' submission is that there may have been a gas leak elsewhere but evidence of it was destroyed in the explosion. That is, a submission put without the support of any of the experts who gave evidence in the proceedings. 146The respondents submitted that the appellants' submission based on the results of the testing of a similar filter box by Mr Kerruish depended upon the Court being satisfied that the zone 1 and zone 2 filter boxes had been exposed to full tank pressure, i.e. 850 kPa. 147The respondents put their submission as follows. It was known that the deformation of the filter box lid occurred at a pressure of 975 kPa in the test carried out by Mr Kerruish and at 800 kPa in the test carried out by Mr Donnelley. There was no evidence of any deformation of the filter lid at a lower pressure (See [110] and [112-113] hereof). For a comparison between the zone 1 filter box and the filter boxes which were tested to be valid, they had to have been exposed to the same pressure, i.e. approximately 850 kPa. The preponderance of evidence was that the zone 1 and zone 2 filter boxes either had not been exposed to full tank pressure or it was not necessary for them to have been exposed to full tank pressure for them to have failed and leaked gas. 148The respondents accepted that in two places Mr Kerruish said that full pressurisation of the gas trains and in particular, the zone 1 and zone 2 filter boxes, had occurred. In his report of 21 June 2010 he said: "96(j) Once failure of the FSR seating assembly took place, full tank pressure, estimated at 850 kPa was applied to a filter rated for only 100 kPa and this led to a failure of this component allowing an uninterrupted flow of LPG from the storage tank into the factory." (Blue 140P-Q) 149At trial he gave this evidence: "Q. What can you say, if anything, about whether there would have been over pressurisation of the lines downstream of the regulator as a result of the regulator being in that condition? A. Yes the first stage regulator seat had failed in the condition we saw it up there, it would have been full tank pressure put on to the second stage regulators downstream. They were designed for 140 kPa and they would have got about 850 kPa pressure on them." (Black 217P-Q) 150The respondents pointed out that this evidence of Mr Kerruish was inconsistent with other evidence which he gave and with his theory of gradual over-pressurisation: "96(h) As the pressure increased in the downstream line all the filter boxes on the upstream side of the second stage regulators (these are not protected by OPSOs), suffered an increase in pressure beyond their rated pressure and ultimately beyond their yield strength (i.e. the point at which permanent deformation of the material will occur). (i) The filter box lids on the two lowest rated boxes which were fitted to the dryer ovens began to bow upwards much the same as a simply supported beam evenly loaded would. This in turn allowed a leakage of gas out of the filters through their O-ring seal to spill into the plant below." (Blue 140J-N) 151At Blue 139D-H Mr Kerruish explained his opinion that the FSR was leaking for some time before the explosion and that the build up of pressure in the gas lines took place between the closing down of the appliances on the Friday afternoon until the explosion. He gave this evidence in relation to the second stage OPSOs, but it was relevant to the pressurisation of the gas lines generally: "Once the escape of gas commenced at the point of least resistance, it is unlikely that the line pressure at this furthest regulator ever reached 454 kPa. In all likelihood the inlet side pressure would have reduced to zero once an upstream breakage or major leak occurred." (Blue 59I-J) 152Mr Pearson expressed a similar opinion when he said: "The cause of the failure of the regulator was that the sealing/regulating surface of the regulator is likely to have become compromised and so could not seal completely. This had allowed a pressure of several bar to be applied to the filters and ultimately their failure." (Blue 402K) 153Mr Donnelley was consistent in his opinion that the build up of pressure would have been gradual. He said: "As to paragraph 537, if the FSR simply allowed a slow leakage past the sealing disc, which is my view, then the pressure in the gas line could take some time, possibly minutes, to fully pressurise the downstream gas lines. I would describe this as a first stage failure of the FSR. This would only apply if the first stage regulator completely failed in one single step. Based on my years of experience with these regulators, in my opinion the regulator allowed gas to pass to the downstream side and thus caused a build up of pressure in the downstream pipe work. The dryer filters began to leak thus causing a build up of free gas in the building. This free gas found an ignition source and an explosion occurred. In my opinion, the failure of the first stage regulator occurred in two stages. 1) The gas leaked past the sealing disc thus over pressurising the downstream pipe and fittings. The dryer filters began to leak gas. Uncontained LP Gas permeated throughout the area and was finally ignited. ... The time taken for the downstream gas pressure to build up was, in my opinion, probably some hours rather than seconds." (Blue 250J-251E) "Mr Cox says that the FSR failed as a result of a sudden application of a sudden excessive load. In relation to this comment: (a) the cause of the 'sudden excessive load' is not identified; and (b) there is an assumption that an FSR failure is a single stage event and without explanation excludes wadding." (Blue 257H-J) "Mr Cox appears to have misunderstood my comments. In my report in discussing the failure of the filters, I was referring to the fact that either or both began to leak gas, when the seal between the lid and the body discharged gas, as demonstrated by Mr Kerruish in his testing of the filters. Thus in my opinion there were two levels of failure of the dryer filters: (a) The failure of the filter to contain the LPG at pressures above 100 kPa rated working pressure of the filters; and (b) the failure by fracture of the filters due to the mechanical damage caused by the explosion. On the assumption that the observation of Mr Kerruish as described in his report on testing of the filter boxes, namely the distortion of the lid due to internal pressure, is correct, Mr Cox is overlooking obvious evidence which indicates a failure of the filters to contain LPG which demonstrates a leakage of gas to the atmosphere prior to the fracture of the filters." (Blue 258H-N) 154Professor Hoffman, although he opined a failure of the disc holder leading to a failure of the disc causing an unregulated flow of gas downstream, also accepted leakage of gas past the sealing disc of the FSR before that event thereby leading to an increase in pressure downstream. 155The respondents' submissions took into account that the testing of similar filter boxes by Messrs Donnelley and Kerruish had produced leakage of gas at 300 kPa (Mr Donnelley), 400 kPa (Mr Kerruish), 690 kPa (Mr Kerruish), 800 kPa (Mr Donnelley) and 975 kPa (Mr Kerruish). The only evidence of deformation of the filter lids in those tests was that in the case of the filter tested by Mr Donnelley, it became deformed at 800 kPa and in the case of that tested by Mr Kerruish it deformed at 975 kPa. There was no evidence of any deformation before those pressure levels were reached but there was evidence of gas leakage at lower pressures and quite significant gas leakage at 690 kPa. 156The respondents submitted that the evidence of gas leakage up to and including pressures of 690 kPa without deformation of the lid was consistent with the appearance of the zone 1 filter box after the explosion and was also consistent with the filter box leaking gas sufficient to cause a build up and subsequently an explosion. That was certainly the opinion of Messrs Donnelley and Pearson and was supported by some evidence from Mr Kerruish and was in part supported by evidence of Professor Hoffman. 157On the basis of the evidence referred to by the respondents, I am satisfied that the respondents have successfully challenged one of the assumptions behind the appellants' submission concerning the testing of the filter boxes. The evidence supports a submission that the filter boxes, rated as they were to 100 kPa, were likely to have leaked gas in a substantial quantity over a considerable period of time as pressure in the gas lines was increasing because of wadding of the FSR but before the failure of the disc holder and disc in the FSR. That process would have occurred without deformation of the lid of the filter boxes and is thus not inconsistent with the test results obtained by Mr Kerruish, the measurements made by Mr Cox and the condition of the lid of the zone 1 filter box. 158The respondents submitted that the appellants' submission in relation to the filter boxes depended upon the zone 1 and zone 2 filter boxes, not only being of a similar kind but being affected in the same way by over-pressurisation. The respondents submitted that this was not supported by the evidence. 159The respondents referred to the difficulty with such a co-relation between the two filter boxes, identified by Professor Hoffman: "12 My reasoned proposition that the zone 2 filter box could have failed through over pressure and possible explosion was in response to Mr Cox's assertion that the failure was not caused by over pressure and explosion. 13 It is not possible for Mr Cox to make this statement as pertinent pieces (i.e. the pieces that would demonstrate there had been an explosion) of the filter box are in fact missing." (Blue 471B-F) 160The respondents also noted that the evidence relied upon by the appellants to the effect that similar pieces of equipment when exposed to the same pressure would be affected in the same way, was not given about the filter boxes but was given by Professor Hoffman in relation to the Bourdon tube pressure gauges. The evidence relied upon by the appellants on that issue was that given by Professor Hoffman at the conclave of experts when he affirmatively answered the proposition: "15.4 Do you agree that if the Bourdon tube pressure gauge located adjacent to the dryer was subjected to the same degree of over-pressurisation as the Bourdon tube pressure gauge located immediately downstream of the first stage regulator it would have suffered similar damage, namely bulging of the Bourdon tube." (Blue 1494S-T) 161The respondents submitted that what might be true of the Bourdon tube pressure gauges was not necessarily true of gas filter boxes. They noted that one of the components of a gas filter box was the rubber O-ring and that the age and quality of such an O-ring might differ from filter box to filter box. They submitted that for the appellants to substantiate their submission, there needed to be direct evidence specifically focused on gas filter boxes. Evidence by implication, such as was relied upon by the appellants here, was not sufficient. The respondents submitted that what was needed but was not available, was expert evidence directly on point. 162In further support of this submission, the respondents drew attention to the difference in test results obtained by Messrs Donnelley and Kerruish. Even though the filters tested were of the same kind, that tested by Mr Donnelley commenced to leak at 300 kPa, whereas that tested by Mr Kerruish commenced to leak at 400 kPa. The lid on the filter tested by Mr Donnelley deformed at 800 kPa, but such deformation did not occur to the lid of the filter tested by Mr Kerruish until it was exposed to a pressure of 975 kPa. The respondents submitted that while the results were generally consistent, the differences were significant, particularly the fact that the lid on the filter tested by Mr Kerruish had not deformed until exposed to pressure of 975 kPa which was more than a bar greater than the tank pressure of 850 kPa. 163The respondents noted that the pressure testing carried out by Messrs Donnelley and Kerruish was conducted under different conditions to those that existed in the factory. In particular, Mr Kerruish's test filter was submerged in water during the pressure testing. Professor Hoffman explained how that of itself was likely to have prevented the fragmentation of all or part of the filter box as a result of over-pressurisation. The respondents submitted that there may have been other consequences of the testing methodology which are not known, but which would account for differences between the filter boxes tested and the state of the zone 1 and zone 2 filter boxes following the explosion. 164I accept the general thrust of the appellants' submission that as a matter of common sense there should be a correspondence in the effect of over-pressurisation between the zone 1 and zone 2 filter boxes. Unless there was such a correspondence, there was little value in Messrs Donnelley and Kerruish carrying out the tests which they did. On the other hand, the test results showed real differences so that there is force in the respondents' submission that the differences were such that in the absence of expert evidence, it would be unsafe to assume an exact correspondence. It follows that care needs to be exercised before basing conclusions on the proposition that the zone 2 filter box would have behaved in the same way as the zone 1 filter box when affected by over-pressurisation. 165My conclusion in relation to the dryer gas filters is that the evidence upon which the appellants rely as constituting objective evidence to conclusively establish the absence of over-pressurisation in the gas lines before the explosion does not reach that standard. The submission depends upon two important assumptions which have not been made out, i.e. over-pressurisation to 850 kPa and a very close correspondence between the zone 1 and zone 2 filters when affected by various levels of over-pressurisation. Finally, and perhaps most importantly, the submission is a lawyer's construct based on pieces of evidence taken from the respondents' case, but it lacks the endorsement of any of the experts who gave evidence in the proceedings. That difficulty is highlighted by the appellants' inability to identify on a rational basis, an alternative source of the gas leak. This is not surprising when all of the expert evidence at trial, including that of Mr Cox, identified the dryer gas filters as the source of the gas leak. The FSR 166The appellants submitted that the failed state of the FSR following the explosion did not assist either side in establishing whether the FSR had failed before or after the explosion. I agree with that submission. The nature of the failure (to the extent that this could be ascertained from its condition after the explosion) could be used to validate various theories but otherwise did not favour either side. 167The appellants submitted that there were differences in opinion between the respondents' experts as to what was the sequence of events leading up to the failure of the FSR and therefore what was the precise mechanism of that failing? The appellants submitted that this meant that the respondents were not putting forward a single expert case as to how the FSR had failed, but rather three or four positive expert cases which were inconsistent each with the other. 168There was certainly a material difference between the theory put forward by Professor Hoffman on the one hand, and those of Messrs Donnelley, Pearson and Kerruish on the other. Professor Hoffman opined that the disc holder had failed leading to the sealing disc failing causing a substantial flow of gas past the FSR and over-pressurisation of the downstream pipes and fittings. There was, however, significant unanimity between the other three experts in that they identified the existence of wadding in the FSR, allowing gas to leak past the FSR causing a gradual over-pressurisation downstream. 169Even Mr Cox agreed that the deterioration of the sealing disc and disc holder was such that it was only a matter of time before they failed. That evidence left open the possibility that damage to the seat ring had already allowed gas to leak through the FSR before the explosion (Black 878-880): "Q. What I was going to ask you, having taken you through your, the observations as to the disc holder and the seat ring, I understood that you had concluded that before the explosion there were fatigue cracks in these pieces of equipment? A. Yes, in the disc holder and also in the sealing disc. Q. Would I be right in concluding that given what you observed it was, these items were undergoing a process of deterioration? A. Yes. Q. And would it be fair to conclude that it was inevitable that either or both would fail at some time? A. Yes. Once the fatigue cracks in the sealing disc or the fatigue cracks in the disc holder reached a length that was critical for the normal downstream pressure, then overload fracture could occur. Q. And it would also be fair to say that it would be impossible to know when, of course, that failure would take place? A. In terms of time, yes. ..." (Black 879W-880I) 170What the appellants' submission fails to have due regard to is that all of the theories put forward by the respondents' experts had in common that the FSR had failed to regulate pressure before the explosion leading to over-pressurisation in the downstream gas line. Even Professor Hoffman accepted some leakage of gas through the FSR prior to the failure of the disc holder as a result of wadding. It was for that reason that the primary judge qualified his acceptance of the opinion of Professor Hoffman as to the failure of the FSR by saying: "199 The common thread through the theories of the plaintiff's expert is that the immediate cause of the failure of the FSR was the failure of the disc to effect a seal against the inflow of LPG. The evidence established that this is what happened. To reach this conclusion it is not necessary to identify the precise cause of the failure of the disc, and the point at which the disc holder fractured ..." (Red132S-V) 171Accordingly, I do not accept the appellants' submission that the respondents were putting forward expert theories to explain the failure of the FSR which were not only materially different, but which were inconsistent. There was unanimity on the essential feature, i.e. a failure by the FSR to regulate pressure before the explosion, which would have been prevented had an OPSO been fitted. It does not matter whether the pre-explosion failure of the FSR which led to over-pressurisation was failure as a result of wadding, or failure as a result of the disc holder breaking. The Missing Gas 172At trial the respondents submitted that somewhere between 1.2 and 2.3 tonnes of gas escaped from the tank between shutdown on Friday 24 January 2003 and the time of the explosion. They submitted that a loss of that amount of gas could only be explained by a failure of the FSR before the explosion. His Honour did not make a finding in relation to that matter, but the parties joined issue on it in the appeal. 173The evidence relied upon by the respondents was as follows. The volume of the tank was 43,000 litres. The last delivery of gas was made to the factory on 24 January 2003 when the tank was filled to 84 percent capacity. The delivery was completed at 11.06am. One litre of liquid Propane weighs .51 kg. Consequently, when the tank was filled on 24 January 2003 there were approximately 18.42 tonnes of Propane in the tank. After the explosion, 14.88 tonnes of Propane were removed from the tank. This left approximately 3.54 tonnes of gas unaccounted for. 174The escape of gas from the tank following the explosion would not account for such a substantial quantity of gas. Mr Verning stopped the flow of gas from the tank between 10 and 13 minutes after the explosion. Taking the upper band of 13 minutes, the unrestricted flow of gas through the FSR would have yielded approximately 132 kg of Propane after the explosion. 175The respondents then took into account how much gas was likely to have been consumed in the factory's operations between the time the tank was last filled (at 11am) and the time the factory shut down its appliances for the last time on 24 January 2003. Without setting out the detail, there was persuasive evidence that in that period the factory would have consumed more than one tonne of Propane but much less than two tonnes. This evidence was based on the average daily consumption of Propane by the factory at the time. The calculations by the respondents' experts produced a range of between 1.24 and 2.34 tonnes of Propane which was unaccounted for and by inference, should be regarded as having leaked before the explosion. 176On the basis of those calculations, the respondents submitted that the only way such a large quantity of Propane could have been lost would have been if the FSR had failed before the explosion causing over-pressurisation and leakage over a significant period of time. 177At trial Mr Cox challenged those calculations. He asserted that 4.4 tonnes of gas could have been consumed during operations following the tank refill and the cessation of operations at the factory. This would have involved a rate of consumption of around .4 tonnes per hour. Mr Cox's evidence was challenged by persons such as Mr Goldring, a director of BestCare. He said that for the evidence of Mr Cox to be accepted, the factory would have needed to have almost trebled its highest previous rate of consumption of gas. 178There was other evidence which supported the respondents' submission that a substantial amount of gas must have escaped before the explosion. Mr Gollan smelt gas at about 5.30pm on the afternoon of the explosions when he was about 300 metres away from the factory. That evidence is consistent with the escape of an amount of gas which could only be accounted for by the failure of the FSR. 179There was evidence from Professor Masri that it required approximately 760 kg of gas to cause a fireball of the size observed by witnesses (Blue 501H). I accept that only the failure of the FSR could explain a leak of gas of that order of magnitude. 180The respondents' missing gas theory was challenged on appeal on the following bases. There was evidence that the gauge on the LPG tank was faulty and could be moved by hand by 4 percent (Blue 163S). As a result, the appellants submitted that the opening premise that the tank was filled to 84 percent could not be made out. If the tank were only filled to 80 percent, that of itself would account for 1.6 tonnes of the missing gas. 181The LPG tank was not emptied until six months after the explosion. The respondents submitted that in the absence of evidence that the LPG tank was secured between January and July 2003 so that no LPG gas was drawn from it, the calculations based on the LPG removed from the tank would be questionable. In that regard, the appellants noted that there was evidence that LPG, which powered forklifts used to clean up the factory after the explosion, could have been taken from the tank. 182The evidence to support the use of LPG by the forklifts is thin in the extreme. It relies upon a photograph of a forklift, which is almost entirely obscured, which may be powered by LPG. There is no evidence of any use of the LPG in the tank between January and July 2003. Moreover, it seems that steps had been taken to prevent the use of the tank after the explosion. Mr Donnelley in his report of 9 April 2008 said: "5.3 The vapour phase gas line from the tank to the buildings had been physically disconnected from the gas tank by removal of the connecting steel pipe between the tank outlet valve and the inlet flange to the first stage regulator. I believe this had been done to ensure the safety of the LPG system after the explosion. The piece of pipe which had been removed was lying on the base of the tank footing close to the vapour outlet valve of the gas tank. The gas tank outlet valve had been plugged. 5.4 The LPG tank contents gauge dial was loose and could be moved by hand some 4 percent of the gauge scale. The reading on the contents gauge at that time was 71 percent." (Blue 163P-S) 183Interestingly, the 14.88 tonnes of Propane which was removed from the tank, represents 75 percent of its capacity which confirms Mr Donnelley's observation of a 4 percent movement in the gauge. It can be seen from the above evidence that at the time when Mr Donnelley checked the gauge, the potential 4 percent error operated to under record the contents of the tank. 184The difficulty with the challenge made by the appellants to the amount of LPG which was in the tank after the delivery had taken place on 24 January 2003 is that the filling of the tank to 84 percent of its capacity was part of an Agreed Statement of Facts concerning gas delivery. The agreed fact recorded was: "55 On 24 January 2003 there was a gas delivery of 10.30 tonnes. The tank was filled from between 35 percent and 38 percent to 84 percent. The delivery commenced at 9.30am. The delivery was completed at 11.6am." (Blue 1237B) 185I infer from the way in which the agreed fact is recorded that it was known at the time of the delivery that the contents gauge was capable of being moved 4 percent, hence the uncertainty as to the amount of LPG in the tank at the start point. The unequivocal statement that the tank had been filled to 84 percent, however, allows me to infer that the author of the Statement of Agreed Facts had satisfied himself or herself that it was accurate to say that the tank had been filled to 84 percent capacity at the conclusion of the delivery. 186The appellants' most significant challenge to the "missing gas" theory was that at the conclave it was agreed that the maximum flow through the FSR with the portion of the sealing disc embedded (as it was found after the explosion) was 55 kg per hour (Blue 1487S - 1488H). On that basis, the appellants submitted that the failure of the FSR before the explosion and an unrestricted flow of LPG, beginning sometime shortly after 2pm when the last workers had left the factory, could only have produced a maximum amount of gas of 360 kg. This, the appellants submitted, directly contradicted the "missing gas" theory. 187The riposte by the respondents relied upon the evidence of Professor Masri that there may have been other positions for the embedded portion of the sealing disc in the FSR which would have allowed a greater flow. That evidence was: "10 We agree that, on the models proposed by Professor Masri at paragraphs 3 and 4 of his report dated November 8 2010 the flow through the seat ring orifice would not be unrestricted but may be greater than that through an orifice with an equivalent diameter of 3mm by an amount which would require further work to determine." (Blue 1115D-E) 188Unfortunately Professor Masri did not subsequently determine what the increased flow rate through the FSR could have been. Accordingly, that evidence goes no further than to suggest that the final position of the sealing disc embedded as it was after the explosion, may not have been its position during much of the period before the explosion. If that be so, it would have allowed for a much greater flow of gas but we do not know how much. 189A further response was based on the report of Mr Pearson of July 2004 where he opined, in relation to the "missing gas", that the FSR may have failed progressively over a period of 24 hours between the shutting down of the bakery area at about 9pm Friday 24 January and the time of the explosion. At Blue 394I-395L Mr Pearson set out a number of scenarios which would have accounted for the missing gas. If, as Professor Masri suggested, the final position of the sealing disc in the FSR was not the position which it had occupied for much of the time during which the FSR was progressively failing, then the report of Mr Pearson explains how such a substantial quantity of gas could have escaped. 190In their joint report of February 2011, Professor Masri and Mr Cox agreed that the maximum amount of gas responsible for the fireball was 605 kg, but they did not agree as to the actual amount in the fireball. They did agree that the reason why more gas (if such were present) did not explode was because it could have either leaked out of the extrusion building before the explosion or could have been too rich to burn. (Blue 1118M-T) 191Even allowing for the uncertainties associated with how much gas the FSR in a failed state could have allowed to pass downstream, the evidence of Professor Masri as to the size of the fireball which was observed remains unanswered on the appellants' case. 192I have concluded that the combination of the amount of LPG which was unaccounted for, and of the amount needed to produce the fireball observed, indicates that much more gas leaked from the system than could be accounted for by the time between the explosion and when Mr Verning isolated the tank. This is despite the challenge which has been made by the appellants to the amount of gas which was capable of passing through the FSR if it did in fact fail before the explosion. The submissions by the appellants depend upon it being established that the condition of the FSR with the portion of the sealing disc embedded, as found after the explosion, was its condition from the time of failure until the explosion and that the period during which the FSR had failed so as to allow the passage of excessive quantities of gas downstream was no longer than six and a half hours. Both those assumptions are controversial. The totality of the evidence on this issue, although not without its difficulties, favours the FSR having failed before the explosion. Conclusion as to liability 193The appellants have challenged the primary judge's findings at [194] - [201] on the basis that his Honour failed to properly engage in the scientific controversy which was before him. Their complaint is that his Honour set out a number of conclusions without properly analysing the evidence leading to those conclusions. The appellants submit that had his Honour properly assessed the evidence, he would have found that the respondents had failed to establish causation. In those circumstances, they submitted, judgment should have been entered in their favour. 194There is some force in that complaint. As his Honour pointed out at the commencement of his judgment, he did not propose, nor did he regard it as necessary, to summarise and replicate the very large amount of evidence which was before him. In the circumstances of this case, his Honour was justified in that approach. To have exhaustively set out the evidence, would have made the judgment virtually unreadable and would have served no useful purpose. 195What his Honour did was to identify the key pieces of evidence, to focus on the issue in dispute in relation to that evidence and to resolve that dispute. In doing so, the basis of his reasoning was clear. In that regard, it should be remembered that the areas of dispute before his Honour were many and diverse. It is a tribute to his Honour's fact finding and the professionalism of counsel appearing before this Court, that most of those findings by his Honour have not been challenged. This has allowed the parties and this Court to focus on the limited issues specifically raised in the appeal. 196That his Honour did not extensively analyse the competing opinions on all these issues does not mean that his conclusions were wrong. The evidence before the Court in the appeal, particularly having regard to the unchallenged conclusions by his Honour and to the proper concessions made by the appellants, has enabled this Court to reach its own conclusion and provide more detailed reasons than were given by his Honour. 197At the commencement of the appeal, the respondents submitted that there was an air of unreality about the appeal. The basis for this apparently bold submission was that at trial the major issue to be decided was causation, i.e. whether the acknowledged breach of duty by the appellants, had caused the explosion. At trial, the respondents' theory was that the FSR had failed, the downstream gas lines had become over-pressurised, the filter boxes could not withstand the high pressure and a large amount of gas had leaked through them into the factory and subsequently ignited, causing the explosion. The only contrary theory was that the FSR had not failed before the explosion, there had been a hole in the diaphragm of the OPSO on the regulator to the main burner of the boiler, a small amount of gas had leaked through that hole and into the factory and subsequently ignited, causing an initial explosion, that initial gas explosion then caused massive explosions of combustible dust and finally, the FSR was broken by a pressure wave generated by the explosion that travelled backwards through 77 metres of pipe work and reached the FSR with sufficient force to break it. This was labelled the "pressure wave theory" at trial. 198The only expert who supported the "pressure wave theory" was Mr Cox. In his evidence, he accepted that if his theory were wrong then on the available evidence the alternative theory must be correct. While that concession did not absolve the respondents from proving their case, it meant that if the pressure wave theory were rejected, all of the expert evidence supported the alternative theory. 199The trial judge rejected the pressure wave theory. In the appeal, the appellants did not challenge that finding and abandoned any reliance upon the theory. What they sought to do was to identify pieces of evidence and by reference to that evidence persuade the Court that the respondents had not made out their case on causation. There was no expert evidence to support that approach in a comprehensive way, i.e. there was no expert who had in a holistic way, incorporated that evidence into a theory which explained how the accident occurred. 200Even Mr Cox, whose measurements of the tested gas filter box and the zone 1 filter box formed the basis of the appellants' challenge to his Honour's findings as to causation, was not able to identify any theory of causation other than the "pressure wave theory" or the respondents' alternative. 201What the Court was confronted with was a legal construct, which challenged an important finding, but which offered no cohesive alternative theory, i.e. no alternative theory of causation which explained all the pieces of evidence which had been carefully examined by experts over 46 days of hearing and which was set out in over 1000 pages of expert reports. The best that the appellants could come up with was that there must have been a gas leak at a location other than the zone 1 and zone 2 filters which had allowed gas to escape in sufficient quantities to cause the fireball and explosion. No attempt was made (nor could it be made since this was an approach not followed at trial) to identify where such a leak had occurred, why it had occurred at that time and if it had occurred at this unknown location, what the downstream and upstream effects would have been. That latter consideration was necessary to test the theory of an alternative source of the gas leak by reference to other pieces of evidence. 202Nothing of that kind could be done on an appeal. What was done was to focus on three particular pieces of equipment, i.e. the zone 1 and zone 2 filters, the Bourdon tube pressure gauges and the OPSOs on the second stage regulators to establish that their condition did not support the respondents' theory of causation and in the case of the filter boxes, that it positively disproved it. No attempt was made to explain how if the appellants' interpretation of the evidence were accepted, the explosion occurred. This was a problem for the appellants and justified the criticism by the respondents that the whole focus of the appeal was too narrow and failed to take into account the totality of the evidence. 203The position of the respondents on appeal was the same as that which they had adopted at trial. Their theory of causation depended upon the FSR failing before the explosion. While there were differences between their experts as to the precise sequence of events leading to the explosion, there was unanimity on the fundamental proposition that the downstream gas trains had become over-pressurised, leading to a failure of the zone 1 and zone 2 filters. Given the extent of the destruction, not every aspect of that theory could be proved, but there was a general consistency in all of the evidence and, as the primary judge found, those pieces of evidence taken together, operated as the strands of a cable leading to the necessary state of persuasion which allowed him to find in favour of the respondents. 204While it is true that his Honour, having reviewed the evidence and submissions, set out his conclusions without fully analysing all of the competing pieces of evidence, that exercise has been able to be carried out by this Court. The result of the Court's analysis is to support his Honour's ultimate conclusion. 205While the measurements of Mr Cox of the zone 1 filter box and of the filter box which was tested cast some doubts on the zone 1 and zone 2 filter boxes being the source of the gas leak, they do not for the reasons identified, invalidate that proposition. 206The evidence concerning the Bourdon tube pressure gauges and the OPSOs on the second stage regulators, looked at in isolation is at best neutral and does not positively persuade one way or the other. Importantly, however, the evidence in relation to those pieces of machinery is not inconsistent with the respondents' theory. 207Although the evidence concerning the "missing gas" is not without its difficulties, on balance, it favours a failure of the FSR before the explosion. 208Once the three specific challenges by the appellants to his Honour's conclusions are shown to not invalidate them, the positive evidence in support of his Honour's conclusions is overwhelming. All of the expert evidence, including that of Mr Cox, supports the fundamentals required to substantiate the respondents' theory of causation. There is no rival theory of causation. It follows that the appellants' challenge to his Honour's findings as to primary liability must fail. Contributory Negligence 209As indicated, the appellants only relied upon one particular of contributory negligence, i.e. that the plaintiffs failed in their common law duty to isolate the storage tank when the factory was closed down and left unattended on 24 January 2003. 210Section 5R of the Civil Liability Act 2002 (CLA) applies: "(1) The principles that are applicable in determining whether a person has been negligent also apply in determining whether the person who suffered harm has been contributorily negligent in failing to take precautions against the risk of that harm. (2) For that purpose: (a) the standard of care required of the person who suffered harm is that of a reasonable person in the position of that person, and (b) the matter is to be determined on the basis of what that person knew or ought to have known at the time." 211It was common ground that the tank isolation valve was not closed when the site was unattended on 24 January 2003. Mr Heness was the factory manager. His evidence was that there was never a requirement to close the isolation valve at the tank and that the respondent had never done so. Under cross-examination he agreed that if required, it could have been done but it would delay start up and be impractical. He agreed that there was nothing to stop the respondents isolating the gas tank had they wanted to on 24 January 2003 (T.186). 212Mr Goldring was a director of each of the respondents. It was his evidence that gas was turned off at the individual appliances, but not at the gas tank. His evidence at the coronial inquiry, which he confirmed at trial, was: "Q. Did anyone turn off the gas supply then? A. No. Q. That you know of? A. No, not to my knowledge, no. It wasn't a regular practice, sorry. We certainly turned the gas off at the appliances otherwise gas would continue to leak into the plant. But there, there wasn't a standard practice to turn off the gas at the tank, in fact the tank was out of bounds to our people." (T.629.38) 213Other evidence given by Mr Goldring was: "Q. But that risk could have been avoided, couldn't it, if you had a procedure to turn the gas off at the tank, isn't that right? A. I don't know about that, I wasn't aware of any need to turn the gas off at the tank. Q. Mr Goldring, common sense would tell you that if you are not going to be in the factory for a couple of days and you have a 43,000 litre gas tank, the proper procedure would be to turn the gas off at the tank, wouldn't you agree? A. No, I don't agree. I considered it really to be like electricity or water. You don't turn those off at the mains when you leave the factory. Q. So you see 43,000 litres of gas in much the same way as water? A. Or electricity, yes. They are utilities. Q. But if you light 43 gallons of water it won't blow up, will it? A. No, it can leak into the factory and cause problems. Light power electricity can cause major problems as well." (T.630.41-631.7) "Q. You see Mr Goldring, you didn't shut down the tank or isolate it when you were shutting down the factory because it caused a problem in starting up, isn't that the reason? A. I think that's something that was passed onto me from the factory manager, he talked about that. I wasn't familiar with that. Q. Do you agree with me? A. I don't really know, that's information I was given by the factory manager at some stage. Q. You don't really know? A. No, I don't know. I didn't have experience. The other factories I worked at we carried out the same sort of philosophy, we never shut down the tank." (T.633.37-48) 214His Honour resolved the question of contributory negligence at [317] - [321] (see [75] hereof). Not only did his Honour find as a fact that neither Mr Heness nor Mr Goldring, nor anyone else in the respondents' employ, was aware of a requirement under the Australian Standard to isolate the gas tank when the factory was unattended, he found that no-one from the appellants had provided that information to them. His Honour found that the likely explanation was that under the contract between Origin and BestCare the responsibility for keeping the equipment safe was undertaken by Origin. 215The appellants submitted that the primary judge erred in his approach to contributory negligence in that he focused on the actual knowledge of the respondents, rather than upon what a reasonable person in the position of the respondents would have done in relation to the isolation of the tank. In other words, his Honour had not addressed the correct legal question. The appellants submitted that had his Honour done so, he would have concluded that, because of the inherently dangerous nature of the installation, a reasonable factory operator in the position of the respondents would have isolated the tank. This was particularly so when all that was required to do so was the closing of an isolation valve. 216As the terms of s 5R(2)(b) CLA make clear, actual knowledge has a part to play in the assessment of contributory negligence. The combination of the statutory requirement and of the common law was considered in Council of the City of Greater Taree v Wells [2010] NSWCA 147; 174 LGERA 208 where Basten JA said: "107 The assessment of the plaintiff's conduct involves a quite different exercise. A critical difference between the assessment of negligence and the assessment of contributory negligence is that the purpose of the latter assessment is to allow for an apportionment of responsibility for the injury by a reduction in the damages recoverable by the plaintiff "to such extent as the Court thinks just and equitable having regard to the claimant's share in the responsibility for the damage": Law Reform (Miscellaneous Provisions) Act 1965 (NSW), s 9(1). That is a different exercise from the determination of whether or not the defendant has been negligent. 108 A further important difference in approach in assessing the negligence of the defendant, as against the contributory negligence of the respondent, involves the degree of precision by which the activity, including relevant states of knowledge and understanding, is to be identified. ... With respect to the plaintiff, the focus of the evidence is often quite different. Although the ultimate question is what a reasonable person in [the plaintiff's] position would have known and done, it is inevitable that the evidence will focus upon the knowledge, understanding and actions of the plaintiff himself, shortly prior to the accident, in part to determine whether he exercised reasonable care, but also to assess what would be reasonable care in the specific circumstances." 217As s 5R and that extract from the judgment of Basten JA in Wells indicate, the correct legal question is whether a reasonable person in the position of the respondents, i.e. having the knowledge which the respondents had or ought to have had, was negligent. In this case, the factual finding of his Honour was that the respondents did not have actual knowledge of the need to isolate the tank. The only inquiry is whether they ought to have had such knowledge. 218There was no expert or other evidence as to what persons operating a factory such as this should have done with respect to isolating a gas tank. There was an Australian Standard but it had only come into effect in 2002 and only applied to installations which commenced operation after 10 May 2002. The only evidence as to a general practice came from Mr Golding when he said: "The other factories I worked out we carried out the same sort of philosophy, we never shut down the tank." Neither that evidence nor any other evidence supported the proposition that the respondents ought to have known that the tank should have been isolated during a shut down. 219On that state of the evidence, I am not satisfied that a reasonable person with the knowledge of the respondents, and in their position, would have isolated the tank before the shut down on 24 January 2003. This is particularly so when Origin, the entity responsible for the safe operation of the installation, had not provided any information or advice to that effect. His Honour was correct to reject the appellants' defence of contributory negligence. CONCLUSION 220The orders which I propose are that the appeal is dismissed and that the appellants pay the respondents' costs of this part of the appeal. 221WARD JA: I agree with Hoeben JA.