SECAN LTD v. WONG PING WAI

HCCT93/1999

IN THE HIGH COURT OF THE

HONG KONG SPECIAL ADMINISTRATIVE REGION

COURT OF FIRST INSTANCE

CONSTRUCTION AND ARBITRATION PROCEEDINGS
NO.93 OF 1999

(transferred from High Court Action No.5227 of 1996)

---------------------

----------------------

-----------------------------

J U D G M E N T

-----------------------------

1. About an hour or less before midnight on 8 May 1993 a flood started in Chiller Plant Room B underneath the commercial buildings at the centre of the development known as South Horizons, which at the time was under construction. The flood continued for about an hour but in that time the water in the plant room rose to about two-feet in depth and damaged machinery installed therein, much of it electrical, to a value in excess of $4 million.

2. By these proceedings, now over 10 years later, the developer, ("Secan"), blames the building consultant engineers ("ACE") for the flood. It is a subrogated claim brought in Secan's name on behalf of the insurers. Secan is the sole plaintiff, ACE, the sole defendant.

BACKGROUND

3. Secan had appointed ACE as building services engineer in relation to the seawater supply and control system and flushing water system in July 1989. Later, Secan appointed Ryoden Engineering Co. Ltd ("Ryoden") as the main contractor for the supply, installation, testing and commissioning of water chilling Plant Room B ("PRB"). They also appointed Hsin Yieh Architects & Associates Ltd as the architects for the whole development. Other parties involved in work related to the PRB were as follows. Defoe Fong Construction Co. Ltd ("Defoe Fong") built the superstructure, Carrier Hong Kong Ltd ("Carrier") supplied the chillers and "Inchcape Boag" supplied a central control monitoring system ("CCMS"). The CCMS was not in operation at the time of the flood. It had not yet been constructed. Once completed and installed the system became fully automated. However, that time had not yet arrived. We are concerned therefore in this case, with the system without a CCMS.

4. Defoe Fong had completed its work in 1992 and had handed over to Ryoden. Ryoden had completed the major installation works in early 1993. The practical completion date for Ryoden's contract was certified to be 8 April 1993. The flood occurred about a month later.

THE SYSTEM

5. It is helpful to have an outline understanding of the water system. The PRB contains four chillers. Sea water is pumped into the PRB through 600mm sea water pipes. Each chiller has a condenser inside. The sea water passes through the condensers in the chillers.

6. The workings of the condensers, chillers and air-conditioning system do not concern us in this case. This case concerns the flow of water which is returned back to the sea after it has been through the system. The water which is discharged is warm water. It flows back along the sea water return pipe and some of it is diverted into the flushing water tank instead of going directly back to the sea. The flushing water tank provided up to 75,000 litres of water for over 10,000 flats in South Horizons. The agreed evidence as to how the flushing water system operates is as follows :

(1) When the out-going warm water gets to a "T connection" along the 600mm return pipe, it may be diverted to the flushing water tank along a 300mm pipe. The water that is not so diverted is discharged back to the sea.

(2) Whether water would be discharged back to sea or taken off to the flushing water tank depends on the water level in the flushing water tank.

(3) If the water level inside the flushing water tank is low the ball float valve will be open and draw water into the tank.

(4) If the flushing water level inside the flushing water tank is high the ball float valve will close and the water will "back up" in the connecting pipe; as a result, no more water can enter the pipe at the T connection and it will be discharged back to sea.

(5) In normal operation, water level drops due to consumption of flushing water. The valve in the flushing water tank usually open/closes many hundreds of times every day. It provides flushing water for 34 residential blocks, the East and West Commercial Complex, Hong Kong Electric and the Club House.

(6) If the ball float valve fails, water will continue to fill the tank through the 300mm pipe and the water level inside the tank will continue to rise. In normal circumstances, the rise in water level will be detected by the ultrasonic sensor which should then close the motorized valve to stop water flowing into the tank and also send out alarm signal to security officer.

(7) If the ultrasonic sensor does not close the motorized valve and water continues to enter the tank, it will overflow through the overflow pipe. This has a diameter of 350mm. It is an open ended pipe like a tap. The overflow water leaving the tank via the overflow pipe will be discharged onto the floor of PRB.

(8) During time when all chillers are off line and sea water is supplied direct via the 600mm supply main, this ultra-sonic level sensor would transmit a signal to shut down the sea water supply pump when the flushing water storage tank has reached full capacity.

7. Because the system went into operation before the CCMS was functional, ACE sent a letter to Secan's Holding Company (H.K. & Whampoa Dock Co. Ltd) on 19 March 1992 confirming Secan's agreement for the manual operation of the plant before the CCMS was completed. That required amongst other things, the manual switching off of the chillers each night and the manual turning off of the sea water pumps at about the same time or soon after.

8. On the night of the flood, the chillers were switched off but the sea water pumps were not. On the night of the flood, the ball float valve malfunctioned, the ultrasonic sensor may not have performed as it should have done and the motorized valve did not close. As a result, the flushing water tank overflowed, the water gushed on to the floor of the PRB. The room was unattended so nobody saw or heard the water flooding onto the concrete floor and the water could only seep away through floor drains and closed doors. The doors opened inwards so did not burst open with the weight of the water which rose to a level of about 24 inches before the water was eventually turned off.

THE LAW

9. Counsel for the two parties to this action, Mr Thomas Lee for the plaintiff and Mr Colin Wright for the defendant differ only slightly as to the applicable law in this case. Mr Lee's primary position is that ACE are in breach of their common law duty of care in tort. It is true that the parties have a contractual relationship as well and thus, an implied contractual duty exists as well. Mr Lee, however, for good reason, pursues only the duty in tort. He places the plaintiff's eggs in the tort basket, acknowledging that if he cannot succeed in tort he could not succeed in contract either. The only questions which arise are therefore, was ACE under a duty of care and if so what was the standard of care? Then the plaintiff must prove that ACE failed to exercise the degree of care and skill imposed on them. The plaintiff has narrowed its case down to two areas, the ultrasonic sensor and the overflow pipe. The plaintiff aims to prove firstly, that ACE failed to exercise due care and skill in the design of the flushing water tank overflow pipe. Secondly, that the incorporation of the ultrasonic sensor into the system and the testing and commissioning of it, was also carried out without the requisite care and skill.

10. Mr Wright submits, as a preliminary argument that, as the design included the CCMS and as the CCMS had not been installed at the time of the flood, then no duty had, at that time, arisen. Their duty was in relation to the finished design only. I reject this argument. ACE knew that the system for which they were responsible was put into operation without the additional benefit of the CCMS. It was because of this that manual attendance and manual operation was provided for in the early stages. The agreed provision of manual attendance was to do the very tasks which the CCMS would do once it was ready. The documentary evidence (from minutes of meetings) supports ACE's contention that at the time manual attendance was agreed, namely in early 1993, it was anticipated that the CCMS would be ready in about July 1993. The fact that it did not actually start running till 1995 is irrelevant.

11. At the time of the flood therefore, ACE was responsible for the design of the system which had been put into operation. Their duty was as set out in the scope of work schedule in the contract namely, inter alia the "design tender documentation and supervision of a ... 'Sea Water Supply System'." There can be no doubt that the word "supervision" refers to the supervision of the installation, testing and commissioning of the system not the supervision of its operation ad infinitum. It was therefore plainly not their job to supervise the tasks being done by the manual operators temporarily, pending the start of the CCMS. Those people were employees of Ryoden and the Estate Management Office.

12. The "slight difference" between counsel to which I referred earlier concerns the appropriate standard of care of which Secan must prove that ACE fell short. Mr Wright contends that the test in Bolam v. Friern Hospital Management Committee [1957] 1 WLR should be applied in this case. The standard advanced in that case is the reasonable skill and care of an ordinary skilled person professing to have the special skill in question. Further, such a person would not be negligent if he acted in accordance with a practice accepted at the time as proper by a responsible body of professional opinion, even though other members of that body may adopt a different practice. Bolam was a jury case and the above is from the approved direction to the jury on the issue.

13. Mr Lee for Secan submits that the approach adopted in the Australian case of Rogers v. Whitaker (1992) 175 CLR 429 should be followed. That was a medical negligence case in which it was stated that the views of members of the profession may be relevant and important but it is for the court to decide what is the appropriate standard of care (and whether it was breached) in any given case. In my judgment the difference between the two is wider in theory than it is in practice.

14. Ultimately it is always the court's task to make a decision on what the professional man should have done. To that end the court will invariably seek guidance from expert opinion. The court will frequently have to accept or reject conflicting expert testimony. Having considered all the professional opinion evidence and having, where necessary, rejected parts of it and accepted other parts of it, the decision for the court to make is, on the facts of the case in issue, did the defendant fail to carry out his responsibility competently, carefully and skillfully to that high standard which his profession expected of him. Perfection is not required, an ability to forsee and cater for all eventualities is not required. However, in this case, the court must ask itself the question would the relevant professional body, on examining the water system in place on 7 May 1993 have said that is a competently designed and skillfully installed system? Mr Wright, in this context, referred the court to two extracts with which I agree and adopt in my consideration of the evidence in this case. They were Bingham LJ in Eckersley v. Binnie & Partners (1988) 18 Con LR at p.80.

"The law does not require of a professional man that he be a paragon combining the qualities of polymath and prophet."

and from Jackson & Powell on Professional Negligence at 8-163 :

"The standard of reasonable care and skill is not a standard of perfection. It does not make an architect, for example, the insurer or guarantor that work has been properly done."

THE WITNESSES

15. As the outcome of this case will depend, to a large degree, on the opinions of experts, it is appropriate at this stage to refer briefly to witnesses, particularly the experts, who gave evidence at trial.

16. Secan called three witnesses as to fact. Mr George Ryder was the site engineer at the time. He was in fact employed by ACE but had been seconded to Secan at the material time. Mr K.L. Wong was a manager from the developer's estate management office and Mr David Woo was a loss adjuster who attended and reported on the flood. The most important was Mr Ryder. He was a man of great experience who gave evidence in a helpful and straightforward manner. He gave the impression of complete impartiality. His only concern was in the search for the truth. When he was unable to recall events because of the passage of time since the incident he plainly said so. He did not shirk from accepting matters put to him by the defence, when he believed it was right to do so. At the end of the day, the defence were able to pray in aid much of Mr Ryoden's evidence. Although called as a witness of fact, he also had the advantage of considerable expertise. The defence factual witness was Mr Jeffrey Leung. He was an assistant director of ACE and had been in their employ since 1987. He also was impressive and honest. He was heavily involved in the works at the time and had a surprisingly good (considering it was 10 years ago) recollection of events. I felt able to rely on Mr Leung's evidence on controversial issues with some confidence.

17. The plaintiff's expert was Mr W.K. Lai, a mechanical engineer with Binnie Black and Veatch H.K. Ltd. The defendant had engaged the services of Dr W.H. Au, a director of Ove Arup & Partners H.K. Ltd. I make no specific criticism of either expert. Both were professionally orientated and desirous of assisting the court. In general terms however, Dr Au appeared the more experienced, was slightly more highly qualified and adopted a more measured and pragmatic approach to the issues. His evidence remained in tact after cross-examination. It would perhaps be helpful at this stage therefore, before embarking on factual matters relating to the events of 8 May 1993, to set out in outline Dr Au's professional opinion, upon which the court has placed considerable reliance. The crucial parts of Dr Au's reports are :

"I am of the opinion that the overflow of the flushing water tank was not caused by any failure in the design. Rather, the overflow was caused by a combination of a number of events: (1) the malfunctioning of the ball float valve; (2) the malfunctioning of the ultrasonic sensor ; (3) non-completion of CCMS and (4) the general failure of Ryoden to provide attendance or to arrange for the switching off of the seawater pumps....

...

Based on the above findings, it is clear that Ryoden was still carrying out testing and commissioning work on the date of incident. Ryoden should have provided attendance throughout the whole testing and commissioning period. Being an experienced contractor, Ryoden should also have arranged for the switching off of the seawater pumps after testing as it had done on the other days of testing. In my opinion, it was totally unforeseeable that Ryoden would fail to provide this attendance or that they would fail to arrange for the seawater pumps to be switched off.

Conclusion

It is clear that the accident was caused by the combination of the above 4 events. The Defendant had exercised great care in the design to ensure that a safety provision was in place in case any of these failures did occur. In my opinion, it was totally unforeseeable that the 4 events occurred simultaneously and caused the flooding on 8 May 1993. The chance of these 4 events occurring simultaneously was extremely remote."

DOCUMENTARY EVIDENCE

18. A final preliminary matter concerns the admissibility of one document. There was a considerable volume of uncontested documentary evidence, much of which was of great significance because none of the witnesses who testified had any first hand knowledge of the events on the night of the flood.

19. ACE, however, strongly challenged the admissibility of a fax dated 10 May 1993 handwritten by a Matthew Chan which set out a brief timetable which purported to set out the sequence of events on the night. Mathew Chan therein recorded information that was given to him from another source. To say the least, it is a document of little value. It is however, arguably, admissible in law under the provisions of section 49 of the Evidence Ordinance. However, this is of little comfort to Secan because its vague provenance and dubious accuracy makes it evidence upon which no weight can be safely attached. The timings contained in the fax conflict with common sense, the probabilities and other evidence. For this reason, it would serve no useful purpose to set out its contents.

THE EVENTS OF 8 MAY 1993

20. What actually happened has been the subject of reports and debate between loss adjusters, insurance companies, employees, experts and lawyers over the last ten years. The standard of proof in this court is a balance of probabilities. What follows therefore is an attempt to set out the most likely scenario based on the evidence. If anyone could be 100% certain what actually happened this litigation would not be taking place.

(a) Firstly, the sea water pumps should have been turned off in the late evening after the chillers had been switched off. Had the CCMS been in operation this would have been done automatically. They should have been switched off manually, the responsibility for which was not ACE's, but they were not. ACE had, in fact, advised Secan that manual attendance from Ryoden should be arranged if the system was to proceed without the CCMS. ACE's actions in the matter of the provision of manual attendance cannot be faulted. To state the obvious, if the pumps had been switched off the flood would not have occurred.

(b) Secondly, the pumps being on meant that excess unwanted water was being pumped into the system. The computer print out shows that two sea water pumps were switched on at 0741 hours but not switched off till 0043 the next morning, i.e. after the flood. Normally, at the beginning of night time operation the flushing water tank is designed to be full and the ball float valve at the top of the tank is closed. During the night, the level of water would slowly fall whilst residential toilets were being used. The capacity of the tank, 75,000 litres, was regarded as being more than enough for one night.

(c) Thirdly, the ball float valve malfunctioned. This caused two problems, firstly excess water went into the tank and secondly, it went into the tank at an abnormally fast rate. Before considering these two problems, the reason that the ball float valve malfunctioned must be considered.

21. It is common ground that the valve itself was not defective. It was a good quality industrial ball float valve. In many many water systems, the ball float valve together with an overflow pipe are the only precautions against flooding. It is a tried and tested reliable mechanical device. The only explanation for its failure to close on the night of the flood was the fact that foreign particles and debris became lodged in the mechanism and jammed its normal operation. The reason that the quantity of debris was sufficiently great so as to, effectively, break the valve was, again, twofold. The abnormally fast flow of water caused excess debris to be moved along the pipes from the sea. Additionally, the pipes were new and were concrete lined on the inside. Small bits of concrete from the newly installed pipes probably became dislodged by the fast moving water and added to the debris.

22. Thus far, Secan lay no blame at the door of ACE. They do not say that the failure to switch off the pumps was ACE's fault, neither do they say that the failure of the ball float valve due to debris, was ACE's fault.

23. I find as a fact that the situation which soon developed as a result of the failure to switch off the pumps was one described by the witnesses as a "dead head" situation. It seems logical to me to accept, on the evidence, that because the pumps kept on pumping, the chillers were off and pressure therefore built up. The ball float valve would become a weak point. It became jammed for reasons already stated and water flooded into the flushing water tank under greater pressure than normal and therefore at a much faster flow rate than normal.

(d) The fourth event concerns the ultrasonic sensor. Normally this was designed to measure the water level inside the tank by conventional sonar pulses. It was designed to be set so that when the water reached a certain level, it would automatically operate a motorized valve which would avert any danger of flooding. It also sent out an alarm signal if the water reached a danger level.

24. Much of the evidence at trial centred on the sensor. It was plainly a sophisticated piece of equipment which provided an additional fail-safe in ACE's design.

25. Whether or not it actually sent out a buzzer alarm to the control room on the night of the flood (which it was designed to do) is not clear on the evidence. It may have done. If it did, the buzzer alarm was not heard because (i) there was no-one there to hear it and (ii) the flood situation would have caused excess noise. If the buzzer did go off, it may have later stopped because of the failure of the electrics caused by the flooding water. These matters are speculative and I make no specific findings about them.

26. However, it is possible, on the evidence, to make the following finding. The situation which developed prior to the point when the sensor would normally do its job had become abnormal in two respects. Firstly, water was coming into the tank at an abnormally fast rate and secondly, the fast inflow caused abnormal turbulence on the water surface. As already pointed out, neither of these two factors had been caused by a failure on ACE's part.

27. It is thus now necessary to consider the degree of abnormality. Cogent evidence leads me to conclude that the flow rate into the tank prior to the start of the flood was four times greater than the normal situation at the material time. This high flow rate, caused by the dead head situation, caused excessive turbulence. The question thus arises can the sensor be blamed for not responding (if indeed it did not respond) to such an abnormal situation.

28. Two factors require consideration to answer this question. Firstly, its installation and secondly, its capabilities. I am satisfied that it was installed precisely in accordance with the manufacturer's recommendations. It would not send out an accurate reading if it was too close to the water surface, accordingly it had to have a minimum distance between it and the water. Much evidence was given on this issue. The outcome of that evidence is that I find that ACE did not fail in its duty of care when it approved the positioning of the sensor. It is true that it was adjusted after the flood. This does not mean that it was wrongly positioned before the flood. Any professional, looking at the installation before the flood, would have approved its positioning.

29. What were its capabilities? In the context of this case, it is better to answer this question by putting it into the negative what was it not capable of and not designed for?

30. Any sophisticated piece of electrical engineering has its inherent limitations. This sensor was designed to deal with those sorts of situations which it was reasonable to expect and foresee within a large industrial water tank of the type in the basement of South Horizons. Normally, the water would rise and fall at a moderate pace and the water surface would remain reasonably flat. I am satisfied that given such conditions, the sensor would always operate in accordance with expectations. I am also satisfied that the conditions which prevailed, which may have caused the sensor not to perform as required were plainly outside those which one could reasonably have expected or foreseen.

31. The fact that the motorized valve did not function was the final event prior to the spillage of water commencing. At this stage, the following differences may be noted, as correctly submitted by Mr Wright, between what happened on the night of the flood and a normal operation of ACE's design :

(1) The pumps were not automatically switched off at the same time as the chillers.

(2) The sea water pumps created a deadhead situation.

(3) Above normal water pressure was exerted on the ball float valve.

(4) On the opening of the ball float valve, the water flow rate past the ball float valve into the flushing water tank would have been approximately four times the maximum rate experienced during operation in accordance with ACE's design.

(5) The high pressure and flow rates forced particles and debris in the system along the pipes until they met an obstruction.

(6) All the particles and debris flowed towards the ball float valve, whereas during daytime operation in accordance with the ACE's design the majority of the particles and debris would have been carried harmlessly along the sea water outfall to the sea.

(7) The high flow rate of water into the flushing water tank resulted in the signal generated by the ultrasonic sensor becoming unstable with the result that the motorized valve failed to close.

(8) The control rooms of both the sea water pumphouse and the PRB were unmanned, with the result that if alarms signals were made, they went unheeded.

(9) The PRB was unattended with the result that the discharge of water onto the floor of the PRB went unnoticed.

(e) The fifth event is the fact that water started to spill onto the floor. Again, two matters require consideration, firstly, ACE's design and positioning of the overflow pipe and secondly, the effect of human error, if any, after the flood had started.

(i) The overflow pipe

32. The pipe itself was 350mm in diameter (approximately 14 inches). It was shaped like a domestic tap in that it emerged from the side of the tank and then turned towards the floor. The end of the pipe was about three feet above the concrete floor. It is part of Secan's case that this design was a contributory cause to the damage which resulted.

33. A conventional overflow pipe is normally sited in a conspicuous place. If water overflows it can then be seen and heard. Secan submit that this overflow pipe was not conspicuous and thus ACE's design was flawed. They submit that either, it should have been routed into the adjourning car park so that security guards or residents would notice the overflowing water or, it should have been routed directly into a storm water drain. In my judgment Secan have failed to discharge the burden of proof on them to satisfy me that either of these contentions are correct.

34. It is true that after the flood the pipe was extended and re-routed directly into the storm water drain. However to have done this at the time of the original installation would not have been consistent with the Waterworks Regulations, Cap.102. Regulation 6 in Part III provides :

"An overflow pipe .... shall be fitted to each cistern and shall be extended to terminate in a conspicuous position. No overflow pipe shall be connected to a drain, sewer or to the overflow pipe from any other cistern."

35. Under the revised design, the end of the pipe is visible to no-one. It is not possible to hear or see water overflowing. It is not possible therefore to know if the tank is overflowing. Had such a scheme been incorporated in the original design it would rightly have been criticized.

36. The alternative suggestion that the pipe should have ended up in the car park is equally untenable. A developer would rightly object to a 350mm pipe pouring water into an area where residents frequent and their cars are parked. In any event, at the time of the flood the flats were unoccupied and thus there would have been no residents to notice the water.

37. The choice of having the overflow pipe where it, in fact, was was, in my judgment the correct one for the following reasons.

38. Firstly, it was conspicuous. The PRB should have been manned at the time and would have had staff in attendance nearby even after the completion of the CCMS. Simply because the CCMS created an automatic system did not mean there would be no human involvement or presence in the room at all. The adjacent control room was plainly intended for human occupation and it was an integral part of the PRB. Secondly, as intended and designed, it was a "nuisance" feature. The sight of copious amounts of water flooding onto the floor creates the natural response of fast and urgent action to stop it.

39. In short, the position of the overflow pipe was entirely in accordance with the prevailing regulations at the time, a skilled and competent professional would have approved it at the material time and thus it cannot be said to have been negligently designed or positioned.

(ii) Human involvement after the flood started

40. It is clear from the evidence that the flood started after 11:00 p.m. At that time, Ryoden staff had left the room unattended. It is not clear exactly when the staff left or how long he was away. It is known however that the level of water rose to about 24 inches maximum. Given the known flow rate of the water escaping from the pipe, the dimensions of the room and making an allowance for some water escaping through the floor drainage channels and seepage through the closed doors, it is more likely than not that the flooding continued for about one hour.

41. It is also not clear if, when the flood was first discovered, the water was at its maximum height or lower. If it was at its maximum height and steps were immediately taken to stem the tide, it follows that the staff were absent for about one hour plus the time between the staff first leaving and the flood starting. That would be an unacceptably long time. If the staff member returned earlier but failed to deal with the problem immediately, by either forcing open the door or contacting the EMO to turn off the pumps or both, then much of the damage caused by the fact that the water was two feet deep would have been caused by the negligent conduct of the staff member. If for example the water had been only one foot deep it would have been easily possible for an average working man to push open the door, inwards, and allow the water to escape. It was also noted that the equipment in the PRB was constructed on concrete plinths about 14 inches high. It is fair to say that as far as that equipment was concerned little or no damage would have been done to it until the water level rose above the concrete plinths.

42. In short, I find that Secan were at fault in failing to ensure that Ryoden properly manned the room, and that they failed adequately or at all to alleviate the flood on discovering it.

CONCLUSION

43. Secan have concentrated their claim against ACE on the ultrasonic sensor and the overflow pipe. In neither case have they shown it to be more likely than not that the design or positioning of the sensor or the design or position of the overflow pipe material contributed to the cause of the flood. On the contrary, I am satisfied that both the sensor and the overflow pipe were necessary and adequate constituent parts in an overall system which contained a number of precautions against flooding and which would have been regarded, on 7 May1993, by any experienced engineer with particular skills in this field, as being competently designed, competently installed and fit for its purpose.

44. The entire sequence of events was described by Mr Ryder as a "one in a million" situation. In other words, the design cannot be criticized for failing to cope with a combination of circumstances which were plainly unforeseeable. It was a freak accident but one about which it is appropriate to make the following final comment. The sequence of events was probably, in abbreviated form, as follows :

- failure to switch off sea water pumps due to communications breakdown between Ryoden and EMO.

- PRB left unmanned by Ryoden staff for over an hour

- deadhead situation occurs

- ball float valve malfunctions

- tank fills fast

- ultrasonic sensor gives unstable reading

- ultrasonic sensor may not have given out alarm buzzer

- ultrasonic sensor alarm, if sounded, is not heard

- motorized valve does not operate

- tank overflows

- PRB, still unmanned, fills with water because inward opening doors are closed.

- staff return either too late or if earlier, fail to alleviate continuation of flooding.

45. The happening of each of these events depended entirely on the one which preceded it. The human errors are the first and last in the sequence. Human error set the ball rolling at the start and human error failed to bring it to an early end. But for human error, ACE's design would have worked and no damage would have been caused. It was unforeseeable human error, by non ACE employees, which "lit the blue touch paper" and set in motion an unforeseeable sequence of events. Inevitably, the unforeseeable sequence of events led to an unforeseeable situation. It was an extreme situation with which ACE's competently designed and installed system could not have been expected to cope. The water therefore overflowed. All the resulting damage was caused directly by (a) the fact that the PRB remained unmanned so that no-one saw or heard the water and (b) when the non ACE staff did return they had either been away for too long or failed to deal with the situation as they found it adequately or at all, or both. If, as Mr Lee submitted, they may have panicked, it was panic caused by a situation created at their own hands.

46. In all these circumstances, I give judgment for the defendants with costs.