Environmental Protection Authority v Caltex Refineries

Environmental Protection Authority v Caltex Refineries - [2006] NSWLEC 335 - NSWLEC 2006 case summary — Zoe

[12]

Foul condensate system 5 Foul condensate (or sour water) is a waste by-product of various processes at the refinery, generated when water comes into contact with sulphur rich liquid or gaseous hydrocarbon streams. Foul condensate looks like grey water and is characterised by its rotten egg odour. 6 At the premises foul condensate is stored in large fixed roof storage tanks 216 and 217 prior to being processed. These tanks are located in Area 1 of the refinery. They operate on a weir/loop seal system which allows excess oil to be skimmed from the surface of tank 216 and decanted to tank 218. Sometimes an aqueous layer, called a "water bottom", can build up below the oil layer in tank 218. If an unusual amount of water bottom builds up in tank 218, it may be transferred to a slops tank, such as tank 12, although such a transfer (from tank 218) is not a regular occurrence. 7 Foul condensate is processed in plant known as the "sour water stripper" ("SWS") (or the back up plant known as the "foul condensate unit" "FCU"). The SWS and FCU are essentially distillation columns, their primary function being to process foul condensate that has been generated as part of the oil refining process. The FCU is used when the SWS is being maintained or when there is excess foul condensate. It operates in much the same way as the SWS. 8 Foul condensate is first pumped from tanks 216 and 217 to the SWS accumulator. The accumulator incorporates a series of baffles and weirs which allow any residual oil to be separated out from the foul condensate. The foul condensate is then preheated and fed into the SWS column. The column uses steam to strip hydrogen sulphide and ammonia from the foul condensate. The mixture of gas and steam produced in the column is then routed to the waste gas incinerator. Incineration of hydrogen sulphide and ammonia produces, among other things, water vapour, sulphur oxides and nitrogen oxides. The stack serving the incinerator (as well as four other emission streams) is known as F10. The stripped foul condensate/water from the SWS column is then piped so that a proportion of it goes back into the crude oil feed lines while the majority goes to the desalters where it is used to strip salt from the crude oil.

[16]

Wet slop transfer 16 At 8.15am on 17 October 2004 Caltex commenced a transfer of 550kl of wet slop from its Banksmeadow Terminal to wet slop tank 11 at the premises. The transfer was completed by 2.20pm on 17 October 2004. The slops were intended to be transferred exclusively to tank 11 but, owing to the fact that jumpover valve A had been mistakenly left open after the water bottom transfer, about half of the slops accidentally flowed to foul condensate tank 216. Because the OMC (which manages transfers of this type) monitored a steady rise in tank 11 throughout the slops transfer, they did not suspect that some of the slops transfer was mistakenly flowing to tank 216. 17 Tank 216 has a capacity of 840kL. For four and a half hours from 8.15am the level in tank 216 rose at an unmanageable rate of 10 per cent or 70kL per hour. On noticing the rapid rise in level in tank 216, the Area 3 technician responsible for monitoring the tank incrementally increased the feed rate from tank 216 to the SWS. He did this in order to try and manage the rising level of tank 216, unaware that slop (not foul condensate) was entering tank 216 and thus containment of the material entering tank 216 was beyond his control. From 9.19am to 9.45am the technician increased the feed rate from tank 216 to the SWS in steps from 16 to 22m3/h. At 12.05pm the technician again increased the feed rate in steps from 22 to 32m3/h. 18 Between approximately 1pm and 2.20pm tank 216 overflowed. As a result, approximately 100kl of slop and foul condensate overflowed into the tank bund surrounding tanks 216, 217 and 218 ("the overflow liquid"). The overflow liquid was isolated in the tank bund and did not reach the main refinery oily water sewers or the waste water treatment plant. 19 A high level alarm set on tank 216 to detect overflows was set at a rate that did not detect this overflow. Black smoke was emitted from the furnace because slop had entered the SWS causing material to be burnt which caused black smoke. This was the first indication that there was a problem in the plant and action was taken within minutes of seeing the smoke to stop its emission. The smoke lasted for about nine minutes.

[24]

Analysis of environmental harm 29 The Prosecutor relied on the evidence of a consultant physician Dr Kenyon, and the Defendant relied on the evidence of an environmental chemist Dr Howell. They prepared a joint statement of matters about which they agreed and disagreed, a useful process in which the Defendant fully cooperated. They agreed that the symptoms experienced by residents resulted from the foul condensate spill on 17 October 2004 and that the psychological impact of the offensive odour could cause the physical reactions reported. They disagreed on whether some of the adverse physical reactions were the result of exposure to toxins in the odour, that is, as to whether there were toxicological effects. Dr Kenyon stated: … (ii) A second type of health effect experienced by the residents is most likely to be due to the action of the chemical components on tissues, particularly mucous membranes, eg, the eyes, the throat and respiratory tract. Examples of such effects are asthma attacks, irritated eyes and sore throats. These are regarded as true toxicological effects. (iii) It is my opinion that the health effects suffered by Marion Gerard's son (asthma attack and home from school), Pamela Rugless (asthma attack and away from home two nights), Janice Wentworth-Perry (watery eyes and asthma symptoms) and Caroline Howard (sinusitis) are more likely than not to be caused by the action of hydrogen sulphide gas, or other contaminant gases, on mucous membrane tissues. (iv) In the absence of any quantitative data on the composition of the spilt material, I rely more heavily on the residents' reports of health effects and conclude that it is probable that the concentrations of hydrogen sulphide in the community were sufficient to have caused these effects by a toxicological mechanism, that is, by the direct action of hydrogen sulphide (and other gases) on tissue. 30 Dr Howell stated at par 25: … (ii) Highly odorous compounds have been shown to cause a range of symptoms including those described in the residents' complaints without having a direct toxicological effect on the tissue. (iii) The specific circumstances of the incident, namely the distance from the residents and the actions taken to reduce the rate of volatilisation by Caltex are such that it is more likely that the concentrations of hydrogen sulphide were less than those associated with direct toxic effects but greater than the odour threshold.

[27]

first set of instructions When Plt2 SWS is shut down, & Plt 45 SWS is on line, carry out the following: …5 (OMC) Check the Plt 2 SWS feed line & jumpover to slop is open… second set of instructions When the transfer is complete… …6 (OMC & SEPS) Place all lines & systems back to normal. (ii) Safety Performance Self Assessment (SPSA) procedure for risk assessment for the non-routine transfer of water bottom tank 218 to tank 12 33 Mr Knight advised that non-routine uncontrolled procedures are written by experienced and knowledgeable operators in specialist roles. He stated the team of refinery technicians responsible for completing the task and reviewing the procedure are expected to raise any issues of concern with the author. The technicians would be expected to perform SPSAs and apply their knowledge and experience as they complete the task. Further, the SPSA is an undocumented process that is intended to make the person think about, and act on, the hazards or risks associated with each task. Mr Knight considered the potential downstream consequences of the failure to close valves are outside the scope of the generation and review of non-routine uncontrolled documents. 34 On the other hand, Mr Court considered the SPSA process is inadequate in itself for the purpose of preventing incidents such as occurred, because: (a) it is not documented - an impediment to thoroughness in assessment; and (b) it is unlikely to identify cascading-type problems such as occurred, that is, the transfer of slops from Banksmeadow to tank 11 resulted in the flow of slops to the SWS line via the open valve A which resulted in overloading the SWS plant producing black smoke and overflow of the foul condensate from tank 216 causing odours, (c) there is no evidence that it was conducted before the transfer from tank 218 to tank 12, (d) the SPSA, if performed prior to the transfer from tank 218 to tank 12 is not adequate, even taken with the RFC process, to assess the risk of the incident which occurred since it involved a cascading-type incident having downstream consequences.

[28]

(iii) Whether the RFC procedure assessed the risks involved in the non-routine transfer adequately 35 Mr Knight advised that the RFC process is a formal review and approval process that is intended to focus on identifying and addressing new risks produced from the change and to ensure that all changes are documented. He further stated that the RFC for the transfer of water bottoms from tank 218 to tank 12 used recognised risk assessment techniques and identified the need for a non-routine uncontrolled procedure. The potential downstream consequences of failure to close an existing valve is not a new risk associated with a change and therefore outside the scope of the RFC. 36 Mr Knight considered that there would have been some discussion of the method of transfer during the RFC, based on: (a) the need to consider if the change is technically possible during Pre-Approval, (b) his involvement with other refinery RFCs and (c) the checklists used in the Health and Safety Evaluation 37 Mr Court however considered the RFC process was inadequate because: (a) Some documented process for the assessment of risk for the transfer of foul condensate to tank 12 was warranted, since it involved interconnecting the main SWS line and the slops line, systems which are not normally connected and therefore subject to greater risk by the interconnection than experienced in normal operation. (b) It was general rather than specific in its consideration of how transfers of foul condensate would be made. (c) It did not consider the risks involved in interconnecting the SWS feed line and the slops line, although it did consider the emission of odour from tank 12 in some detail; an effect of the RFC was to highlight in the Non-routine Uncontrolled Document the need to preserve the "sponge layer" in tank 12 to control odours. (d) It specifically did not call for a HAZOP or equivalent analysis of the transfer operation. (e) It was not oriented to investigation of cascade-type incidents such as occurred. (f) There was no requirement to tag and sign valves in the tank 218 to tank 12 transfer, as would apply for a safety procedure. 38 Mr Court could not form an opinion on the adequacy of the process as identified by Mr Knight as there was no evidence of the discussion Mr Knight considered would have occurred. (iv) Risk assessment process generally 39 There was disagreement about whether the existing risk analysis procedures in place were adequate for the refinery in general. Overall, Mr Knight stated that Caltex's risk assessment processes were industry best practice world wide and that there was no legal obligation to undertake any greater level of assessment. Mr Court disagreed that the procedures in place were best practice. He considered that had appropriate systems been in place then the multiple causes giving rise to this offence may have been prevented as multiple defences would have been in place to protect against errors such as the failure to close jumpover valve A, and the failure to install adequate monitors on tanks to warn of overflow. Mr Court stated that a Process Hazard Analysis (PHA) should have been carried out promptly after the issue of the Australian Standard for Risk Management AS/NZS 4369:1999.

[32]

42 On a plea of guilty the Prosecutor has to prove matters adverse to the Defendant beyond reasonable doubt; R v Olbrich (1999) 199 CLR 270. The Prosecutor accepted its onus of proof under Olbrich but argued that it had proved beyond reasonable doubt actual harm to residents in relation to psychological impacts and that toxicological effects were likely to have been caused. Defendant's submissions 43 Dr Howell considered it was unlikely there was sufficient concentration of chemicals in the odour to cause an actual physical reaction, that is, a toxicological effect (see par 30). Applying the onus required of the Prosecutor in Olbrich there was not proof beyond reasonable doubt that there was a toxicological effect caused or likely to be caused to residents. This suggests the extent of the environmental harm is not as serious as the Prosecutor argued. The Defendant admitted that the offensive odour caused suffering to residents, which was not long lasting. No one is reported as requiring hospitalisation or medical treatment as a result of the incident. Finding 44 Dr Kenyon's evidence (par 29) is that the health effects of some complainants were more likely than not to be caused by the action of certain gases on mucous membrane tissues. Dr Howell considered it was more likely that chemical concentrations were less than those associated with direct toxic effects. In the absence of data to analyse concerning the concentration of chemicals it is difficult for either expert to be definitive on this issue. However Dr Kenyon stated that he relied on the residents' reports of health effects to conclude that it was more probable that these were toxicological impacts on human tissue. 45 It is clear the Prosecutor has proved there was a psychological impact on numerous residents which caused immediate and unpleasant physical reactions. The evidence does not demonstrate to the requisite standard of proof that toxicological effects were actually caused nor does the Prosecutor argue that. In the absence of actual data about the composition of the gas causing the odour has the Prosecutor discharged the onus it bears of proving there was a likelihood of actual toxicological harm beyond reasonable doubt? Dr Kenyon considered it was more probable than not that some of the physical symptoms were caused by a toxicological impact based on the reports of the residents' complaints. I do not accept that the Prosecutor has proved beyond reasonable doubt by the evidence of Dr Kenyon that there was a likelihood that toxicological impacts were caused in light of that evidence. The harm caused is reasonably serious but short term with no evidence of long lasting impacts and no evidence that medical assistance was sought by the complainants. (b) the practical measures that may be taken to prevent, control, abate or mitigate that harm Prosecutor's submissions 46 The Prosecutor submitted there were a number of practical measures that could have been taken including physical inspection of jumpover valve A to ensure it was closed after the transfer of foul condensate from tank 218 had taken place. Defendant's submissions 47 The first cause of the incident was the failure by a Caltex technician to close jumpover valve A despite a specific request to do so. 48 The Defendant submitted that as there were 23,000 valves at the Kurnell refinery it was difficult to implement additional measures such as physical inspection. In this case a specific telephone call was made to ask that the jumpover valve A be closed after the transfer procedure but this was not done. It is not practical to expect a physical inspection of every valve.

[33]

49 Measures were undertaken immediately after the incident to reduce the impact of the spill by covering the spill with sand. Finding 50 Clearly there were practical measures which could have been taken to prevent the incident, in relation to the failure to close jumpover valve A and the failure to have an adequate warning system in place in relation to the overfilling of tank 216. 51 I note that in the statement of agreed facts, set out at par 27 and 28, a number of measures have been implemented since the offence in terms of a procedure requiring physical checking of valves and additional monitoring equipment on tanks. The Defendant's counsel suggested these were in part a response to the prosecution rather than practicable measures given the size of the Defendant's operation. 52 The Defendant runs a large complex industrial plant and should have in place appropriate measures to ensure this kind of offence does not take place. Particularly where non-routine procedures are undertaken, the measures now implemented to prevent a recurrence appear reasonable. I do not accept that these are unduly onerous given the nature of the operation, the responsibility of the Defendant to ensure that pollution does not escape from the refinery, and that the Defendant is a large well resourced corporation. (c) the extent to which the person who committed the offence could reasonably have foreseen the harm caused or likely to be caused to the environment by the commission of the offence 53 The Defendant accepted that it could have reasonably foreseen the harm caused as a result of the offence. (d) the extent to which the person who committed the offence had control over the causes that gave rise to the offence 54 The joint report of the experts on risk assessment identified a number of areas of disagreement in relation to risk assessment processes as set out at par 31 - 39. The processes undertaken by the Defendant were a request for change (RFC) process and Safety Performance Self Assessment ("SPSA") process. The only written procedure produced was the Non-Routine Uncontrolled Document "T/F of Water Bottom Tk218 to Tk12". See par 14, 32. Prosecutor's submissions 55 The Prosecutor argued that the Defendant had inadequate risk assessment processes in place which had resulted in a failure to have proper procedures to identify particular risk areas in the conduct of this non-routine procedure. The only written evidence of any risk assessment is the special work procedure. That document contains no risk assessment processes. The direction in the second set of instructions (see par 32) to "return everything to normal" is inadequate as it lacks specificity and does not specify what is "normal". The Prosecutor relied on the evidence of Mr Court, set out above in relation to (ii) and (iii) (par 34, 37, 38), to demonstrate the inadequacies of the risk assessment processes of the Defendant in relation to the particular process undertaken and, at (iv) (par 34), partly generally at the refinery.

[36]

(ii), (iii) - RFC/SPSA processes (see par 33 - 38 above) 58 Mr Knight was not directly involved in the processes the RFC and SPSA about which he was giving evidence, having assumed his position after the offence dates. He stated that the purpose of the RFC and the SPSA is to identify risks for the transfer of the water bottom from tank 218 to tank 12 and a written procedure was required as a result, hence the special work procedure document. I have held this document is inadequate. As identified by Mr Court, there is otherwise no evidence of what risk assessment analysis was undertaken in the RFC and SPSA processes which can be considered. 59 The Prosecutor bears the onus of proving beyond reasonable doubt that the Defendant's risk assessment processes are inadequate. I consider that based on the evidence of Mr Court there are inadequacies in the Defendant's processes in relation to the non-routine procedure which gave rise to this offence, as he outlines in par 34, 37 and 38. In weighing up this evidence with that given by Mr Knight about processes he was not involved in, I am unable to conclude that these processes were adequate for a non-routine procedure undertaken in a large, complex refinery operation given the lack of specific evidence about these. (iv) risk assessment generally (see par 39 above) 60 The Defendant argued in part that the refinery's processes were best practice and that there was no legal obligation to do more. The Defendant's practices accord with best practice worldwide according to the evidence of Mr Knight. Mr Court disagrees with that opinion and gives extensive reasons why, including identifying a specific procedure, PHA (see par 39) which should have been undertaken. Given the disagreement between the experts on the adequacy of the risk assessment processes generally at the refinery and the necessity for the Prosecutor to discharge the onus of proof beyond reasonable doubt, I make no finding here on this issue but make the following observations. I do not consider the argument of the Defendant that it was not legally obliged to undertake any further risk assessment such as a PHA procedure as persuasive. There is a responsibility on operators of large industrial plant to undertake risk assessment processes adequate to their operation. Whether such procedures are specifically required by law is an inadequate measure of their necessity. It is common for there to be industry codes of practice or national/international standards which are adopted widely within an industry without necessarily being required by law. Mr Court referred to the Australian Standard for Risk Management AS/NZS 4369:1999 as one example of this.

At a glance

Court

Decision date

Before

Source

Judgment (56 paragraphs)

Parties

Cases Cited (15)

Cited By (7)