StocExpo 2012 Ross Coulman 15th March 2012 History has told us multiple events occur Multiple Tanks and Bund / Dykes = Natural escalation Worse case is the Bund/Dyke Not your largest tank not enough foam not enough foam 430k for 65 minute application not enough foam 559k for 85 minute application not enough capacity INSERT Dyke Volume Calculation BUND Holds XXXXX litres not enough capacity INSERT Dyke Volume Calculation Fire water alone is 11,272,300 litres not enough capacity INSERT Dyke Volume Calculation Fire water plus rainwater@ xx% = XX,XXX,XXX litres not enough capacity INSERT Dyke Volume Calculation Fire water plus rainwater@ xx% = XX,XXX,XXX litres PLUS Product 100% largest tank, 10% of all others = XX,XXX,XXX litres IP 19 HSG 176 CIRA 598 EN 1992 BS EN 14015 vertical, cylindrical, non-refrigerated, above-ground storage tanks typically designed to standards BS 2654,3 BS EN 14015,4 API 620,5 API 6506 Single case pre-plan methodology Boilover potential provides escalation – potential to create lots of multiple events Single case pre-plan methodology Pessimism is good – saying singles case is acceptable due to frequency level is naïve Single case pre-plan methodology Have you taken the shared learning and applied it post buncefield? Single case pre-plan methodology Negatives History has told us multiple events occur (Buncefield) – Multiple Tanks natural escalation Worse case is the Dyke not your largest tank – shortcomings; - not enough foam - not enough water - not enough manpower - not enough ….. Boilover potential provides escalation – potential to create lots of multiple events Single case pre-plan methodology The factors to be taken into account when sizing bunds to deal with multi-tank failure scenarios and fire water management, plus the capacity for remote [tertiary] containment systems, are: Primary capacity - 100% of primary capacity. Consider the possible failure modes and where appropriate, include the capacity of all primary tanks in multi-tank installations, incidence of multiple tanks in one bund and where tanks are hydraulically linked in which case they should be treated as if they were a single tank. Rainfall - subject to operational procedures, in order to calculate the volume to be contained, allow for a 10 year return, 8 days rainfall prior to the incident, and a 10 year return, 24 hour rainfall, plus an allowance for rain falling directly on to remote containment and areas of the site draining into it, immediately after the incident. The post-incident component and the allowance for dynamic effects (see e) are not additive. Fire fighting and cooling water - Allow for the volume of extinguishing and cooling water delivered through fixed and non-fixed installations based on BS5306, VCI, CEA, ICI and Institute of Petroleum methodologies, with appropriate adjustments in the light of the particular circumstances. Consultation with the regulators and the fire service is essential. Foam - Allow a freeboard of not less than 100 mm. Dynamic effects - this is to allow for the initial surge of liquid and for wind-blown waves. In the absence of detailed analysis, allow 250mm (750 mm for earth walled bunds). Health and Safety Executive (UK) recommendations on multi tank and bund/dyke modelling post Buncefield disaster Tailored to risk, not Prescriptive Swiss Cheese Model People – Process – Plant People are the weakest link Identify Vulnerability Compare and contrast single source and multi-source incident scenarios No excuses – it’s easy to model multi-source Q&A
© Copyright 2026 Paperzz