The basis for determination of the relief load for a particular system is the worst single case contingency. A system for this purpose is all of the equipment/piping a given safety relief valve service is designed to protect. The worst single case contingency corresponds to the largest expected relief load because of a single event. The event may be a fire or failure/mal-operation of a single component of the system (e.g., fire around a single vessel, inadvertent operation of a single valve, etc.). If a single contingency could trigger further contingencies (e.g., fire around a single vessel spreading to other vessels within the same fire zone, failure, of a control valve causing inadvertent response of other control valves, plant-wide power, or instrument air failure, etc.), then the total effect of the contingency should be taken in to account in arriving at the system relief load. For non-fire cases, credit may be taken for normally open flow paths out of the system. For fire cases, it should be assumed that all the feed to and product from the system has ceased to flow.
Safety relief valve sizing should be generally based on API RP 520 PT I and API RP 521. However, specific requirements for basic categories of safety relief valves (e.g., full-flow, fire, and thermal) are typically given per project specific. Sizing requirements for relief of backflow due to check valve failure are also given. In addition, examples of typical systems that often require each type of protection are included. This can be taken only as a starting point since each relief situation needs to be analyzed carefully to make sure all potential relief scenarios have been screened.