Of all the overpressure scenarios a vessel can face, an external pool fire is the one that most often ends up governing the size of its relief valve. The reason is physical: a fire dumps heat into everything at once, and on a vessel holding liquid that heat boils the liquid into a large volume of vapor that has to escape. This guide explains how the fire case relief load is calculated, why the wetted area and heat input drive the answer, and why liquid-full vessels so frequently need a larger valve than any process upset would demand.
Fire Case Relief Load in one line: The fire case relief load is the vapor generation rate a relief valve must handle when an external fire heats a vessel and boils off its liquid contents. It is found by estimating the heat a pool fire puts into the vessel's wetted surface using the API 521 heat input equations, then dividing that heat by the latent heat of vaporization of the fluid to get a vapor rate. Because a fire adds heat across the whole wetted area rather than through a single process path, this scenario often produces the largest required relief rate on a liquid-holding vessel.
The fire case starts from a simple chain of physics: a pool of burning hydrocarbon surrounds a vessel, radiant and convective heat crosses the vessel wall, and that heat boils the liquid inside into vapor. The relief valve then has to pass that vapor fast enough that pressure never climbs past the allowable accumulation. To turn this into a number, the engineer estimates the heat input rate, converts it to a mass of liquid boiled per unit time, and treats that vapor rate as the required relief load for the fire contingency.
The heat input depends on the area of the vessel that is in contact with liquid - the wetted area - because only the wetted surface transfers heat efficiently into boiling liquid. A dry vapor space above the liquid heats too, but it does not generate the same rush of vapor, so the calculation focuses on the wetted portion, typically up to a defined elevation above grade where a ground-level pool fire can plausibly reach. API 521 provides heat input equations that relate this wetted area to a heat rate, with different forms depending on whether the vessel has adequate firefighting and drainage, which reduce the credited heat, or lacks them, which increases it.
Once the heat rate is known, dividing it by the latent heat of vaporization of the fluid gives the mass rate of vapor produced. A fluid with a low latent heat boils off faster for the same heat, producing a larger vapor rate and therefore a larger required valve. That vapor rate, at the relieving temperature and pressure, is what the relief device must be sized to pass - and it is this figure that so often turns out to be the governing case.
Process overpressure scenarios usually push fluid through one specific route: a control valve fails open, a pump keeps running against a closed outlet, or tubes rupture in an exchanger. The relief rate from such a scenario is bounded by the capacity of that single path. A fire is different in kind - it does not flow through a pipe, it wraps the entire vessel, so the heat input scales with the whole wetted surface rather than the throughput of one line. On a large vessel with a lot of liquid inventory and a lot of wetted area, that distributed heating can generate far more vapor than any single process path could ever deliver.
This is why the fire case so frequently governs on vessels that run liquid-full or hold a large liquid volume: separators, surge drums, storage bullets, and treater vessels. The more liquid surface exposed to a fire, the more vapor generated, and the bigger the valve required. On a mostly vapor-filled vessel the picture can flip, because there is little liquid to boil and the fire produces mainly thermal expansion of gas rather than a torrent of new vapor, so a process case may govern instead. Knowing which regime a vessel is in tells an engineer whether to expect the fire case to dominate.
The consequence is practical. When a relief valve on a liquid service looks oversized relative to normal operation, the fire case is usually the reason, and shrinking the valve to match a process case would leave the vessel unprotected in the one credible event that matters most. This is also why measures that reduce credited fire heat - fireproof insulation, adequate drainage that keeps a pool from pooling under the vessel, and firewater - can meaningfully cut the required relief load, because they act directly on the heat input term that sizes the valve.
A relief valve sized for the fire case is a passive last line of defense; it does nothing until pressure is already high. Everything that reduces the chance of a fire reaching that point, and everything that gives operators warning, sits in the monitoring and control layer around it. On an unmanned oil and gas site, a cloud SCADA platform is often the only continuous eye on the equipment, watching separator and vessel pressures, flagging a rising trend long before a relief device would lift, and correlating it with fire-and-gas detector states so an operator can act on a developing hazard rather than discovering it after the fact.
A platform such as Merobix reads vessel pressure, level, and temperature from field devices over Modbus and DNP3 and can log the exact conditions during and after any abnormal event. If a relief device does lift, the history shows how fast pressure rose and how long the vessel relieved, which supports the engineering review that follows and helps confirm the vessel behaved within its fire-case design. That record also feeds the wider safety picture: a fire that boils a vessel's contents is exactly the kind of event where a timestamped trend of pressure, level, and detector alarms is worth far more than a single after-the-fact snapshot.
None of this replaces the relief valve or the fire-case calculation behind it. The load is set by physics and by API 521, and the valve must be able to pass it regardless of any instrumentation. What the monitoring layer adds is the difference between a site where an operator sees the wetted-vessel scenario developing and one where the first sign of trouble is a relief valve already blowing to flare - a distinction that matters most on remote sites where no one is standing at the fence line.
Because a fire heats the entire wetted surface of a vessel at once and boils its liquid into vapor, rather than pushing flow through a single process path. On a vessel with a large liquid inventory and a lot of wetted area, that distributed heating can generate more vapor than any process upset, so the fire case produces the largest required relief rate and sets the valve size.
Wetted area is the portion of a vessel's surface in contact with liquid inside, usually counted up to a defined elevation above grade that a ground-level pool fire could reach. It matters because only the wetted surface efficiently transfers fire heat into boiling liquid, so the heat input and the resulting vapor rate scale with it. The dry vapor space above the liquid is treated differently because it does not generate the same rush of vapor.
Yes. Fireproof insulation, adequate drainage that keeps a burning pool from settling under the vessel, and firewater all reduce the heat that API 521 credits as reaching the vessel. Because the required relief rate is proportional to that heat input, cutting the credited heat directly lowers the vapor generated and can reduce the size of relief device the vessel needs.
This page references the protocol specifications published by the organizations below. Editions, product capabilities, and documentation change over time - confirm current requirements and specifications directly with the source.
Last reviewed: July 27, 2026. Merobix is not affiliated with, endorsed by, or sponsored by these organizations; their names are used only to identify the standards and products discussed.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.