A pressure relief valve (PRV/PSV) is a spring-loaded, self-actuated device that opens when pressure reaches its set pressure (≤ the vessel’s MAWP) and relieves enough flow to stop the pressure climbing past the code accumulation limit (typically 110% MAWP, 121% in a fire case).
It must be sized for the worst credible relief load (blocked outlet, fire, control failure, thermal expansion…). Undersized, the pressure blows through the limit even with the valve wide open — the model shows exactly that.
It is a safety device, not a control device: never isolate it, never let it be the normal pressure control, test it on a schedule, and use a rupture disc where speed or zero leakage is needed.
1 · Why a purely mechanical last resort
Process pressure is normally held by controls, interlocks and operators. But all of those can fail — a control valve sticks shut, a downstream block valve is left closed, a fire heats a vessel, a tube ruptures in an exchanger. The relief valve is the independent, last-resort barrier that doesn’t depend on any of them: it senses the process pressure directly, and the spring force is the only thing holding it shut. When pressure wins, it opens. That mechanical simplicity is the whole point — it works when the power is off and the logic is down.
It exists because the alternative is catastrophic. A vessel taken past its strength doesn’t leak politely; it ruptures, releasing its contents and stored energy at once. Codes (ASME VIII, the PED) therefore require overpressure protection on pressure equipment, and the relief device is how that requirement is met.
2 · The language: set, accumulation, overpressure, blowdown
Relief is described with a precise vocabulary, all referenced to the vessel’s MAWP (maximum allowable working pressure):
| Term | What it means |
|---|---|
| MAWP | Maximum allowable working pressure — the vessel’s rated ceiling. The reference for everything else. |
| Set pressure | The inlet pressure at which the valve starts to open. Set at or below MAWP. |
| Overpressure | The pressure rise above set needed for the valve to reach full lift & rated capacity (typically 10%). |
| Accumulation | The peak pressure reached during relief, as % above MAWP. Code limit: 110% (single device, non-fire), 116% (multiple), 121% (fire case). |
| Blowdown | How far below set the pressure must fall before the valve re-seats (typically ~7%). Stops chatter. |
The job of sizing is to guarantee that, during the worst relief scenario, the valve passes enough flow that the accumulation stays under the code limit. If it can’t, the protection has failed even though the valve opened.
3 · Sizing — matching capacity to the worst case
Sizing is a two-step discipline set out in API 520/521:
- Find the governing relief load. Work through the credible overpressure scenarios — blocked outlet, control-valve failure, external fire, thermal expansion of trapped liquid, exchanger tube rupture, loss of cooling — and take the largest required relief rate. Fire often governs for vessels; thermal expansion needs only a tiny valve.
- Size the orifice for that load at the allowable accumulation, using the API gas or liquid equations (which give a required effective orifice area, then round up to a standard API letter orifice D…T).
The consequence is simple and unforgiving: a valve sized for a smaller scenario than the one that actually occurs will be wide open and still losing the race. Run the event with an under- and over-sized valve:
Interactive — Relief event
Live modelVessel pressure during the upset
4 · Valves, discs and the rules
- Relief vs safety valve. A relief valve opens proportionally (liquid service); a safety valve pops fully open (compressible gas/steam); a safety relief valve does either. Loosely, all are called PSVs.
- How they lift — progressive vs snap. This is the practical difference between the two technologies. A conventional spring-loaded valve lifts progressively: it cracks at set and opens further as the pressure climbs, reaching its rated capacity only at the rated overpressure (≈10% for gas/steam, up to ~25% for liquid). The valve therefore needs that overpressure margin to pass full flow — which the sizing and accumulation budget must allow for. A snap-acting pilot-operated valve behaves quite differently: the pilot senses set pressure and loads/unloads the main valve’s dome so the main disc snaps essentially full open at set, reaching rated capacity with almost no overpressure. (The opposite pilot type, a modulating pilot, deliberately opens only as far as the flow demands, to minimise the relieved volume — so “pilot” is not synonymous with “snap.”)
- Conventional vs balanced vs pilot-operated. Conventional spring valves are affected by back-pressure and need overpressure to reach full lift; balanced-bellows valves resist back-pressure; pilot-operated valves use the process pressure itself to seal tighter as set is approached — so they leak less, can run much closer to set (≈95–98%), and, in the snap configuration, reach full capacity at minimal overpressure. That low overpressure requirement is a real advantage where the gap between operating pressure and MAWP is tight.
Rupture discs — the non-reclosing alternative
The other primary overpressure device is the rupture disc (bursting disc): a thin membrane that bursts at a set differential pressure, opening the full bore in milliseconds. It has no moving parts, is bubble-tight right up to burst, is cheap, and reacts faster than any spring valve — but it is non-reclosing: once it goes it stays open and the equipment vents until it is shut down and a new disc fitted.
That trade-off decides where each is used. A PSV recloses and is the default for repeatable relief; a rupture disc wins where speed matters (a fast deflagration or runaway reaction a spring valve can’t catch), where zero leakage is essential (toxic, costly or fugitive-emission fluids), or as a cheap primary on simple equipment. The two are most powerful combined: a disc installed upstream of a PSV isolates the valve from corrosive, fouling or polymerising fluid and gives the leak-tightness the valve lacks, while the PSV still recloses after the event. Codes then require the space between them to be monitored (a pressure gauge or excess-flow telltale), because a pin-holed or already-burst disc would quietly shift the combined set pressure.
- Operate well below burst. Discs fatigue, so you run below the rating — roughly ≤70% of burst for older forward-acting (tension) discs, up to ~90% for modern reverse-buckling (compression) discs, which resist fatigue and don’t fragment, so they can sit under a PSV.
- Its failure is hidden too. You can’t non-destructively test the actual burst, and a disc fitted backwards, corroded or fatigued will let go at the wrong pressure with no warning — so integrity rests on correct installation, the interspace monitor, and time-based replacement: the same hidden-failure logic as the PSV itself.
And the operating rules that keep relief devices effective are non-negotiable: never isolate a PSV from the equipment it protects (or use a car-sealed/locked-open, interlocked arrangement); never use it for routine pressure control (that’s the control valve’s job — repeated lifting damages the seat and causes leakage); and test and recertify on a schedule — because a PSV’s failure is hidden (see below), a valve that has sat untested for years may be silently seized, corroded or set wrong, and nobody would know until the day it was demanded.
A PSV is a classic hidden-failure device. Like other protective devices — process trips, the SIS, standby pumps, firewater — a relief valve has a hidden function: in normal running you cannot tell whether it still works, because it does nothing until it is demanded. A seized, fouled or mis-set valve therefore sits silently failed until either an overpressure event arrives (too late) or someone tests it. That is why its scheduled pop-test is not ordinary PM but a failure-finding task in the RCM sense — and its interval is set by how rarely you can tolerate the valve being failed-and-undetected when an upset coincides, the very same logic as the SIS proof-test interval that keeps a SIL valid. The population is prioritised by criticality, and the valve is almost always a credited protection layer in LOPA.
Key takeaways
- The PSV is the independent, mechanical last resort — it opens on pressure alone when controls and people have failed.
- Set ≤ MAWP; keep accumulation under the code limit (110% non-fire, 121% fire) — that is what sizing must guarantee.
- Size for the worst credible relief load (API 520/521); an undersized valve loses the race wide open.
- It protects, it doesn’t control — never isolate it, never use it for routine control.
- Its failure is hidden — a PSV does nothing until demanded, so the scheduled pop-test is a failure-finding task (RCM), with the interval set by tolerable failed-and-undetected risk.