A pressure relief valve does not fail the way most mechanical equipment fails. It does not make noise, produce vibration, or show a trend on a condition monitoring chart. It sits on its seat, holding pressure, giving no indication of its condition — until the moment it is called upon to open. At that moment, a valve that has drifted from its set point, accumulated seat deposits, or corroded internally will either fail to open at the correct pressure, open and fail to reseat, or not open at all. Any of those outcomes, in any credible overpressure scenario, is a loss of containment event.
This is why overpressure protection is not a maintenance problem. It is a process safety problem — one that can only be managed through a disciplined inspection and testing program applied to every pressure relief device in the register. What follows is a forensic account of how the failure chain develops, and what an execution-grade program requires to interrupt it.
The Protection Hierarchy
Before examining how PRVs fail, it is necessary to establish what they are and what they are not. A pressure relief valve is the last line of passive overpressure protection in a layered defence system. It is not the primary safeguard. The protection hierarchy runs as follows, from first line to last:
Overpressure Protection — Defence Layers in Order
- Process design and operating limits — Equipment designed to a maximum allowable working pressure (MAWP) that provides margin above normal operating conditions. The first layer is design adequacy.
- Operating procedures and controls — Operator actions, control systems, and alarms that prevent exceedance of operating limits under normal and abnormal conditions.
- Safety instrumented systems (SIS) — Automated protective functions that intervene when operating limits are exceeded and control systems alone are insufficient.
- Pressure relief devices (PRDs) — The final passive layer. PRVs, rupture discs, and pilot-operated relief valves that open to discharge pressure when all upstream layers have failed or been bypassed. A PRD that does not function is not a degraded safeguard. It is an absent one.
Every overpressure event that reaches a PRV has already passed through three failed or exhausted layers of protection. The PRV is not a backup. It is the last barrier between a controlled pressure exceedance and a catastrophic loss of containment. A maintenance program that treats PRVs as routine rotating equipment — scheduled on calendar intervals without a risk basis, tested on a bench and reinstalled without a documented programme — has misunderstood what it is managing.
The Failure Chain
PRV failures are not sudden. They develop through a predictable sequence that begins long before the valve is called upon to operate. Understanding the chain is the foundation of any credible program — because every link in the chain represents an intervention point where the program either acts or fails to act.
Set Point Drift
Spring relaxation, corrosion of internal components, and product buildup on the disc or seat cause the actual opening pressure to deviate from the stamped set pressure. A valve set at 150 PSI that opens at 180 PSI is not providing protection at 150 PSI. It is providing unverified protection at an unknown pressure — which is no protection at all. Set point drift is silent, progressive, and detectable only through testing.
Seat Leakage and Simmering
Contaminants in the process stream — solids, polymerised product, corrosion products — accumulate on the valve seat and prevent full closure after the valve has lifted. A leaking PRV wastes product and creates a process upset, but the more dangerous consequence is that the operator or maintenance team tightens the valve to stop the leak, effectively increasing the set point beyond the stamped value and beyond the MAWP. This action — taken routinely in facilities without a PRV programme — eliminates the protection layer entirely.
Valve Isolation Under Operating Conditions
A PRV that is leaking or simmering creates operational pressure to isolate it. A vessel operating with its relief valve isolated — even temporarily — has no overpressure protection. ASME Section VIII and API RP 576 both address the use of block valves under PRVs and the strict administrative controls required when such arrangements are used. In most facilities where PRV isolation occurs, those controls do not exist. The valve is closed, the tag-out is verbal, and the isolation duration extends until the PRV is eventually returned to service — or until an overpressure event occurs.
The most dangerous PRV in any facility is not the one that fails to open. It is the one that was isolated three months ago for a leaking seat and has not been returned to service — because no one built a system to ensure it would be.
Darryl Mohammed — Principal, Rock Industrial Solutions LimitedRupture Disc Failure or Bypass
Where rupture discs are installed in series with PRVs — either as upstream isolation or downstream back-pressure protection — disc fatigue, corrosion, or incorrect installation can result in premature rupture or failure to rupture at the design pressure. A ruptured disc upstream of a PRV leaves the vessel protected only by the PRV, which may have been selected and sized assuming the disc was intact. A disc that has failed to rupture at the design pressure means neither layer of protection activated.
Inadequate Discharge Path
A PRV that opens correctly at the correct pressure and fails to protect the vessel because its discharge path is blocked, undersized for the credible relieving scenario, or discharges into a system that imposes back pressure exceeding the valve's allowable limit. The valve functioned. The system failed. Discharge path design and maintenance are part of the overpressure protection program — not a separate piping problem.
Loss of Containment
Vessel or system pressure exceeds MAWP with no functioning relief path. Catastrophic failure, personnel exposure, fire, explosion, or toxic release. Every link in the chain above was an intervention point. A program that does not address each link has accepted the terminal outcome as a possibility.
What the Standards Require
Three primary standards govern PRV inspection and maintenance in pressure equipment service. They must be applied together.
Governing Standards — Pressure Relief Device Inspection and Maintenance
- ASME Section VIII, Division 1 (UG-125 through UG-136) — Establishes the overpressure protection requirements for pressure vessels: required relief capacity, set pressure limits relative to MAWP, accumulation limits, and the requirement that every vessel in pressure service be protected by a PRD sized for all credible relieving scenarios.
- API RP 576 — Inspection of Pressure-Relieving Devices — The primary practice standard for PRV inspection, testing, and maintenance. Defines inspection intervals, testing requirements, records management, and the criteria for acceptance or replacement. The risk-based inspection methodology in API RP 576 allows interval extension — but only with a documented risk basis, not administrative convenience.
- API 510 — Pressure Vessel Inspection Code — Governs in-service inspection of pressure vessels including PRV suitability verification as part of the vessel inspection program. A vessel inspection that does not address the condition and test status of its PRVs is an incomplete inspection.
- TSSA O. Reg. 220/01 — Ontario regulatory requirements for pressure equipment in service, including operating engineer obligations and inspection record requirements for all pressure-retaining components.
- NBIC NB-23 (National Board Inspection Code) — Governs the repair and replacement of PRVs. NBIC-certified repair organisations only. Field modification of PRV spring settings by non-certified personnel is a code violation — not a maintenance activity.
What an Execution-Grade PRV Program Requires
A PRV maintenance program is not a testing schedule. It is a closed-loop system that connects testing results to set-point verification, to interval determination, to isolation management, to discharge path integrity — on every valve in the register, with no exceptions. The following elements are mandatory.
- PRV Register — Complete and Current Every pressure relief device in service registered individually: tag number, vessel it protects, set pressure, MAWP of protected equipment, design standard, date of last test, test result, next due date, and current status (in-service, isolated, spare). A device not in the register is not in the program. It is unmanaged.
- Risk-Based Inspection Intervals Test intervals assigned on a documented risk basis — service severity, process fluid characteristics, corrosion potential, operational history — not defaulted to the maximum permitted interval for administrative convenience. API RP 576 Table 1 provides the framework. The interval must be justified, not assumed.
- As-Found Testing Before Any Adjustment Every PRV tested in the as-found condition before any cleaning, adjustment, or repair. As-found test results are the only data that tell you what the valve was actually doing while it was in service. A valve cleaned and adjusted before testing has destroyed the evidence of set-point drift, seat leakage, and internal condition. As-found data is non-negotiable.
- Isolation Management System Any PRV isolation — planned or emergency — must be governed by a formal isolation management system: documented authorisation, defined maximum isolation duration, vessel operating restriction during isolation (if any), and mandatory reinstatement procedure. Verbal isolation with no reinstatement tracking is not a system. It is a latent risk accumulator.
- Discharge Path Verification PRV discharge system inspected and verified for each credible relieving scenario: no blockages, correct sizing for peak relieving flow, back pressure within valve design limits, discharge point confirmed. This verification is part of the PRV program — not deferred to the piping inspection program.
- NBIC-Certified Repair and Adjustment Only All PRV set-point adjustments, internal repairs, and spring replacements performed by an NBIC-certified organisation with documented repair records. Field adjustment of spring tension by maintenance personnel — however routine it may appear — is a code violation and a liability event. The stamp on the valve nameplate is only valid after a certified test and seal.
Field Observation — The Tightened Valve
In field reviews across multiple industries, the most consistently encountered PRV deficiency is not the valve that has never been tested. It is the valve that was leaking — for reasons that would have been identified and corrected under a proper programme — and was tightened by a maintenance technician to stop the simmer. That valve is now set above its stamped set pressure. It may be set above the MAWP of the vessel it protects. No one recorded the adjustment. The work order says "PRV seat leak — corrected." The vessel has no overpressure protection.
The Consequence of an Absent Program
The consequence of an absent PRV program is not a failed valve. It is a failed assumption. Every operator who has not tested their PRVs is operating on the assumption that the valves will perform at their stamped set pressure when called upon. That assumption has never been validated. It may be correct. It may not be. The program exists to replace the assumption with data.
Texas City. Bhopal. Piper Alpha. Every high-consequence pressure equipment incident in the public record contains, somewhere in the failure sequence, a protection layer that was assumed to be functional and was not. The PRV program is not the layer that prevents the incident. It is the layer that ensures the last line of passive protection is not the one that fails.
Commission a PRV Program Gap Audit
If your pressure relief devices do not have as-found test records, documented set-point verification, and a formal isolation management system, your overpressure protection program has gaps. RISL identifies them before an incident does.
Request an Engagement