“Fire-safe” does not mean the valve will not burn. The PTFE in a soft seat starts degrading around 250 °C and is consumed within minutes in a hydrocarbon flame above 750 °C. What fire type-testing actually verifies is what happens next: after the soft seal is destroyed, can the valve’s metal backup seat and graphite packing hold leakage to a defined magnitude, so that a burning plant is not continuously fed by this valve? ISO 10497 and API 607 are the two entry points to that verification — one covering metallic valves of all types, the other focused on soft-seated quarter-turn valves. The procedures share the same origin, but the temperature criteria, acceptance details and certificate coverage rules differ, and citing the wrong one in a technical agreement stalls acceptance.
This article separates the systems in order: what fire testing verifies, how ISO 10497 runs, how API 607 runs, how to cite them on one valve, what the test bench must provide, and what to check at acceptance. All figures follow the official texts referenced by the project; the tables are for quick orientation in sourcing and contract discussions.
1. Fire testing verifies the “after burnout” state
Fire testing is a destructive type test: a representative valve is pressurized, burned for 30 minutes in a controlled flame, and leakage is measured throughout. The standard’s premise is that the soft seat will burn away — the purpose is not to prove the valve is fireproof, but to verify that three backup structures take over:
- Metal backup seat: after the soft seat (PTFE/RPTFE/PEEK) is destroyed, the ball or disc lands on the metal secondary sealing surface, and the metal-to-metal contact limits leakage to the allowed magnitude;
- Graphite packing: PTFE packing decomposes above 260 °C; fire-safe designs include graphite rings in the stuffing box that still seal the stem at flame temperatures;
- Body joints and anti-static structure: the body gasket must not become an external leak path in the flame; API 607 additionally mandates a stem-to-body anti-static bond.
A common scoping error: answering a fire-test clause with an ISO 5208 leakage class or an API 598 seat-leak criterion. Ambient pressure tests judge the soft seal in normal service; fire tests judge the backup seals after flame exposure — different magnitudes, different measurement chains, not interchangeable. Also note that fire testing burns one design: the certificate covers a design family, and the size, pressure class and seat-material configuration of the delivered valve must fall inside that coverage to be quotable.
2. ISO 10497: fire type testing for all valve types
ISO 10497:2015 (3rd edition) applies to gate, globe, ball, butterfly and plug valves — metallic industrial valves of all types. It is the international version that merged BS 6755-2 and the early API 607 route. The procedure:
- Ambient-temperature baseline seat-leak test;
- The cavity is filled with water and pressurized (test pressure per the table below), valve horizontal with stem horizontal;
- Flame surrounds the body, calorimeter temperatures roughly in the 750–1000 °C band, burn duration 30 minutes; through-seat and external leakage measured during the burn;
- After flame-out the valve is quenched with water (thermal shock is part of the procedure), and leakage continues to be measured until the body drops below 100 °C;
- After cooldown, one full open-close operation using only the standard handwheel or gearbox (operability check);
- Seat and external leakage re-tested, plus a 1.5× rated-pressure hydrostatic retest of the body.
| Item | ISO 10497 requirement |
|---|---|
| Scope | Industrial metallic valves of all types (multi-turn and quarter-turn) |
| Test pressure | Soft-seated valves rated ≤PN40/Class 300 may use a low pressure of 0.2 MPa; otherwise 75% of the maximum permissible working pressure at 20 °C |
| Burn temperature | Calorimeter temperature roughly 750–1000 °C (flame fully enveloping the body) |
| Burn duration | 30 minutes |
| Leakage during burn | Seat and external leakage metered separately; limits scaled by valve size (e.g. for DN50 under high-pressure burn: seat ≤200 mL/min, external ≤50 mL/min band) |
| After cooldown | Below 100 °C, one full open-close cycle; 5-minute pressure hold, leakage judged against size-scaled limits (typically the ≤200 mL/min band) |
| Body retest | 1.5× rated pressure hydrostatic — no external leakage, no rupture |
| Qualification coverage | Testing one size qualifies smaller sizes of the same design; pressure rating coverage up to about twice the tested rating; ferritic results may be extended to austenitic/duplex with intermediate test points |
Watch the scope boundary: ISO 10497 covers the valve body only — electric, pneumatic or hydraulic actuators are not covered (manually operated gearboxes excepted). For actuated ESD valves, the actuator needs its own fire qualification or a separate clause in the specification.
3. API 607: the American-system path for soft-seated valves
API 607 (7th edition, 2021) covers valves equipped with nonmetallic seats — in practice mainly quarter-turn designs (ball, butterfly, plug) — and is the de-facto fire-certification standard in global oil and petrochemical projects. Key requirements:
| Item | API 607 (7th Ed.) requirement |
|---|---|
| Burn temperature | 760 °C ±38 °C (1400 °F ±68 °F), flame enveloping the body |
| Burn duration | 30 minutes, test pressure held at about 75% of CWP throughout |
| Operation during burn | The valve is cycled once between fully closed and fully open during the burn |
| After flame-out | Immediate water quench |
| Post-fire leakage | High- and low-pressure seat leakage and external leakage measured separately; limits scaled by size (external leakage typically about 750 mL/(h·DN); seat leakage per Table 1 of the standard) |
| Body retest | 1.5 × CWP hydrostatic — no external leakage, no rupture |
| Anti-static test | Stem-to-body and closure-element-to-body resistance measured in fully open and closed positions: <10 Ω at 12 VDC / 250 mA |
The API system has three companions: API 6FA covers API 6A/6D valves (gate, ball, check — pipeline and wellhead), API 6FC addresses fire testing of valves with an automatic backseat, and API 6FD covers check valves. A frequent mismatch is citing API 607 for a multi-turn gate valve — multi-turn valves should go through ISO 10497 or API 6FA unless the specification explicitly accepts otherwise.
4. Citing the two systems on one valve
ISO 10497 and API 607 share the same test logic (30-minute pressurized burn, metal backup seals taking over), and most project specifications accept either — but three differences must be spelled out: the temperature criteria (ISO gives a calorimeter band, API 607 a 760 °C ±38 °C point value), the leakage-limit scaling method, and the qualification coverage rules (size and material extrapolation follow different tables). Cite whichever the specification names; when both are required, accept against the stricter one.
When checking a certificate, confirm four things: the design family (structure, size, pressure class, seat-material configuration) covers the delivered valve; burn temperature and duration records (calorimeter data) are complete; leakage records separate during-burn from post-cooldown, and high from low pressure; and the anti-static value is on the certificate or its annex. Any missing item means the certificate cannot be quoted directly for that valve’s acceptance.
5. What the test bench must provide
A fire-test bench differs from a conventional pressure-test bench in four areas:
- Burn system: fuel supply with metered burners, calorimeters (the basis of the temperature criteria), flame guards and air-cooled shielding — the flame must envelop the body completely and reproducibly;
- Water circuit: cavity filling, pressurization to 75% CWP held through the 30-minute burn, then an immediate switch to the quench circuit after flame-out;
- Leakage metering: during-burn seat and external leakage need independent collection and metering (graduated vessels, weighing or flow meters), and both chains must be calibrated, since leakage is re-measured after cooldown;
- Safety and environment: an open-flame test needs a dedicated test cell, fuel-supply interlocks, flue extraction and fire-fighting provisions — the main cost gap versus a hydrostatic bench.
In addition, API 607 makes the anti-static test routine: a micro-ohmmeter checks stem-to-body continuity at 12 VDC/250 mA, both on the test bench and at final inspection.
6. Six things to check at acceptance
- The design family on the certificate (size, pressure class, seat-material configuration) covers the delivered valve;
- Burn temperature records (calorimeter point data) and the 30-minute duration are complete;
- Leakage records separate during-burn from post-cooldown and high from low pressure, with consistent units and limit scaling;
- The post-cooldown operability check (one full open-close) is documented;
- The anti-static value is below 10 Ω;
- Whether the actuator is inside the certificate scope — if not, the actuator fire requirement in the specification must be closed out separately.
FAQ
Is a “fire-safe” valve immune to fire?
No. Fire-safe is a test conclusion: after the soft seat burns away, the metal backup seat and graphite packing hold leakage within the allowed magnitude. The soft seal will always be destroyed in a flame — that is the standard’s premise, not a quality defect.
Can an API 607 certificate substitute for ISO 10497?
Depends on the specification. The test logic is homologous and most projects accept either, but the temperature criteria, limit scaling and qualification coverage rules differ. If the specification names ISO 10497, an ISO 10497 certificate is required; only a clause like “API 607 or equivalent” allows substitution — and “equivalent” must be confirmed in writing by the purchaser.
Can a conventional pressure-test bench be used for fire testing?
No. Fire testing needs a burn system, quench circuit, during-burn leakage metering and a dedicated test cell, and it is destructive (the sample valve is scrapped after the test) — a completely different facility class from hydrostatic/pneumatic benches. Most manufacturers delegate the type test to an accredited laboratory and keep the anti-static and similar checks as in-house routine.
Are “fire test” and “fire-resistant test” the same thing?
In the valve context, yes. China’s current standard GB/T 26479-2011 “Fire test for resilient-seated quarter-turn valves” is a modified adoption of API 607:2005 and ISO 10497:2010, and its title uses the “fire-resistant” wording; the older JB/T 6899 “Fire-resistant testing of valves” covers the same verification. The content corresponds exactly to what ISO 10497 and API 607 call fire type testing: a 30-minute burn after which the soft seal is destroyed, the metal backup seat and graphite packing take over, and leakage is held within allowed limits. Do not confuse this with “fire resistance” in buildings and materials — fire-resistance ratings for construction elements (GB/T 9978) or fire-resistant cables (GB/T 19216) measure how long a structure or circuit survives a fire, with entirely different subjects, criteria and methods. When a procurement document says “fire-resistant test”, confirm first whether the object is a valve or a construction element, then match the standard.
Specific parameters, criteria and scopes of application follow the official ISO 10497:2015, API 607 (7th Ed.) and API 6FA/6FC/6FD texts referenced by the project; the tables above are for quick orientation in sourcing discussions. How fire testing fits among the other test types and standards is covered in ISO 5208 leakage classes and API 598 vs GB/T 13927; the current editions and scopes of all referenced standards are summarized in the Valve Test Standards Index, and the pressure-rating check logic is in the PN vs CWP checklist.
Related test equipment
Each test above maps to a bench. Pick by valve type and by the item you need to prove:
- Valve pressure test machine — shell at 1.5x and seat at 1.1x, with a report per valve
- Valve comprehensive test bench — pressure, tightness, flow, torque and life cycling on one clamping system
- Valve torque test machine — break-to-open and seating torque curves for actuator sizing
- Valve testing equipment range — browse all models by valve type and test item