The failure mode on water-supply valves is rarely “it will not hold pressure” — it is quoting the wrong pressure basis. In the EN 1074 family, PN is only a nominal class. What actually sets the test pressure is three letters: PFA, PMA and PEA. The same PN16 gate valve may declare PFA 16 bar, or only 12 bar because of seat-material and temperature derating. A shell test run at 1.5 × PFA versus at PEA is several bar apart. The buyer writes PN in the specification, the maker tests against PFA, and acceptance turns into an argument.
EN 1074 is also not one document but five parts: Part 1 general requirements, Part 2 isolating valves, Part 3 check valves, Part 4 air valves, Part 5 control valves. It governs whether a valve is fit to sit in a supply network — a fitness-for-purpose standard. EN 12266 governs how the pressure test is run and how leakage is judged — a method standard. They divide the work; neither replaces the other. This article runs: the three pressure levels, eleven type tests, the two-tier seat test, endurance at 250/2500 cycles, Kv and disinfection, which part to use, and the Chinese mapping.
1. PFA, PMA, PEA — the three numbers that belong in the contract
EN 1074-1 splits “how much pressure can this valve take” into three service conditions, each with its own purpose:
| Code | Meaning | Condition | Where it is used |
|---|---|---|---|
| PFA | Allowable operating pressure | Continuous, long-term | Selection and normal operation ceiling; the high tier of the seat test is based on it |
| PMA | Allowable maximum pressure | Short-term, occasional (surge, pump start/stop) | Transient check; PMA ≥ PFA |
| PEA | Allowable test pressure | Factory or site pressure test | The highest pressure this valve may ever see in test; site testing must not exceed it |
Three relations are enough: PMA is not below PFA; the shell strength test pressure is the higher of 1.5 × PFA and PEA; the high tier of the seat test is 1.1 × PFA. PEA is commonly set at 1.5 × PFA on network projects, following the EN 805 site-test convention, so the shell test usually lands on the 1.5× rung — but that is not an identity you can invert. PEA is a declared value: whatever the manufacturer declares is what you verify against.
Why PN cannot stand in for PFA: PN is a nominal class at 20 °C, a classification label. PFA is a performance promise for that specific product across its declared temperature range. Within one PN16 family, elastomer-seated and metal-seated versions can carry different PFAs, and DN 200 can differ from DN 50. A specification that names PN16 and stops there leaves the acceptance criterion open.
2. Eleven type tests, grouped the way a bench is laid out
Clause 5 of EN 1074-2 groups performance into mechanical strength, leak-tightness, hydraulic characteristics, resistance to disinfection products and endurance. Regrouped by how they are fixtured, there are four groups and eleven items:
| Group | Clause | Test | How it is done | Typical failure |
|---|---|---|---|---|
| Strength | 5.1.1 | Internal pressure on shell and all pressure-containing parts | max(1.5 × PFA, PEA), hold; no visible leakage, no permanent deformation | Porosity weeping, crack at flange root |
| 5.1.2 | Obturator resistance to differential pressure | Differential across the gate/disc, zero on the other side | Gate distortion, guide jam | |
| 5.1.3 | Resistance to bending | Bending moment applied at both ends, simulating settlement and misalignment | Neck cracking, flange face skew | |
| 5.1.4 | Operating loads (Annex A) | Specified force/torque applied per operating device type | Handwheel deformation, stem bending | |
| Tightness | 5.2.1 | Shell leak-tightness | Same pressure as strength, judged “no visible leakage” | Gland drip |
| 5.2.2 | Seat tightness, two tiers | Low 0.5 bar + high 1.1 × PFA, both directions, rate A | Will not shut off, reverse seepage | |
| 5.2.3 | Maximum operating torque (MOT) | Within MOT the valve must close and reach rate A (Annexes B/C) | Cheater bar on a DN 300 | |
| 5.2.4 | Gearbox tightness to external pressure | External pressure / immersion check on the operator | Buried valve floods and seizes | |
| Function | 5.3 | Kv to EN 1267 | Measured Kv ≥ 0.9 × declared; full-bore gate and clear-way valves exempt | Reduced bore sold as full bore, 20 % Kv short |
| Medium | 5.4 | Resistance to disinfection products | Chlorinated water (order of 50 mg/L free chlorine) then re-check seat tightness and appearance | Rubber swells, seat face goes tacky |
| Endurance | 5.5 | Cycling under pressure (Annex D) | See section 4; strength, tightness and torque re-tested afterwards | Torque doubles after 100 cycles |
A type test is run against a design; once the prototype passes, it covers products of the same design and size band. A routine test is run on every unit and usually covers only shell strength and seat tightness. Treating a type-test report as a per-unit certificate is the most common document error on water projects.
3. Why the seat is tested twice — 0.5 bar is the real network condition
EN 1074-2 requires seat tightness at 0.5 bar low pressure and 1.1 × PFA high pressure, in both directions, to rate A — no visible leakage. Note this is not “pick a class from ISO 5208 A to G”. A supply isolating valve has one criterion: if you can see it, it fails.
The low tier is not a formality. Night-time network pressure is low; a new main is filled from empty; an isolated section is drained and refilled. In those moments the differential across the valve is a few tenths of a bar. A resilient-seated gate with a poor seat contact leaks most visibly exactly there; at high pressure the rubber is pressed harder and hides the defect. A report that only shows 1.1 × PFA is technically incomplete.
Related to this is how leakage is expressed: EN 12266-1 uses mm³/s plus bubbles/min, ISO 5208 uses classes A to G, EN 1074 simply writes “no visible leakage”. These are not interchangeable — “meets class A” and “no visible leakage” belong to different systems, and a cross-reference must name the source document. For the class table see ISO 5208 leakage classes; for the European test designations see EN 12266-1.
4. 250 and 2500 cycles — two lines, not one number
Endurance (5.5, method in Annex D) is banded by operating mode:
| Operation | Cycles | Size range | Notes |
|---|---|---|---|
| Manual (handwheel, T-key, ring key and bar) | 250 | DN 50 – DN 500 | One full open–close is one cycle, under PFA |
| Actuated (gearbox, electric, pneumatic) | 2500 | Same band | Actuated valves move far more often — 10× the count |
250 sounds low. Convert it: an isolating valve on a trunk main is operated fewer than ten times a year, so 250 cycles is roughly 25 years of operation. The standard is counting lifetime operations, not running an accelerated reliability test — a different animal from the tens-of-thousands fatigue runs described in valve life testing.
What decides pass or fail is the re-test afterwards: shell strength, seat tightness (both tiers) and MOT are all repeated, and degradation in any of them fails the valve. Suppliers who state “survived 250 cycles” without post-cycle data have not completed 5.5.
5. Kv and the full-bore exemption
EN 1074-2 requires Kv to EN 1267 with a measured value not below 0.9 × declared — but full-bore gate valves and clear-way valves are exempt, because with no reduction in the waterway Kv follows from geometry and needs no bench verification. That exemption is where cost gets saved: a reduced-bore valve declared full bore, with the Kv bench and report skipped. The detection method is crude and effective — measure the bore. If the minimum waterway of a DN 100 valve is clearly under 100 mm, it is not full bore and Kv must be measured.
Reynolds number, differential stability and pressure-tap locations are specified in EN 1267; get them wrong and the number reads high. See Cv versus Kv boundaries and the valve flow-coefficient test bench.
6. Disinfection resistance — the requirement unique to supply valves
Clause 5.4 requires the wetted interior to keep sealing and functioning after contact with chlorinated disinfecting water. In practice the interior is exposed to chlorinated water on the order of 50 mg/L free chlorine for about 48 h, after which seat tightness is re-checked and the elastomers inspected for swelling, tackiness and cracking. The rubber itself is selected to EN 681-1, and must be a compound that does not impart taste or odour to drinking water in long-term contact.
There is no direct Chinese equivalent: GB/T 17219 evaluates the hygiene indicators of a soak solution — water safety. Clause 5.4 evaluates whether the valve still works after disinfection — function. You need both, and neither substitutes for the other. On export projects they are two separate documents.
7. Choosing the right part
| Part | Subject | Typical size / pressure | Watch for |
|---|---|---|---|
| EN 1074-1 | General: design, materials, test rules, marking | — | PFA/PMA/PEA definitions and checks, coating, hygiene |
| EN 1074-2 | Isolating valves (gate, butterfly, ball…) | DN 50 – DN 2000, PFA 6 – 25 bar | Two-tier seat test, MOT, 250/2500 cycles |
| EN 1074-3 | Check valves | Network sizes | Reverse tightness, cracking differential, surge behaviour |
| EN 1074-4 | Air valves (inlet/exhaust) | Network sizes | Large air intake, small air release, closing tightness |
| EN 1074-5 | Control valves (pressure reducing, sustaining, flow) | Network sizes | Regulation accuracy, steadiness, fail position |
One hard procurement rule: EN 1074-2 is about isolating, not regulating. Using a gate valve as a control valve, or shipping an EN 1074-2 report against an EN 1074-5 order, is a wrong-standard situation. Equally, EN 1074 does not cover potable-water ball valves inside buildings (that is EN 13828) or building gas valves (EN 331).
8. The Chinese mapping
| Dimension | Chinese counterpart | Difference |
|---|---|---|
| Pressure test method | GB/T 13927-2022 (MOD ISO 5208) | Test procedure and classes only; no PFA/PMA/PEA structure |
| Product | GB/T 12232 and other product standards | Materials and pressure classes by product family; no fitness-for-purpose type-test layer |
| Hygiene | GB/T 17219 | Soak-solution hygiene indicators, not disinfection resistance |
| Project acceptance | GB 50268 | Site pressure test, flushing and disinfection — a construction step, not a type test |
| Endurance and MOT | Product standard + customer spec | No unified 250/2500 count or MOT criterion |
For valves exported to Europe, the correct posture is a two-layer report: strength and tightness recorded in-house to GB/T 13927, conformity issued third-party to EN 1074-1/-2, with an explicit clause-mapping table. Hand over a GB-only report and the first question will be “what is the declared PFA, and did you run the 0.5 bar tier?”.
9. Configuring the test equipment
- Large-size pressure testing: DN 200 – DN 2000 needs blind flanges with hydraulic clamping, shell pressure set to max(1.5 × PFA, PEA). PFA16 means 2.4 MPa, PFA25 means 3.75 MPa — leave headroom in both range and fixture stiffness
- Low-tier resolution: judging “no visible leakage” at 0.5 bar needs a pressure sensor at 0.1 % FS or better, otherwise the small decay is invisible to the instrument
- Torque and over-torque: must close and hold within MOT, then apply operating loads per Annex A; gearbox output torque above DN 300 needs its own calibration
- Endurance bench: cycling under pressure, automatic count to 250/2500, stop on failure, and a route straight into the re-test
- Bending rig: moment loading above DN 200, clamps clear of the flange sealing face
- Kv bench: large-flow loop above DN 50, taps located per EN 1267
For sharing one frame across several test types see the valve comprehensive test bench; individual machines include the valve pressure test machine, valve life test bench, valve torque test machine and the air valve test machine (for EN 1074-4 large-intake and small-release behaviour).
FAQ
The specification only says PN16 — what PFA do I test to?
Test to the manufacturer’s declared PFA, and put PFA, PMA and PEA in the bid. PN16 declares a class, not a test pressure. If the buyer insists PN16 means 16 bar continuous, check whether the seat material and temperature derating support it — many resilient-seated valves carry PFA of only 12–14 bar in continuous service.
Can an EN 1074 report and an EN 12266 report substitute for each other?
No. EN 12266-1 covers pressure testing only (P10 shell strength, P11 shell tightness, P12 seat tightness). EN 1074 also requires bending, operating loads, MOT, Kv, disinfection resistance and cycling. Conversely, the EN 1074 shell pressure is derived from PFA/PEA and does not fully match the EN 12266 calc from the 20 °C rating; each report must name its own basis.
Can the 0.5 bar tier be skipped?
No. It is a required tier of EN 1074-2 seat tightness, and it is the condition that most reliably exposes poor resilient-seat contact. A report showing only 1.1 × PFA will be marked “5.2.2 not completed” in technical review.
250 cycles sounds far below the tens of thousands industrial valves require — is it lenient?
Different purpose. The 250 cycles represent a supply isolating valve’s lifetime operation count (about 25 years at under ten operations a year), testing whether it still moves and still seals after long idle periods. The tens of thousands on an industrial valve are a high-frequency fatigue duty. The numbers are not comparable across the two intents.
Specific clauses follow EN 1074-1:2000 and EN 1074-2:2000 (including A1:2004); the tables above are for orientation in selection, quoting and test discussions. For the US counterpart on the same valve see AWWA C515 and C509; for the hygiene gate into US municipal networks see NSF/ANSI 61 and 372; and the full map is in the valve test standards index.
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 flow test system — flow and differential pressure per opening to derive Cv/Kv
- Valve comprehensive test bench — pressure, tightness, flow, torque and life cycling on one clamping system
- Valve testing equipment range — browse all models by valve type and test item