Valve Comprehensive Test Bench: Five Tests on One Frame
A valve comprehensive test bench brings shell strength, seat tightness, flow coefficient, operating torque and life cycling onto one frame: one clamping system, one media circuit, one data record — changing products means changing fixtures, not machines. It pays off for manufacturers and test houses that run several valve types, several test items and need a full report per valve. For a single valve type and a single test item, a dedicated machine is cheaper and faster.
Not sure about the test pressure yet? Use the valve test scope configurator — pick valve, size, class and standard, and it returns shell and seat pressures, hold time and a bench list you can email.
By test item: five cluster entry points
These five entry points are organised by the test you need to run, each gathering the standards, criteria, bench configuration and common misjudgements for one line of work:
- Pressure and tightness — shell 1.5x, seat 1.1x, 6 mbar and 0.5 bar tiers
- Flow and control performance — Cv/Kv, PRV curves, TMV response
- Life, torque and fatigue — cycles, torque, impulse waveform, failure criteria
- Special and safety tests — fire, cryogenic, emissions, type test
- Certification and market access — PED/CE, NSF 61+372, TSG D7002
What “comprehensive” actually covers
Comprehensive does not mean stacking five machines together; it means sharing one pressure source, one clamping system and one data output. A bench earns the name only if it can run the following items without breaking pipework or swapping pumps:
| Test item | Pressure / condition | Reference | Bench requirement |
|---|---|---|---|
| Shell strength | 1.5 × CWP (PN16 → 2.4 MPa) | GB/T 13927-2022, ISO 5208, API 598 | No leakage, deformation or cracking at body, body joint and stuffing box |
| Seat tightness (high-pressure liquid) | 1.1 × CWP | GB/T 13927, API 598, EN 12266 P12 | Zero visible leakage for soft-seated; rate-based limits for metal-seated |
| Low-pressure closure (gas) | 0.4–0.7 MPa air | GB/T 13927, API 598 | Bubble immersion or pressure-decay; needs a separate low-pressure circuit |
| Backseat test | 1.1 × CWP, valve fully open | GB/T 13927 | Only for gate, globe and throttle valves with a backseat |
| Flow coefficient Kv / Cv | Flow and differential pressure per opening step | GB/T 30832-2014 | Stable water supply, adequate straight runs, precision DP transmitter |
| Operating torque | Full open/close torque trace | Product standard + customer spec | Verify peak torque against actuator sizing; record dry vs pressurised |
| Life cycling | Cycles per product standard or customer spec | Product standard | Automatic counting, cycling under pressure, auto stop on failure |
| High-sensitivity leak test | Per leak-rate target, e.g. ≤50 ppmv | Customer spec, ISO 15848 (fugitive) | Helium detection needs its own chamber and vacuum circuit |
Hold times are defined, not improvised
GB/T 13927-2022 sets minimum hold times by nominal size. The bench should derive the duration automatically from the entered size — not rely on an operator with a stopwatch. This is one of the first things a second-party auditor checks.
| Nominal size | Shell | Backseat | Seat (other valves) | Seat (check valves) |
|---|---|---|---|---|
| ≤ DN50 | 15 s | 15 s | 15 s | 60 s |
| DN65–DN150 | 60 s | 60 s | 60 s | 60 s |
| DN200–DN300 | 120 s | 60 s | 60 s | 120 s |
| ≥ DN350 | 300 s | 60 s | 120 s | 120 s |
Pressurisation rate matters as much as hold time. Ramping too fast creates a transient differential across the soft seat; the valve then fails the seat test that follows the shell test, and the “leak” is an artefact of the ramp, not the valve.
When a comprehensive bench makes sense
| Scenario | Recommended setup | Why |
|---|---|---|
| One valve type, one test item, dozens per day | Dedicated hydrostatic or air-tightness machine | Most bench functions stay idle; fixture cost never amortises |
| ≥3 valve types, ≥3 test items, one consolidated report | Comprehensive bench | Shared clamping and data output, report generated in one pass |
| Needs Cv/Kv plus torque and life cycling | Bench plus a separate flow rig | Flow needs stable supply and straight runs; conflicts with the pressure circuit |
| Frequent DN≥500 work | Top-press design in a pit | Vertical clamping saves floor space; keep large sizes off the flow rig |
| Type testing only, few samples | Dedicated machine plus outsourcing | Wide item mix but low frequency; utilisation stays poor |
Configuring the bench: five modules, costed separately
1. Pressure sources
Water for shell and high-pressure seat tests; air or nitrogen for low-pressure closure and leak tests. Keep the high- and low-pressure circuits separate with their own gauges: reading a 0.5 MPa test on a 25 MPa gauge gives an error larger than the acceptance limit itself. Add an accumulator to smooth pump cycling.
2. Clamping
Top-press (blind-flange) clamps change over fastest and suit DN15–DN300 batch work; longitudinal flange clamping is stiffer and suits large sizes and high pressure. Do not try to cover both on one frame — the mechanical compromises hurt both.
3. Media circuit
Water needs filtration, drainage and drying. Chloride content must be controlled for stainless bodies; leaving a valve wet overnight can start pitting. A good share of “mysterious seepage” complaints trace back to post-test handling rather than the test itself.
4. Control and pressurisation
PLC with an HMI is the norm: closed-loop pressure control, settable ramp rate, automatic relief on overpressure. Provide dual mechanical locking and a relief valve, and size the pressure cavity with a pressure-vessel mindset.
5. Data and reporting
One pressure–time trace per valve, plus leakage, peak torque, verdict, operator and timestamp. Fields should drop straight into a customer’s quality system instead of saying “passed”.
The bottleneck is usually the fixture, not the pump
The same bench can change over in three minutes or thirty, and the difference is entirely tooling. Provide three to four quick-change fixture sets by size band — ≤DN50, DN65–150, DN200–300, ≥DN350 — with locating pins and centring, plus adapter plates for GB, HG and ASME flanges. Calculate blind-flange thickness per pressure band rather than using one plate for everything.
Traceability: what the report should contain
- Per valve: serial or batch, DN/PN, standard applied, test pressure, actual hold time, leakage, verdict, operator, timestamp
- Retain traces per the customer’s quality system; file factory tests and type tests separately
- Push only key parameters and verdicts to MES; keep raw traces local so shop-floor networks are not choked by large files
- Failures must be replayable: which second pressure dropped and by how much beats “retested OK”
How it maps to the standards
- Chinese: GB/T 13927-2022 (pressure testing), GB/T 30832-2014 (flow coefficient and resistance), GB/T 26480 (inspection and testing)
- International: ISO 5208 (leakage rates), API 598 (factory tests), API 6D (pipeline valves), EN 12266-1 (P10 shell strength / P11 shell tightness / P12 seat tightness)
- Special: GB/T 26479, ISO 10497 and API 607 (fire testing), GB/T 24925, ISO 28921 (cryogenic testing), ISO 15848 (fugitive emissions)
- Regulatory: TSG D7002-2023 type testing for pressure piping components
Pressures and acceptance criteria differ between standards for the same valve, so the recipe library should be indexed by standard, valve type and size — not by pressure alone. See our comparisons of API 598 versus GB/T 13927, EN 12266-1 P10/P11/P12, API 6D testing, ISO 5208 leakage rates and the valve test standards index.
Five common specification mistakes
- Checking only the maximum pressure. A 0–25 MPa gauge cannot resolve 0.5 MPa; low-pressure closure needs its own instrument.
- Sharing one pump between pressure and flow. Pressure testing wants a stiff circuit, flow testing wants a stable one; sharing makes the flow data wander.
- Buying one fixture set at the largest size. Every small valve then costs half an hour of bolt turning.
- No ramp-rate control. Fast ramps damage soft seats, and the following seat tests fail in series.
- Verdicts without data. Auditors want traces and raw values; “passed” does not survive a second-party audit.
Acceptance checklist
- Shell, seat and backseat pressures separately settable, within metrological tolerance of indicated values
- Hold time derived from size automatically and logged, not hand-timed
- Adjustable ramp rate, automatic relief, dual mechanical and electrical interlocks
- Independent low-pressure circuit with a dedicated 0.4–0.7 MPa gauge and air supply
- Fixtures in size bands with measured changeover time
- One complete trace per valve, exportable report fields
- Recipe library maintained by standard, valve type and size
- Drain and dry steps built into the cycle rather than left to the operator
Dedicated machines when you only need one or two items
Not every item belongs on a shared frame. If you only need one or two, a dedicated machine is cheaper:
- Valve flow test system — flow and differential pressure per opening step for Cv/Kv, including balancing and pressure-differential valves
- Valve hydrostatic test machine — shell and seat only, fast cycle, simple structure
- Valve torque test machine — full open/close torque traces for actuator sizing checks
- Valve life cycle test bench — cycling under pressure, automatic counting, stop on failure
- All valve testing machines — the full range by valve type and test item
FAQ
How is this different from a hydrostatic test machine?
A test machine performs shell and seat tests only — simpler, faster. A comprehensive bench adds flow, torque and life cycling on shared clamping and data systems, which raises mechanical and control complexity by an order of magnitude. If you only test pressure, buy the machine.
What size range can one bench cover?
Typical practice is DN15–DN300 on a top-press frame and DN350–DN600 in a pit or with longitudinal clamping; larger sizes need separate evaluation of lifting and foundations. The wider the span, the higher the share of cost in fixtures and civil work — two benches usually beat one that covers everything.
Should flow coefficient be measured on the bench?
Not recommended. Cv/Kv needs a stable supply, long straight runs and low pulsation, which is the opposite of what a pressure circuit is designed for. Most plants keep flow and pressure on separate rigs and consolidate data in one system.
Can the bench run fire or cryogenic tests?
No. Fire testing needs a furnace with a 750 °C class temperature field; cryogenic testing needs a cold box and −196 °C media. Both are dedicated type-test installations. The bench handles routine re-checks before and after, such as seat tightness after a fire test.
Should we integrate with MES from day one?
Get local data retention working first, then talk about MES after three months of stable running. Integrating before the field list is settled usually ends with a slow shop-floor network and data nobody uses.
Why do quotations for the “same” bench differ twofold?
Clamping design, number of pressure stages, number of fixture sets and control level. One fixture set versus four, single versus dual pressure stages, manual records versus automated traces — the cost structures are simply different. Put those four items into the technical specification before comparing prices.
Configuration depends on valve types, size range, test items and line takt. Send the sample list and the standards you test to, and we will respond with a bench scheme including station layout, fixture list and acceptance criteria.