- API 598 defines the shell test, which verifies the valve body’s structural integrity under pressure, and the seat test, which verifies shutoff performance, as two distinct procedures with different acceptance criteria.
- Shell test pressure is set well above rated working pressure, typically 1.5 times the Class 150 rating equivalent at ambient temperature, to demonstrate margin against body failure rather than to simulate normal operating conditions.
- Seat leakage acceptance criteria vary by valve type and specified shutoff class, and a valve passing its shell test with no visible leakage can still fail its seat test if internal leakage exceeds the permitted rate for its class.
Two Tests, Two Different Questions
API 598 is the industry-standard reference for valve inspection and testing, and its core value comes from separating two conceptually distinct questions that a valve needs to answer before it’s accepted into service. The shell test asks whether the pressure-containing boundary, the valve body and bonnet, can withstand pressure without leaking to atmosphere or failing structurally, essentially confirming the pressure vessel aspect of the valve is sound. The seat test asks a completely different question: with the valve in its closed position, does it actually stop flow through the valve to the degree its specified shutoff class requires? A valve can pass one test and fail the other, and understanding why requires recognizing that these tests examine different failure modes: a shell leak indicates a problem with the body casting, wall thickness, or a body joint, while a seat leak indicates a problem with the sealing surfaces, disc or ball fit, or seat material, entirely separate concerns from whether the body itself holds pressure.
Shell Test Procedure and Purpose
The shell test pressurizes the valve body, typically with the valve held partially open to expose the internal cavity fully to test pressure, at a level set by the standard well above the valve’s rated working pressure, commonly around 1.5 times the Class 150 equivalent rating at ambient temperature regardless of the valve’s actual pressure class, though the specific multiplier and hold time are defined precisely in the standard’s tables. This elevated test pressure exists to demonstrate margin against body failure, not to simulate any actual operating condition the valve will see in service, and the test typically requires a minimum hold time, commonly a specified number of seconds per nominal pipe size, with zero visible leakage through the body or bonnet joint being the acceptance criterion for most valve types. A shell test failure, showing visible leakage through the pressure boundary itself, is a serious defect indicating the valve body should not be accepted, since it points to a fundamental problem with the casting, forging, or machining that no amount of trim adjustment can fix.
Seat Test Procedure and Acceptance Criteria
The seat test, performed with the valve closed, measures leakage past the seat when pressure is applied from one or both directions depending on whether the valve is bidirectional, using either a liquid or gas medium as specified. Unlike the shell test’s zero-leakage requirement, the seat test acceptance criteria are graduated by shutoff class: API 598 references classes ranging from relatively permissive rates suitable for basic isolation duty up through much tighter classes, with true bubble-tight shutoff, meaning literally zero measurable leakage over the test duration, reserved for critical applications where any bypass is unacceptable. This is why understanding what shutoff class a valve is actually specified and rated for matters enormously when interpreting seat test results, since a valve permitted a specific small leakage rate under its class isn’t defective for showing that rate during test, while a valve specified for bubble-tight service showing any measurable leakage at all has genuinely failed its acceptance criteria.
Test Medium Considerations
Whether testing uses water, air, or another gas affects both the practical test setup and how leakage is measured and reported. Liquid testing, typically water, allows leakage to be measured by volume, often as a specific number of drops per minute for a given size, and is generally the default medium for shell testing and much seat testing. Gas testing, using air or nitrogen, is required for services where the valve will actually see gas duty and where liquid testing wouldn’t adequately represent actual sealing performance, or simply where water contamination of the process or difficulty drying the valve afterward makes liquid testing impractical, and gas seat testing typically measures leakage as a bubble rate through a water column or with calibrated flow measurement equipment. Specification documents should call out the required test medium explicitly rather than leaving it to the manufacturer’s default practice, particularly for gas service valves where a liquid-only test might not adequately validate performance in the actual service condition.
Documentation and Traceability
A test performed without proper documentation provides little practical value beyond the moment it’s conducted, since there’s no record to refer back to if a question arises about the valve’s history later. Complete API 598 test documentation should include the specific test pressures and durations used, the actual measured leakage rate for the seat test rather than a simple pass/fail notation, the test medium, and traceability back to the specific valve serial number, allowing any future question about a particular valve’s original acceptance testing to be answered definitively rather than reconstructed from memory or assumption. This documentation becomes particularly valuable years later if a valve develops a field problem, since comparing current condition against original test data helps distinguish a valve that’s degraded in service from one that may never have met its stated specification in the first place.