UK industrial valve sourcing, automation packages and practical application support

ALLWAY VALVESIndustrial valve supply & selection Request a Quote

Technical article

SIL Requirements for Valves: Safety Integrity Levels, PFD Calculations and Proof Testing for ESD Valves

Safety Integrity Level (SIL) is now a fundamental framework for safety-critical valve applications in process industries worldwide. Whether you’re specifying an emergency shutdown valve for an offshore gas platform, selecting an SIS pressure control valve for a refinery reactor, or reviewing a functional safety assessment for a new chemical plant, understanding SIL requirements for valves is no longer optional for any engineer working in safety-critical environments. This guide provides a technically grounded explanation of SIL requirements, how they apply to valves, and what valve engineers need to know to specify, test, and maintain SIL-rated valve assemblies.

  • SIL ratings for valves quantify the probability of dangerous failure on demand — a valve in a SIL 2 safety function must have a Probability of Failure on Demand (PFD) between 0.001 and 0.01, meaning it must successfully perform its safety action 99-99.9% of the time when demanded
  • SIL capability of a valve assembly depends on the combined PFD of valve, actuator, and positioner — each component has its own failure rate data that must be summed to determine total SIS element PFD contribution
  • Proof testing interval and diagnostic coverage are the primary engineering levers for achieving required PFD targets — increasing test frequency or implementing partial stroke testing can improve SIL capability significantly

SIL Fundamentals: What Safety Integrity Level Actually Means

Safety Integrity Level is a classification defined in IEC 61508 (the parent standard) and IEC 61511 (the process industry application standard). SIL describes the reliability required of a Safety Instrumented Function (SIF) — a specific safety action performed by the Safety Instrumented System (SIS) in response to a process demand. SILs are defined from SIL 1 (least rigorous) to SIL 4 (most rigorous), with each level corresponding to a specific target range of Probability of Failure on Demand (PFD).

The PFD is the probability that the SIS will fail to perform its safety function when a demand occurs. It is essentially the average unavailability of the safety function over the proof test interval. The SIL targets are:

SIL 1: PFD average ≥ 0.1 and < 0.01 (90-99% reliability on demand). SIL 2: PFD average ≥ 0.01 and < 0.001 (99-99.9%). SIL 3: PFD average ≥ 0.001 and < 0.0001 (99.9-99.99%). SIL 4: PFD average < 0.0001 (above 99.99%). SIL 4 is extremely rare in process industry applications — the analysis and management burden is enormous and most hazard analyses show SIL 3 is the maximum justified level for typical process plant scenarios.

SIL is assigned by a formal hazard and risk assessment process — typically a Layer of Protection Analysis (LOPA) or a quantitative risk assessment — that determines the risk reduction required from the SIS function. A SIL 2 assignment means the SIS function must reduce the frequency of the hazardous event by a factor of 100-1000 compared to the unprotected case.

How Valve PFD Is Calculated

The valve assembly — body, actuator, and positioner — is the “final element” of the Safety Instrumented Function. The SIF chain also includes the initiating sensor(s) and the logic solver (SIS controller). Each element contributes its own PFD to the total SIF PFD. The total SIF PFD must be within the SIL target range.

For the valve assembly, the PFD calculation uses the simplified formula for low-demand mode (where demands occur less frequently than the proof test interval):

PFDavg = λDU × TI / 2

Where λDU is the dangerous undetected failure rate of the valve assembly (failures per hour) and TI is the proof test interval in hours. The factor of 2 reflects that, on average, a dangerous undetected failure will occur midway through the test interval — so the average time the valve is failed-and-undetected before the next proof test is TI/2.

The λDU value for a valve assembly comes from validated failure rate databases. The most widely used sources are EXIDA’s SERH (Safety Equipment Reliability Handbook), OREDA (Offshore and Onshore Reliability Data), and the IEC 61508-2 annexes. Typical λDU values for a spring-return pneumatic actuated ball valve are in the range of 1×10⁻⁶ to 5×10⁻⁶ dangerous failures per hour, depending on the valve type, service conditions, and whether partial stroke testing is implemented.

For a proof test interval of one year (8760 hours) and λDU = 2×10⁻⁶ /hr:

PFDavg = 2×10⁻⁶ × 8760 / 2 = 0.00876 ≈ 0.01

This is at the boundary of SIL 1/SIL 2, meaning this valve assembly alone would just meet SIL 1 but not SIL 2 on annual testing. If the SIF requires SIL 2, either the proof test interval must be shortened (to quarterly), partial stroke testing must be added to reduce the effective λDU, or the valve assembly must be redundant.

Diagnostic Coverage and Partial Stroke Testing

Diagnostic coverage (DC) is the fraction of dangerous failures that are detected by automatic diagnostics before they cause a loss of safety function. Detected dangerous failures are converted from dangerous undetected (DU) to dangerous detected (DD) in the calculation, and DD failures can be repaired much more promptly than DU failures — dramatically improving the PFD.

For a valve assembly without any diagnostics, DC ≈ 0 (no automatic detection). All dangerous failures remain undetected until the next proof test. With partial stroke testing implemented, DC can be 60-90% depending on the PST methodology and the valve failure modes it can detect.

The IEC 61511 formula accounting for diagnostic coverage and automatic repair of detected failures is:

PFDavg = λDU × TI / 2 + λDD × MTTR / 2

Where λDU = λD × (1 – DC), λDD = λD × DC, and MTTR is the mean time to repair for detected failures. With good PST (DC = 0.8) and λD = 2×10⁻⁶ /hr, MTTR = 8 hours, TI = 8760 hours:

λDU = 2×10⁻⁶ × 0.2 = 4×10⁻⁷ /hr; PFD(DU) = 4×10⁻⁷ × 8760/2 = 0.00175

λDD = 2×10⁻⁶ × 0.8 = 1.6×10⁻⁶ /hr; PFD(DD) = 1.6×10⁻⁶ × 8/2 = 6.4×10⁻⁶

Total PFDavg ≈ 0.00175 — now well within SIL 2 territory on an annual proof test, without needing to reduce the proof test interval. This is the direct quantitative benefit of implementing partial stroke testing.

SIL Certification for Valves: What It Means and What It Doesn’t

Valve manufacturers increasingly provide SIL certification for their products from third-party certification bodies (TÜV Rheinland, TÜV SÜD, exida, and others). It is important to understand precisely what a SIL certificate for a valve covers and what it does not.

A SIL certificate for a valve assembly certifies that: the failure rate data (λDU, λDD) published by the manufacturer is based on validated methodology; the safety manual for the valve (which specifies installation requirements, operating limitations, proof test procedures, and diagnostic requirements) has been reviewed and approved; and the valve design has been assessed for systematic failures (design errors and common cause failures) that could compromise the SIL capability.

A SIL certificate does NOT certify that: a specific SIL has been achieved in a specific application (this depends on how the valve is used, the proof test interval, diagnostic coverage, and the contribution of the other SIF elements); the valve will achieve any particular SIL without following the safety manual requirements exactly; the valve alone constitutes a complete SIF (the sensor and logic solver PFD contributions must also be included).

In practice, a SIL 2-capable certificate on a valve assembly means the valve’s λDU is low enough that it could contribute to a SIL 2 SIF — but achieving SIL 2 in the actual application requires verifying the complete PFD calculation for the entire SIF, including proof test interval, diagnostic coverage, and redundancy architecture.

Redundancy Architectures: 1oo1, 1oo2, 2oo3

Redundancy is the most powerful tool for achieving high SIL in valve applications when the single-valve PFD is insufficient. The standard redundancy architectures are:

1oo1 (one-out-of-one): A single valve must operate. Fails to safe if the valve fails. Standard architecture for most SIL 1-2 applications. PFD = PFDsingle (as calculated above).

1oo2 (one-out-of-two, parallel valves): Two valves in series; both must fail for the safety function to be lost. Dramatically reduces PFD: PFD(1oo2) ≈ PFDsingle² / 3 for uncorrelated failures. For PFDsingle = 0.01, PFD(1oo2) ≈ 3.3×10⁻⁵ — a 300× improvement. However, 1oo2 architecture also doubles the probability of a spurious trip (false activation), because either valve failing safe causes the safety function to activate. Used where very low PFD is required and spurious trips are acceptable.

2oo3 (two-out-of-three): Three valves; any two must operate for the function to be active. Provides both improved PFD (compared to 1oo1) and reduced spurious trip rate (compared to 1oo2). Common in high-consequence applications where both high reliability and high availability (low spurious trip rate) are required. Most common in the sensing element architecture but also applied to final elements in critical ESD systems.

Common cause failure (CCF) is the critical concern with redundant architectures. If two valves are identical models from the same manufacturer with the same installation and maintenance, a systematic failure mode (design error, maintenance error, common adverse condition) can cause both to fail simultaneously — eliminating the benefit of redundancy. The beta factor methodology in IEC 61508 accounts for this: a fraction β of failures is assumed to be common cause (typically 1-10%). CCF mitigation strategies include using valves from different manufacturers, different maintenance procedures, different operating principles, and periodic staggered proof tests.

Proof Test Procedures for SIL-Rated Valves

Proof testing is the periodic verification that a safety function operates correctly as-intended. For valve-based safety functions, the proof test must demonstrate that the valve will move to its safe position when demanded, within the specified time, and that the seat sealing meets the required leakage class for the process isolation requirement.

The proof test procedure must be documented, controlled, and performed by qualified personnel. The key elements of a valve SIL proof test procedure are: valve identification and location (ensure the correct valve is being tested); pre-test condition verification (instrument air pressure, control signal status); isolation of the process loop if full stroke testing is being performed during operation; initiation of the test stroke (full or partial) and measurement of stroke time; verification of valve position feedback signal; return to normal operating position; functional check of positioner calibration and communication; and documentation of all results with comparison to acceptance criteria.

The diagnostic coverage credit claimed in the SIL calculation must correspond precisely to the actual test that is performed. If the PFD calculation assumes 80% diagnostic coverage from PST, the PST procedure must demonstrably detect 80% of the dangerous failure modes included in the λDU. EXIDA and other safety assessment organisations provide detailed guidance on diagnostic coverage credits for specific PST implementations.

Common Valve Failure Modes in SIL Applications

Understanding the actual failure modes of valves in safety service informs both the proof test design and the specification of diagnostic coverage requirements.

The most common dangerous failure modes (valve fails to go to safe position when demanded) for spring-return pneumatic actuated ball valves are: valve stem/disc corrosion or scale causing the valve to be mechanically stuck; gland packing over-compression causing excessive stem friction; foreign material wedging the ball off-seat; actuator diaphragm or cylinder seal failure causing loss of spring-return force; solenoid valve failure causing failure to exhaust actuator air; and positioner failure causing failure to exhaust air or maintaining supply pressure when it should exhaust.

Each of these failure modes has a different detectability by partial stroke testing, pressure monitoring, and other diagnostics. A rigorous diagnostic coverage assessment maps each identified failure mode to the diagnostics that can detect it, and the fraction of failure rate accounted for by detected modes gives the achievable diagnostic coverage.

SIL is fundamentally about engineering discipline — defining what safety level is required, calculating whether the proposed design achieves it, verifying the calculation assumptions with actual component data, and maintaining the safety case through the valve’s operating life with appropriate proof testing and documentation. Every engineer specifying valves for safety functions should be literate in the basics covered in this guide. The IEC 61511 and IEC 61508 standards provide the full technical basis, and the safety manual of any SIL-certified valve provides the specific implementation requirements for that device.