Skip to content
SILProof

Functional safety · IEC 61508-6 · IEC 61511

Proof test interval calculator

Find how often a 1oo1 device must be proof tested to stay within a target PFDavg.

FIT
%
Advanced
years

Adds the failures the proof test cannot find, which stay until overhaul or replacement. Leave empty for the simple form.

Try an example:
Proof test intervalSimple form (1oo1)

25.4 years

222,200 h for PFDavg 1 × 10⁻²

PFD over time: each proof test resets the tested share
00.010.020204060YearsPFDTarget PFDavg 1 × 10⁻²
Interval
222,200 h · 304 months
PFDavg at this interval
1 × 10⁻²
λDU
1 × 10⁻⁷ /h (100 FIT)
Not included
failures the proof test cannot find: add a mission time under Advanced

Working

  1. T = 2 × PFDavg / (λDU × PTC) = 2 × 1 × 10⁻² / (1 × 10⁻⁷ /h × 0.9) = 222,200 h
  2. Check: PFDavg = λDU × [PTC × T / 2] = 1 × 10⁻²

What the calculator is for

A proof test finds the dangerous undetected failures that the diagnostics cannot, and returns the device to as-good-as-new. The longer the interval, the longer a hidden failure can wait for a demand, so PFDavg grows with the interval. This calculator turns a target PFDavg into the longest interval that meets it.

  • Simple form: PFDavg ≈ λDU × PTC × T / 2, so T = 2 × PFDavg / (λDU × PTC).
  • With a mission time MT: PFDavg ≈ λDU × [PTC × T / 2 + (1 − PTC) × MT / 2], adding the failures the proof test cannot find.

Worked example

A transmitter has λDU = 200 FIT, the proof test covers 95 % of its dangerous undetected failures, and its share of the SIF budget is PFDavg = 2 × 10⁻³. The simple form gives T = 2 × 2 × 10⁻³ / (2 × 10⁻⁷ × 0.95) = 2.4 years.

That ignores the 5 % never found by the test. Over a 15-year mission they add 2 × 10⁻⁷ × 0.05 × 15 × 8,760 / 2 = 6.57 × 10⁻⁴ to PFDavg, and the interval drops to 1.61 years. Below 100 % coverage, always check the untested share.

Setting the interval in practice

The calculated interval is an upper limit, not a recommendation. Plants usually round down to a shutdown or maintenance cycle, check that the test can really be done safely at that frequency, and keep the interval within the device's useful life. When the result is shorter than practical, look at the other levers: a more complete proof test, partial-stroke testing for valves, a device with a lower λDU, or redundancy.

Quick reference

Interval in years for a target PFDavg (simple form, PTC 100 %)
λDU (FIT)10⁻¹10⁻²10⁻³
5045745.74.57
10022822.82.28
20011411.41.14
50045.74.570.457
100022.82.280.228

Common mistakes

  • Assuming a perfect proof test. Few tests reach every failure mode; a realistic PTC is often 70 % to 95 %.
  • Forgetting the untested share. With PTC below 100 %, the failures the test misses build up over the whole mission and can use most of the PFDavg budget on their own.
  • Applying it to the whole SIF. This is a 1oo1 estimate for one device; the SIF combines sensors, logic and final elements, often redundant.
  • Intervals beyond the device's useful life. A constant failure rate only holds within the useful life stated by the manufacturer.

Basis

The simplified 1oo1 low-demand expressions of IEC 61508-6 Annex B with proof test coverage, as used for screening in IEC 61511 practice. Repair time and detected failures are left out; use SIL verification for the complete function.

SILProof

Part of a safety instrumented function?

SILProof keeps the risk basis, SRS and three-gate SIL verification of each safety instrumented function on one record.

Verify the whole SIF

Frequently asked questions

How do I calculate a proof test interval?

For one device in low demand, T = 2 × target PFDavg / (λDU × PTC). With 100 FIT and a target of 10⁻², that is about 22.8 years at 100 % coverage, before any useful-life limit.

What is proof test coverage?

The fraction of dangerous undetected failures that the proof test reveals. Failures it misses stay hidden until the device is overhauled or replaced.

Why does a lower coverage shorten the interval so much?

The untested share is only renewed at the end of the mission, so it grows for years. Its contribution can approach the whole PFDavg target by itself.

Is the result valid for a redundant subsystem?

No. It is a 1oo1 estimate. Redundant voting, common cause and the other subsystems need a full SIL verification.

All functional safety calculators →