| # | Finding | How to check |
|---|---|---|
| 1 | Failure rate in the wrong unit (per year or FIT entered as per hour) | λDU plausible? First Out AI flags values above 10⁻⁵ per hour |
| 2 | Failure rate of the valve body only, without actuator and solenoid | Source covers the whole assembly? |
| 3 | Proof test coverage assumed 100 % | Test procedure reveals every failure mode? |
| 4 | β = 0 for redundant channels | Common cause checklist done? |
| 5 | Required PFDavg from LOPA ignored; only the SIL band checked | Compare with the LOPA result, not the band |
| 6 | Fault tolerance not checked against IEC 61511 | HFT per subsystem against the target |
| 7 | Systematic capability of an element below the target | Certificates or prior use evidence for every element |
| 8 | Test interval longer than the turnaround allows, or not achievable online | Maintenance plan agrees with T1? |
| 9 | MTTR assumed 8 h when spares and permits make it days | Realistic repair time entered? |
| 10 | Demand mode wrong (frequent demands treated as low demand) | Demand rate stated and justified |
| 11 | Spurious trip rate not considered | MTTFS acceptable to operations? |
| 12 | Solenoid, interposing relays or barriers missing from the groups | All elements in the signal path included? |
| 13 | Same transmitter used for control and trip | Independence from the initiating cause |
| 14 | Calculation not updated after a device change | Version and date of inputs |
| 15 | Assumptions not carried into the SRS | SRS lists T1, PTC, MTTR, β, mission time |
In First Out AI
The automatic input checks cover several of these (units, β = 0, low coverage, missing sources, fault tolerance, systematic capability, required PFDavg). Ask First Out AI "Where is my mistake?" after calculating; it reads the checks and your inputs.
Written for First Out AI as engineering guidance. It supports engineering work and does not replace it; check every value against your standards, certificates and company rules with a competent functional safety engineer.