Definition:
Failure Rate (λ) is the expected number of failures per unit time for a system or component. This is a critical part of calculating probability of failure on demand (PFD) and risk reduction.
Within functional safety, whenever failure rate is used, it is always random failure rate. Sometimes this is called random failure rate, but the “random” is implied. Note “opposite” of that would be systematic failures, which are not random, thus there is no systematic failure rate.
There are four types of failure rates created by combining:
- Dangerous or Safe – meaning the failure impacts the ability of the safety function to perform its function, or to not.
- Detected or Undetected – meaning the diagnostics detects the error, or not.
Failure rate importantly includes both independent failures and common cause failures (CCF). This is why the PFDavg equations are structured and built as they are, especially for NooN architectures (like 2oo2 or 4oo4). See more on that in this article.
See this much deeper dive on failure rate as well.
Key Points:
- Expressed typically as failures/hour.
- Dangerous Detected (λDD), Safe Detected (λSD),
- Dangerous Undetected (λDU), Safe Undetected (λSU)
- Typically reported on the SIL Certificate.
- MTTF is the inverse of failure rate (for a constant hazard rate model, very typical for SIS)
- Note that proof tests attempt to capture λDU
Example:
A valve has a dangerous undetected failure rate (λDU) of 2E-6 failures/hour as reported on the SIL Certificate. This term is used in the PFD calculations.
See Also: PFDavg, random failure, detected failure, undetected failure, MTTF, beta factor, beta factor in depth article, common cause failure
Cited Source:
- IEC 61508-6:2010, Annex B.