Are the terms "Mean Time Between Failures" (MTBF) and "Safety Integrity Level" (SIL) related to each other? Some would think not but read on if you'd like to discover if that's true.
Equipment suppliers have been providing MTBF data for their electrical and electronic equipment for decades. The base information usually comes from a field failure study, or for new products, from an analysis of failures during accelerated testing.
MTBF is a statistical indication that an equipment item will operate within its defined specification for a stated time. The logical assumption is that repair and restoration get included in the MTBF model.
Safety Integrity Level - SIL, is a more recent metric defined in standards IEC 61508 and IEC 61511 to specify the safety performance of devices. The integrity targets apply to diverse groups of equipment, including sensors, programmable logic controllers (PLCs) and final elements such as valves.
When a safety function includes a SIL requirement, there is a need to calculate the predicted performance of the devices included in the design. This calculation can be relatively simple, or quite complex, depending wholly on the system architecture. Read about PFD and PFH calculations if you want to know more about those terms.
What many do not realise is that MTBF and SIL are both related to common underlying data
Rule of thumb
With the underpinning knowledge about failure rates, some "rules of thumb" can be derived, which may prove useful when trying to complete SIL verification calculations, for instance.
MTBF = 1/(failure rate).
(failure rate) can be derived from the MTBF by taking the inverse: 1/MTBF.
Once the failure rate is known, it is plausible to assume a 50/50 distribution of safe and dangerous failure. Usually, this will be conservative from a safety perspective.
Using the 50% Dangerous failures rate, both PFD and PFH can be calculated, and these relate directly to SIL.
In conclusion, MTBF can be related to the underpinning calculations required for SIL conformance, even if no other supporting data is available. The rule of thumb is conservative, provided the MTBF data itself is not flawed.
Don't fall into the trap of thinking that SIL relates only to such calculations. Many accidents occur because of poor procedures, lack of understanding and systematic errors. The achievement of a safe design requires much more than statistical calculations.