Two methods are available to analyze the risk of random HW failures, the first using probabilistic metrics (i.e.
quantifiable from of FTA) and the second being the individual evaluation of every residual, single and dual point
failure that can lead to a violation of a safety goal. This second form of analysis can also be considered to be a cut
set analysis.
The first method, detailed in ISO 26262, vol 5, clause 9.4.2 is the evaluation of the Probabillistic Metrics for
random HW failures (PMHF). In clause 9.4.2 the workflow for applying this method is detailed with some appropriate
examples. Depending on the results of this method it might be necessary to apply dedicated measures, this is typically
when there is a single point failure or the diagnostic coverage with respect to residual faults is less than 90%.
The second method detailed in ISO 26262, vol 5, clause 9.4.3 is a method for the evaluation of each cause of safety
goal violation due to random hardware failures. The process is illustrated by two flowcharts. In essence each single-point fault is evaluated using criteria on the
occurrence of the fault. Each residual fault is evaluated using criteria combining the occurrence of the fault and the
efficiency of the safety mechanism. This method allows identifies criteria for the use of dedicated measures to improve
the safety of the item being developed, in this case it is dependent on a failure rate class, ASIL and % diagnostic
coverage.
It is required that item only needs to comply with the results of one of these assessments
It is highly recommended that the user refers to ISO 26262, vol 5, clause 9.4 for the detailed advice and guidance
required to achieve this goal.
It is highly recommended that users refer to ISO 26262, vol 5 annex D for evaluation of the diagnostic coverage.
|