Risk Determination, Evaluation, and Ranking
From two numbers to one ranking
Risk is the product of the POF and COF results, and is normally displayed on a matrix whose axes are probability categories and consequence categories. Each cell maps to a risk level - commonly high, medium-high, medium, and low - so that items can be sorted and compared. In a semi-quantitative study, the numeric product is a relative index, not an absolute frequency of harm.
Evaluation means comparing to criteria
Ranking alone does not tell you what to do. Evaluation compares each item's risk against the owner/user's risk acceptance criteria and against the risk of other items in the same study. Items above the acceptance line require action; items far below may justify extended inspection intervals, but only after the basis is documented.
Consistency is the whole game
Every item in a study must be scored with the same definitions, the same consequence categories, and the same time basis. Mixing methods, or importing a ranking from a different study with different criteria, invalidates the comparison. Aggregate risk across a unit or plant should also be considered, since many low-risk items can sum to significant total risk.
Traps
- Treating a matrix cell number as an absolute failure frequency.
- Re-ranking items after the fact to make the plan look reasonable.
- Comparing rankings produced with different methodologies or criteria.
- Forgetting that consequence-only changes (new occupied building nearby) can move an item up the ranking without any change in POF.
Rule of thumb
Rank by risk, then evaluate against criteria, then decide. If two items share a consequence category and one has POF an order of magnitude higher, it ranks one column to the right - and its inspection plan should reflect that.