Failure Mode and Effects Analysis (FMEA): Risk Priority and Control Plan

Why it matters

Failure mode and effects analysis, FMEA for short, serves to examine systematically in what form a job can fail and what effect this will have. Failure mode, cause and effect are not the same thing. A part being placed reversed is a failure mode, the function not being achieved is an effect, and a gap in the orientation condition may be a cause explanation to be tested. Filling these three fields with the same statement reduces the thinking value of the analysis. Production, quality, maintenance and the people who do the job must look at the same process boundary. The aim is not to prepare a long list but to determine which risk to act on first and on what evidence. The analysis does not replace daily control or the product decision.

The 1–10 scoring scale used in this lesson is synthetic for training; it does not represent the score table of any particular outside standard. In severity a larger score describes a heavier effect, in occurrence a larger score a less favorable occurrence level, and in detection difficulty a larger score that the defect is harder to catch. Thus a falling detection score means better detection. Scores are not measured probabilities; an occurrence score of 5 cannot be counted as a 50% probability of error. Without keeping the same scoring rationale, the products of different teams are not compared. Boundary conditions, supporting records and uncertainties are explained; choosing a score only to reach a desired result misleads the analysis. This dataset was generated synthetically for training; it does not represent a real business, employee or customer study.

The risk priority number, RPN, is the product of the severity, occurrence and detection difficulty scores. The product is an auxiliary indicator that organizes the review; it cannot carry all risk decisions alone. A high-severity failure can get a small product because of a low occurrence score and an easier detection score. This is not a reason to postpone action against a heavy effect. Also, the same product can arise from different score combinations; these combinations do not require the same measure. The priority decision first examines the protection that the effect and high severity require, and then the occurrence and detection findings. The numbers used here do not form a single universal acceptance threshold or a ready ranking rule to be applied across all sectors.

The control plan turns the analysis into workable action. The failure mode, the characteristic being controlled, the prevention or detection method, the frequency, the owner and the deviation response are linked together. Writing only to be checked is not enough; what will count as conforming, where the result will be recorded and how the product will be protected on failure must be understood. A proposed improvement is not a verified control. A score cannot be lowered just because installation is complete; a record showing that the target failure condition was tested is required. When the control plan version changes, the instruction and training on the floor are updated as well. Proving a cause explanation is the subject of M12; a cause written into the FMEA record is not accepted as real merely because it was written.

A control being skipped, being applied but failing, and its result not being recorded are separate events. The first requires examining the execution of the work, the second the detection power, the third the chain of evidence. These headings are not in themselves evidence of cause. The team must explain which condition it has secured against the heavy effect and by which verification it showed this. Adding a control or lowering the product is not adequacy by itself.

Formula