Failure Mode and Effects Analysis (FMEA)

FMEA is a tool used to

• identify
• estimate
• prioritize
• and reduce the risk of failure in CTQs through the development of actions (process changes and innovations) and contingency plans based on Xs. It is an alternative to the analysis discussed in DMADV

You can use FMEA to

• redesign an existing unit
• design a new unit
• make existing units more robust to environmental conditions; that is, less risky with respect to failure modes.

1 identify the critical parameters and their potential failure modes for each X identified through * the type of analysis shown in Table 4.14, * a cause and effects matrix * diagram through brainstorming * other tools; that is, ways in which the process step (X) might fail 1 identify the potential effect of each failure (consequences of that failure) and rate its severity (Columns 3 and 4 of Table 4.16). The definition of the severity scale is shown in Table 4.17. 1 identify causes of the effects and rate their likelihood of occurrence (Columns 5 and 6 of Table 4.16). The definition of the likelihood of occurrence scale is shown in Table 4.18. 1 identify the current controls for detecting each failure mode and rate the organization's ability to detect each failure mode (Columns 7 and 8 of Table 4.16). The definition of the detection scale is shown in Table 4.19. 1 calculate the RPN (Risk Priority Number) for each failure mode by multiplying the values in Columns 4, 6, and 8 (Column 9 of Table 4.16). 1 Identify the action(s), contingency plans, persons responsible, and target completion dates for reducing or eliminating the RPN for each failure mode (Columns 10 and 11 of Table 4.16). $ Actions: are the process changes needed to reduce the severity and likelihood of occurrence, and increase the likelihood of detection, of a potential failure mode. $ Contingency plans: are the alternative actions immediately available to a process owner when a failure mode occurs in spite of process improvement actions. A contingency plan might include a contact name and phone number in case of a failure mode.
1 identify the date the action was taken to reduce or eliminate each failure mode (Column 12 of Table 4.16). 1 rank the severity (Column 13 of Table 4.16), occurrence (Column 14 of Table 4.16), and detection (Column 15 of Table 4.16) of each failure mode after the recommended action (Column 10 of Table 4.16) has been put into motion. 1 multiply the values in Columns 13, 14, and 15 of Table 4.16 to re-calculate the RPN (Risk Priority Number) for each failure mode after the recommended action (Column 12 of Table 4.16) has been put into motion.

Table 4.17. Definition of "Severity" Scale = Likely Impact of Failure Impact Rating Criteria: A failure could... Bad 10 Injure a customer or employee. 9 Be illegal. 8 Render the unit unfit for use. 7 Cause extreme customer dissatisfaction. 6 Result in partial malfunction. 5 Cause a loss of performance likely to result in a complaint. 4 Cause minor performance loss. 3 Cause a minor nuisance; can be overcome with no loss. 2 Be unnoticed; minor affect on performance. Good 1 Be unnoticed and not affect the performance.

Table 4.18. Definition of "Occurrence" Scale = Frequency of Failure Impact Rating Time Period Probability of occurrence Bad 10 More than once per day > 30% 9 Once every 3–4 days < = 30% 8 Once per week < = 5% 7 Once per month < = 1% 6 Once every 3 months < = .3 per 1,000 5 Once every 6 months < = 1 per 10,000 4 Once per year < = 6 per 100,000 3 Once every 1–3 years < = 6 per million (approx. Six Sigma) 2 Once every 3–6 years < = 3 per ten million Good 1 Once every 6-100 years < = 2 per billion

Table 4.19. Definition of "Detection" Scale = Ability to Detect Failure Impact Rating Definition Bad 10 Defect caused by failure is not detectable. 9 Occasional units are checked for defects. 8 Units are systematically sampled and inspected. 7 All units are manually inspected. 6 Manual inspection with mistake-proofing modifications. 5 Process is monitored with control charts and manually inspected. 4 Control charts used with an immediate reaction to out-of-control condition. 3 Control charts used as above with 100% inspection surrounding out-ofcontrol condition. 2 All units automatically inspected or control charts used to improve the process. Good 1 Defect is obvious and can be kept from the customer or control charts are used for process improvement to yield a no-inspection system with routine monitoring.

Table 4.20 shows a FMEA for a student accounts office process with four steps: student waits for clerk, clerk pulls record, clerk processes record, and clerk files record. In Table 4.20, X2 (clerk pulls record) is a critical failure mode because it exhibits such a high risk priority number (RPN = 700), while X2 (clerk processes record) is not a critical failure mode due to its low risk priority number (RPN = 42). In a brainstorming session, the clerks determined that the most likely reason for not locating a record is because another clerk was using it. Consequently, the recommended action (Column 10) was to "insert a note in a file if it is in use with the user's name." The department manager, Hiram, said that the revised best practice method would take effect no later than 1/31/03 (Column 11); in fact, it was in effect on 12/15/02 (Column 12). The calculation of the revised RPN number is shown in Columns 13 through 16. As you can see, the recommended action lowered the RPN number from 700 to 28.