Appendix 5

Failure Reporting, Analysis and Corrective Action System (FRACAS)

ANALYSES

1. Listing of failures:

   

   Output selectable by system, sub-system, assembly, part. Also by time period and other appropriate feature, for example location, modification status.

2. Pareto analysis of top 10–20 failure modes, selectable as above (Chapter 13). See Note 1.
3. MTBF for system, sub-system, assembly. Show number of failures for each MTBF value. Selectable as above.
4. Trend analyses, selectable as above. See Chapter 13.
5. Probability/hazard plots and derivation of distribution parameters. Selected as above. See Chapter 3.

Notes

1. The Pareto analysis should form the basis for the critical items list (Section 7.4.7). Failure reporting for critical items should be amplified as necessary to aid investigation, and special reporting, for example by phone, fax or e-mail, could be used.
2. Calendar time could be used instead of run time when appropriate. Then trend analysis, probability plots, and MTBF would be calculated on a calendar time basis. This is acceptable if run time is quite closely correlated with calendar time and if run time data are not easily obtainable, for example if the equipment does not have run time indicators. Using calendar time can be easier and cheaper, since it is then not necessary to obtain the run time at each failure and total (fleet) run time. Only the startup date for each unit and the total number of units in use need be ascertained.
3. The FRACAS should be made simple to operate and to understand.
4. Input data should be sufficiently detailed to be meaningful to users. Coding of data such as causes of failure, either by the person filling in the job report or by people at the data centre, can lead to ambiguity and errors.
5. Data input directly into portable computers, rather than onto paper, can greatly improve the quality and speed of data collection and analysis.