Downloaded from https://www.everyspec.com ADS-79D-HDBK This approach involves setting thresholds for faulted components (requiring mandatory maintenance); and un-faulted components (to include components with tolerable defects, for which maintenance is optional). Since it is vibratory CBM being discussed herein for transmissions and engines, there are also numerous faults to consider with different thresholds (e.g. bearing spall, chipped/scuffed gear teeth, and bent shafts). For each threshold, a binomial distribution must be established which means a separate set of component samples must be established for each verification/validation effort on each threshold.
Also, each test for faulted and un-faulted components is assumed to result in a success or failure and the outcome on one test does not affect the outcome on another test. Finally, the configuration of the monitoring hardware (i.e. processor, wiring, sensors, sensor location, amplifiers) is assumed to be identical for each test. An advantage to this approach is that a sample size can be established in advance, for any desired number of allowed failures, without detailed knowledge of how the CI/HI data is distributed.
A disadvantage to this approach is that the latest vibratory condition monitors installed on legacy US Army rotorcraft have not evolved to be the optimal monitor to capture all component failure modes on complex systems such as transmissions and engines. Further, replacing TBOs on legacy systems with On Condition monitoring may not be a valid option to consider unless ―No Build Windows‖ are considered as viable. No Build windows would be implemented to avoid issues when fatigue critical components on a transmission or engine are involved.
A No Build window refers to a depot process of not rebuilding an engine or gearbox incorporating a fatigue life limited component that is within a specified proximity range of the published retirement life. Another disadvantage perceived to this approach is the relatively high cost and lengthiest time among the alternatives discussed within this Appendix. The costs and time are associated with the requisite sample size to achieve 90% confidence and 90% probability of detection (reliability). These percentage levels are used by the Aviation Engineering Directorate to assess risk.
Further substantiation for the use of high confidence and reliability levels in aviation may be found in various statistical books, websites, and documents 48,49,50,51 . Per the reliability methodology, sample size can be calculated by using the following equations 52 : n 1 = ln (1-confidence) / ln (reliability) = faulted (Equation 1) n 2 = ln (1-confidence) / ln (reliability) = un-faulted (Equation 2) 48 Rees, DG, ―Foundations of Statistics,‖ CRC Press, 1987. 49 NIST/SEMATECH e, National Institute of Standards and Technology, ―Handbook of Statistical Methods,‖ Engineering Statistics Handbook, 2011. 50 Olson, E. and T. Olson, ―Real Life Math: Statistics,‖ Walch Publishing, 2000.
51 AIR 5113 Aerospace Information Report, Society of Aerospace Engineers (SAE), ―Legal Issues Associated with the Use of Probabilistic Design Methods,‖ 2002-06. 52 Abernethy, B. Robert, ―The New Weibull Handbook – Reliability & Statistical Analysis for Predicting Life, Safety, Survivability, Risk, Cost, and Warranty Claims,‖ Fourth Edition, 2000. 154