U.S. Army ADS-79D-HDBK — ADS-79D-HDBK.pdf

ADS-79D-HDBK Downloaded from https://www.everyspec.com FIGURE K-8. Mechanical health-based prognostics detailed verification process If the entry criterion is met, the next step is to establish the prognostics requirements and level of accuracy desired. Some suggested prognostics criteria are: Rate of Change of RUL ≈ -1: In other words, if a component has 100 hours of RUL and it is run for 1 hour, the prognostics algorithm should predict 99 hours at the end of the 1 hour operation. Second Derivative of the RUL ≈ 0: The value of the second derivative of the RUL is a measure of the stability of the algorithm. The closer it is to 0, the more stable the estimation.

Next, capture these requirements in a Configuration Specification Document. Upon completion of the Configuration Specification Document, the algorithm and configuration can be implemented. It is important to note that the term verification refers to testing against the requirements as stated in the Configuration Specification Document. The algorithm, configuration and other system elements designed for the component‘s health credit will be tested to assure correct implementation. Whether or not they function appropriately in the real world application is determined during validation.

Given this definition, the ―Verify Prognostics Algorithm‖ step is only used to test the algorithm against its requirements. The tests should verify and validate that the system elements work as designed. Such verification can be conducted on a bench-type or 207