Downloaded from https://www.everyspec.com ADS-79D-HDBK D.4.2.2 False negative rate. CI/HIs designed to monitor flight critical failure modes must have a missed detection rate of not more than 1 in 1,000,000 occurrences of a fault.. In applications where missed fault detection could be flight critical to the aircraft‘s operation, the missed detection rate should be no more than 1 in 1,000,000 occurrences of the fault. D.4.2.3 Fault isolation rate. Once a fault has been detected, the fault should be correctly identified 95% of the time. 18 Since a component may fail in several ways, the system should identify the particular type of failure specifically within that component.
Maintenance actions resulting from HI exceedances should restore component condition with a success rate of at least 95%. D.4.2.4 Software development. Diagnostic software should be developed, at the minimum, to the integrity level required by the system criticality assessment using RTCA DO-178B. A Functional Hazard Assessment (FHA) should determine the Design Assurance Level (DAL) of the CBM system. The Safety Assessment Report (SAR) should define the DAL and rationale. This system-determined level should be a result of the end-to-end criticality assessment. D.4.2.5 Recommended maintenance actions.
A reliable alert generation process should be developed to provide maintenance recommendations and requirements. This will provide maintenance personnel information needed to perform recommended maintenance actions, to perform troubleshooting activities to isolate a fault, or to review data and determine what maintenance actions are required. D.4.2.6 Predictability. The feature to be detected and the CI that the detection updates and supports should be amenable to characterization by a mathematical function that enables prediction of future condition. Prognostics based on this characterization will be updated with usage experience. D.4.2.7 Time horizon guidance.
Prognostic algorithms that predict the time remaining before a required maintenance action and the time until the component will fail should have time horizons of sufficient length to permit the scheduling of maintenance actions and to enhance the safe operation of the aircraft. In some components incipient failures may be detectable only a few flight hours prior to component failure. This is particularly true of components operating under load at high rotational speeds. Consequently, vibration data acquisition (and subsequently data download intervals) for these components should be performed more frequently than for other components. D.4.3 CI and HI Confidence Level Requirements.
CIs and HIs can be classified into several confidence strata based on the ground truth evidence associated with any particular implementation. For example, levels of evidence required for maintenance credits are of course higher than levels required for enhanced maintenance. The following are examples of how to classify the different CIs and HIs respective to their ground truth data base size. Levels 1, 2, and 3 are considered advanced and require limited to no oversight. Levels 4 and 5 are part of engineering development, and Level 6 is essentially the diagnostic graveyard. D.4.3.1 Level 1: Verified and Validated CIs/HIs.
CIs and HIs that are certified through the process outlined in Appendix I for maintenance credit are Verified and Validated. They typically 18 SAE Aerospace, Aerospace Recommended Practice ARP5783, Health and Usage Monitoring Metrics, Monitoring the Monitor, Jan. 11, 2008. 79