Downloaded from https://www.everyspec.com ADS-79D-HDBK One should consider the criticality of the failure associated with a component when specifying a data compromise recovery strategy. A more conservative procedure should be specified when failure consequences are more severe. As a result, the CBM system designer may specify a different recovery procedure for each component part number in the maintenance tracking database. In the worst case, it may be specified that a component be replaced immediately when data loss occurs. B.5.3 Digital source collector validation for structural usage monitoring.
Prior to deploying the flight regime measurement package as part of operational usage monitoring, a test aircraft should be instrumented for demonstration that the algorithms can accurately classify flight regimes. For developmental programs this can be performed as part of the Flight Loads Survey Testing (FLST) where the aircraft will be exposed to the range of flight regimes specified in the design usage spectrum. The bin range of regimes should be set for an aircraft equipped with usage monitoring in order to maximize maintenance credits. The current large bin ranges and associated loads data will not permit maximum benefits for a monitored aircraft.
For legacy aircraft, flight testing should be performed to verify the capability of the usage monitoring system in identifying the regimes of the design usage spectrum. Also, additional FLST may be beneficial to maximize maintenance credits for usage monitoring. These additional flights allow smaller bin ranges that will improve the accuracy of fatigue damage calculations. For example, if the current regimes bins turns into 45 and 60 degree angle of bank (AOB), any turn recognized by the usage monitoring system with an AOB less than 45 degrees would be assigned to the damage accumulated for a 45 degree turn.
Gathering load data for AOB less than 45 degrees and restructuring the bin range for turns will allow more accurate tracking of usage and realistic damage fraction calculations. B.5.3.1 Algorithm validation methodology. A series of flights should be performed with a test aircraft that is fully equipped with the regime measurement package and additional recording systems for capturing data needed to evaluate and tune the algorithms. Engineering should prepare a series of flight cards identifying the maneuvers for which algorithms have been developed. Maneuver descriptions and tolerances should match those used during the flight load survey.
The monitoring flight test engineer should know the sequence in which the pilots are flying the maneuvers and their target severity and duration. After the flight, the data records will be surveyed to determine which maneuvers were sufficiently detected and which maneuvers require improved algorithms. Algorithm optimization will be performed and a subsequent flight made in a totally different sequence using the improved algorithms. The post flight process will be the same. Usually two optimization flights are sufficient but additional flights may be necessary to achieve the desired regime classification accuracy.
For aircraft with a very large range in gross weight (GW) it may be desirable to check the accuracy of the algorithms at very heavy and very light GW. Additionally, an aircraft that has a very high altitude mission may require algorithm validation at both high altitude and near sea level conditions. Finally, after completion of optimization to achieve the designated accuracy, a comprehensive flight card should be developed which incorporates all of the maneuvers for which the algorithms have been developed.
Without being provided knowledge of the flight card content, the regime recognition design team should demonstrate the ability to identify the maneuvers flown, their severity and duration. Sufficient flight time and flight conditions should be properly identified such that any unrecognized regimes would introduce less than 0.5% under-prediction of fatigue damage fraction based on the design usage spectrum. B.5.3.2 Accuracy. To avoid under predicting damage fraction by more than 0.5%, it is recommended that the CBM regime recognitions algorithms should demonstrate that they can identify 57