Fault Diagnosability Evaluation of Key Components in Helicopter Tail Drive System Based on Joint Mean‐Variance Discrepancy

ABSTRACT

Fault diagnosability is a crucial aspect of system reliability. By quantitatively assessing the diagnosability of key components in helicopter drive system, it provides a theoretical basis for optimizing the configuration of tail drive system sensors and designing diagnostic algorithms, thereby enhancing maintenance efficiency and reducing maintenance costs. To address the limitation that existing discrepancy metrics fail to comprehensively and accurately evaluate fault diagnosability, this study proposes a novel discrepancy metric based on joint mean‐variance discrepancy (JMVD) for assessing key components in a helicopter tail drive system. Firstly, a fault diagnosability evaluation model is constructed for the key components in the helicopter tail drive system, where diagnosability evaluation is essentially formulated as a discrepancy assessment between data distributions. Subsequently, the discrepancy metrics within the evaluation model are defined, leading to the construction of the JMVD metric, which quantitatively measures diagnosability. The JMVD simultaneously captures both the mean information (central tendency) and variance information (fluctuation tendency), allowing for a comprehensive assessment of discrepancies between data distributions. Finally, we validate the effectiveness of the proposed method and the rationality of the evaluation results through tests conducted on the tail drive system simulation test bench, comparing it with other classical methods.