基于故障特征信息量的诊断策略优化仿真研究

Research on DMFT Test Method Based on Hybrid Diagnostic Model

针对如何使用较低的测试费用对复杂系统的故障快速诊断和隔离,进而生成诊断策略的问题,提出一种基于故障特征信息量的诊断策略优化算法。算法采用一步前向回溯的方式,预先生成基于故障特征信息熵算法的完全测试序列,然后通过评估测试序列的平均测试费用来确定当前一步的最优测试点。实验表明,在故障检测率和故障隔离率均达到要求的情况下,所提算法要比故障特征信息量诊断算法的期望测试费用更低,验证了上述算法在降低系统检测时测试费用具有一定优越性。

Aiming at how to use the lower test cost to quickly diagnose and isolate faults of complex systems and generate diagnostic strategies, a diagnostic strategy optimization algorithm based on fault feature information is proposed. The algorithm uses a one-step forward-backtracking method to pre-generate a complete test sequence based on the fault feature information entropy algorithm, and then determines the optimal test point of the current step by evaluating the average test cost of the test sequence. Experiments show that the proposed algorithm has lower cost than the fault feature information diagnostic algorithm when the fault detection rate and fault isolation rate are both satisfactory, and the superiority of the algorithm is verified.