基于人工免疫系统的复合故障诊断技术研究

Study on Complex Fault Diagnosis Technology Based on Artificial Immune System

鉴于多缺陷并存的故障对无量纲指标影响不明显这一特性,直接从各个单独故障的无量纲指标值诊断复合故障,得知无量纲指标值对复合故障的诊断整体不是太理想。针对这一点,先利用对单独故障比较敏感的无量纲指标结合人工免疫系统的阴性选择机制构建无量纲指标免疫检测器,再采用将复合故障特征直接生成无量纲指标免疫检测器的方法来诊断复合故障。该方法操作简单,仿真试验验证该方法的有效性。

Due to the characteristic which subsequent fault is not sensitive to non-dimensional parameter, it is not promising to diagnose complex fault directly by adopting non-dimensional parameter of single fault. Non-dimensional parameter immune detector is constructed using some non-dimensional parameter which is more sensitive to single fault combined with negative selection mechanism of the artificial immune system. Then, the complex fault is diagnosed adopting two methods respectively, namely non-dimensional parameter statistic and complex fault diagnosis created directly non-dimensional parameter immune detector. The simulation experiment results prove its validity.