基于自我-非我识别机理的状态监测与故障诊断

基于人工免疫自己非己识别机理(Self—nonself Discrimination—SND)的反面选择算法(Negative-selection Algorithm—NSA)，为复杂机电系统的故障监测和诊断提供了新的思路。针对现有故障监测中，反面选择算法存在的不足，文章提出采用距离匹配的反面选择算法(Negative—selection A1gorithm using distance matching—NSAD)，采用遗传算法，生成适合故障样本的故障种群，并采用欧几里德距离为适应函数，以产生与故障模式匹配的监测器，使监测器对特定故障模式具有更好的敏感性。文章以旋转机械为对象，进行验证，结果表明NSAD较NSA监测准确率高，误诊率低。

Condition Monitoring and Fault Diagnosis Based on Self-nonself Discrimination Mechanism

Negative-selection Algorithm (NSA) based on Self-nonself Discrimination,offers a new method to condition monitoring and fault diagnosis in complex mechatronics systems.For the defects of NSA in condition monitoring,the Negative-selection Algorithm using distance matching-NSAD is presented,in NSAD,GA is adopted to generate the fault population,which best matches the fault pattern.To get an efficient detector,which is more sensitive to the specific fault mode in the real applications,Euclidean Distance is taken as a fitness function to select the best population to match the fault mode. NSAD is verified through the experiments of rotating machinery. The result indicates that the lower mistaken diagnosis rate and the higher accuracy rate can be obtained using NSAD.