免疫粒子群算法及其在矿井提升机故障诊断中的应用

Artificial Immune Particle Swarm Optimization for Fault Diagnosis of Mine hoist

基于人工免疫系统的故障诊断方法是人工智能领域发展起来的一个十分活跃的分支.为了提高免疫算法在矿井提升机故障诊断系统中的执行效率，通过对诊断问题进行更精确的建模和分析,提出了将免疫模型和离散粒子群进化算法相结合的提升机系统的故障诊断方法.该方法在免疫形态空间中采用核主元形式的相似性度量，解决了传统距离判别函数法在故障诊断中存在误差较大等问题.仿真结果表明，该方法能够适应诊断过程中出现的不确定性，并实现多故障诊断.

This paper presents an intelligent methodology for diagnosing incipient faults in mine hoist. In this fault diagnosis system, in order to enh_ance the immune algorithms performance, we propose the improved immune-based symbiotic a new evolu- tionary learning algorithm. This new evolutionary learning algorithm is based on Discrete Particle Swarm Optimization （DPSO） technique to improve the mutation mechanism. Also to solve the problem that exists in fault diagnosis based on the traditional method using distance discfiminant function, an improved method based on immunity strategy with similarity measurement of princi- ple component kernel is presented. The effectiveness of the DPSO based immune algorithms is demonstrated through the classifica- tion of the fault signals in mine hoist. Simulation results show that the new scheduling algorithna can deal with the uncertainty situa- tion and be suitable for multi-faults diagnosis,compared to the traditional scheduling algorithms.