基于扩展记忆粒子群优化支持向量机的汽轮机故障诊断  

Steam Turbine Fault Diagnosis Based on Extended Memory Particle Swarm Optimization Support Vector Machine

作　　者：范汉林 FAN Han-lin(Zhuhai Shenneng Hongwan Electric Power Co.Ltd.,Zhuhai 519000,China)

机构地区：[1]珠海深能洪湾电力有限公司,广东珠海519000

出　　处：《电气开关》2023年第3期68-71,共4页Electric Switchgear

摘　　要：为了提高汽轮机故障诊断正确率,提出了一种基于EMPSO-SVM的汽轮机故障诊断方法。采用扩展记忆系数对PSO算法进行改进,以提高PSO算法的优化性能,采用扩展记忆粒子群算法对支持向量机进行优化,建立了基于EMPSO-SVM的汽轮机故障诊断模型。采用实际算例进行仿真分析,结果表明,EMPSO-SVM模型诊断结果的正确率高达95%,相比PSO-SVM模型正确率提高了7.5%,验证了模型的正确性和实用性。In order to improve the accuracy of turbine fault diagnosis,a method of turbine fault diagnosis based on EMPSO-SVM is proposed.Extended memory coefficient is used to improve PSO algorithm to improve the optimization performance of PSO algorithm.Extended memory particle swarm optimization algorithm is used to optimize support vector machine,and a steam turbine fault diagnosis model based on EMPSO-SVM is established.The simulation results show that the correct rate of the diagnosis results of EMPSO-SVM model is as high as 95%,which is 7.5%higher than that of PSO-SVM model,which verifies the correctness and practicability of the model.

关 键 词：汽轮机 故障诊断 扩展记忆粒子群 支持向量机 

分 类 号：TK26[动力工程及工程热物理—动力机械及工程]