摘要
为提高汽轮机转子故障诊断精度,提出了一种基于萤火虫算法优化最小二乘支持向量机的汽轮机转子故障诊断方法。采用FA算法对LSSVM的惩罚系数和核函数宽度进行寻优,建立基于FA-LSSVM汽轮机转子故障诊断模型。采用转子试验台获取的样本数据进行仿真分析,结果表明,FA-LSSVM模型只出现了1次误诊断,诊断精度为97.5%,诊断效果好于其他方法,验证了所提汽轮机转子故障诊断方法的正确性和实用性。
In order to improve the fault diagnosis accuracy of steam turbine rotor,a method based on Firefly algorithm optimal least squares Support vector machine is proposed.The penalty coefficient and kernel function width of LSSVM are optimized by FA algorithm,and the fault diagnosis model of steam turbine rotor based on FA-LSSVM is established.The simulation results show that the FA-LSSVM model has only one false diagnosis,the diagnostic accuracy is 97.5%,and the diagnosis effect is better than other methods,the correctness and practicability of the proposed fault diagnosis method for steam turbine rotor are verified.
作者
张文静
Zhang Wenjing(Guangdong Machinery Technician College,Guangzhou,China)
出处
《科学技术创新》
2023年第23期10-13,共4页
Scientific and Technological Innovation
关键词
汽轮机
转子
故障诊断
萤火虫算法
最小二乘支持向量机
steam turbine
rotor
fault diagnosis
firefly algorithm
least squares support vector machine