摘要
针对现有方法在汽轮机故障诊断中存在影响诊断结果的样本、诊断准确率不高等缺陷,提出了基于变分模态分解(VMD)和改进的模糊支持向量机(FSVM)相结合的故障诊断模型;采用VMD对信号进行分解,并计算出各模态分量的样本熵值作为特征值,构造特征向量;通过核模糊C均值(KFCM)聚类算法计算出不同特征向量的模糊隶属度值,并将其引入支持向量机(SVM)中形成FSVM故障诊断识别模型,同时采用粒子群算法(PSO)优化FSVM中的参数;通过本特利RK4型转子试验台模拟汽轮机故障,将得到的数据输入诊断模型中进行实验。结果表明,在保证诊断模型性能的前提下,与传统SVM相比,基于VMD和改进的FSVM方法可以有效提高汽轮机故障诊断的准确率,而且诊断时间更短。
In view of the existing methods in the fault diagnosis of steam turbine such as poor samples and low accuracy,a fault diagnosis model based on the combination of variable mode decomposition(VMD)and improved fuzzy support vector machine(FSVM)was proposed.VMD was used to decompose the signal,and the sample entropy value of each modal component was calculated as the eigenvalue,and the eigenvectors were constructed.The fuzzy membership value of different eigenvectors was calculated by KFCM algorithm,and it was introduced into support vector machine(SVM)to form FSVM fault diagnosis and recognition model,and the particle swarm optimization algorithm was used to optimize the parameters of FSVM.The Bentley RK4 rotor test rig was used to simulate the steam turbine fault,and the data was input into the diagnostic model to carry out the experiments.The results show that,on the premise of ensuring the performance of the diagnostic model,compared with the traditional support vector machine,this method based on VMD and improved FSVM can effectively improve the accuracy of the fault diagnosis of the steam turbine,and the time of diagnosis is shorter.
作者
张栋良
黄昕宇
李帅位
ZHANG Dongliang;HUANG Xinyu;LI Shuaiwei(College of Automation Engineering,Shanghai University of Electric Power,Shanghai 200090,China)
出处
《济南大学学报（自然科学版）》
CAS
北大核心
2019年第2期142-149,共8页
Journal of University of Jinan(Science and Technology)
基金
国家自然科学基金项目(61503237)
上海市自然科学基金项目(15ZR1418300)
上海市电站自动化技术重点实验室开放基金项目(13DZ2273800)
上海市科研计划项目(18020500900)
关键词
变分模态分解
模糊支持向量机
故障诊断
variable mode decomposition
fuzzy support vector machine
fault diagnosis
