基于排列熵与IFOA-RVM的汽轮机转子故障诊断  

Fault diagnosis of steam turbine rotor based on permutation entropy and IFOA-RVM

作　　者：石志标[1] 陈斐 曹丽华[2] SHI Zhibiao;CHEN Fei;CAO Lihua(School of Mechanical Engineering, Northeast Dianli University, Jilin 132012,China;School of Energy and Power Engineering, Northeast Dianli University, Jilin 132012,China)

机构地区：[1]东北电力大学机械工程学院,吉林132012 [2]东北电力大学能源与动力工程学院,吉林132012

出　　处：《振动与冲击》2018年第5期79-84,113,共7页Journal of Vibration and Shock

基　　金：国家自然科学基金(51576036);吉林省科技发展计划项目(20100506)

摘　　要：为了提高汽轮机转子故障诊断的识别准确率和效率,提出基于排列熵与改进的果蝇算法(IFOA)优化相关向量机(RVM)的汽轮机转子故障诊断方法。将实验数据进行自适应完备的集合经验模态分解(CEEMDAN),并选取故障特征敏感的IMF分量计算排列熵,以此构造特征样本集,进而建立"二叉树"IFOA-RVM故障分类器对特征集进行分类,其中IFOA通过两个阶段来定义果蝇群体的搜索范围来提高搜索效率,同时避免RVM核函数陷入局部最优。通过ZT-3汽轮机转子模拟试验台获得的故障数据进行实验研究,结果表明与模糊熵对比,排列熵获得的特征样本集的聚类效果明显;IFOA-RVM分类器在故障识别准确率和效率上优于FOA-RVM等其它分类器;证明了基于排列熵与IFOA-RVM汽轮机转子故障诊断方法的有效性和可行性。In order to improve the accuracy and efficiency of steam turbine rotor fault diagnosis,a fault diagnosis method for steam turbine rotor was proposed based on permutation entropy and relevance vector machine(RVM)optimized by improved fruit fly optimization algorithm(IFOA).The experimental data were decomposed with the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN).The permutation entropy was calculated with IMF components being sensitive to fault features,it was used to construct the feature sample set.Then the"two fork tree"IFOA-RVM fault classifier was established to classify the feature set.IFOA was used to define two-stage fruit fly population search ranges to improve the search efficiency,meanwhile RVM kernel function was avoided to fall into local optimum.The fault data obtained on the ZT-3steam turbine rotor analog test rig were studied.The results showed that compared with the fuzzy entropy,the clustering effect of the feature sample set obtained with permutation entropy is obvious;IFOA-RVM classifier is superior to FOA-RVM classifier in fault classification accuracy and efficiency;the validity and feasibility of the fault diagnosis method for steam turbine rotor based on permutation entropy and IFOA-RVM were verified.

关 键 词：IFOA RVM 汽轮机转子 故障诊断 

分 类 号：TK267[动力工程及工程热物理—动力机械及工程]