Fault diagnosis of subway auxiliary inverter based ON LMD AND QPSO-LSSVM

作者：Gao Junwei^*; Cheng Liang; Qin Yong; Leng Ziwen; Zhang Bin

来源：ICIC Express Letters, 2015, 9(1): 37-43.

• Frequency transformations
• Least square support vector machines
• LMD
• Local mean decomposition (LMD)
• LSSVM
• Non-stationary characteristics
• QPSO
• Quantum particle swarm optimization

Focusing on the non-stationary characteristic of the fault signal of subway auxiliary inverter, this paper puts forward a new model about fault diagnosis of subway auxiliary inverter, that can realize the benefits of local mean decomposition (LMD), quantum particle swarm optimization (QPSO) and least square support vector machine (LSSVM) method. At first, LMD method decomposes the original signals to get a series of PF components. Secondly, the method of energy feature extraction to extract feature vectors from original signals is used in this paper. Finally, QPSO optimizes the parameters of LSSVM, and then LSSVM diagnoses three types of fault diagnosis, frequency transformation, voltage fluctuation and pulse transient. According to the analysis results of the fault signal of subway auxiliary inverter, it shows that the model can identify these faults accurately and efficiently.

• 出版日期2015
• 单位北京交通大学