Abstract:  Aiming at the problems of traditional elevator brake diagnosis method, such as complex operation, low precision and vulnerable to environmental influences, a fault diagnosis method of elevator drum brake based on variational mode decomposition (VMD) and support vector machine (SVM) was proposed. This method realized signal feature extraction of normal and fault state by several steps. Firstly, brake vibration acceleration signal during the operation of elevator was collected, and the signal was decomposed by variational mode decomposition (VMD). Then, the sample entropy was calculated with the intrinsic mode function (IMF), which was selected by spectrum analysis to represents the brake condition. In order to enhance the ability of signal feature extraction, the correlation coefficient and energy index were used as fitness function to solve the parameter selection issue of VMD, and the optimal value of signal decomposition was obtained by squirrel search algorithm (SSA). Finally, the fault diagnosis of elevator drum brake was realized by SVM and the optimal value. The result of the research indicated that the two common faults of drum brake can be identified by the fault diagnosis method effectively. The overall identification rate can reach 96%, which meets the requirement of fault diagnosis of elevator brake, and can prove the validity and feasibility of the fault diagnosis method.

Key words:  elevator drum brake; feature extraction; variational mode decomposition (VMD); support vector machine (SVM); vibration feature extraction;parameter optimization

HAO Jia-qi, XU Jin-hai, BAO Chao-chao,et al.Fault diagnosis of elevator drum brake based on VMD and SVM[J].Journal of Mechanical & Electrical Engineering,2022,39(1):112-119.