[3C3] Fault diagnosis of elevators based on empirical mode decomposition and random forest algorithm
T Wang¹, Y Liu¹, M Wang², D Lu² and Y Xu¹
¹Xi’an Jiaotong University, China
²China Special Equipment Inspection and Research Institute, China
To improve the accuracy of elevator fault diagnosis, the paper proposes a fault diagnosis method for elevator system brakes based on empirical mode decomposition (EMD) and the random forest algorithm. The collected data during elevator running is decomposed using EMD to obtain the intrinsic mode functions at different time scales. Then, kernel principal component analysis (KPCA) is applied to reduce the dimension of the decomposed mode functions while retaining the key information of the original dataset. The extracted features are input into a random forest model for training and fault identification. Experimental results demonstrate that, compared to traditional K-nearest neighbours (KNN) methods, the proposed EMD random forest-based fault diagnosis method achieves superior performance in brake fault recognition.
