Fault diagnosis of Elevator Door Machines Based on Deep Convolutional Forest

Abstract

Accurate fault diagnosis can provide decision support to elevator maintenance personnel for troubleshooting and reduce maintenance time and costs. However, the existing research on fault diagnosis based on elevator operation and maintenance data have not gained sufficient attention. To address the problem that fault diagnosis of elevator door machine systems is difficult to implement, a deep convolutional forest algorithm is designed to model the fault diagnosis of elevator door machines based on a large amount of elevator operation and maintenance data. Firstly, the algorithm adopts a convolutional neural network (CNN) for feature extraction. The deep forest algorithm is then used as a classifier to identify faults based on the extracted features. The experimental results based on the elevator door machine dataset collected reveal that the proposed algorithm shows merits compared to traditional convolutional neural network models and deep forests. The proposed algorithm can help operation management (O&M) engineers to judge the health condition of elevator door machine system, reduce the cost of elevator O&M and effectively improve the level of elevator intelligent O&M.