Research of bigdatabased elevator fault diagnosis and predication
                                      CHEN Zhiping1, WANG Zan1, ZHANG Guoan2, LI Chunguang1, LI Zhewei1, HE Ping1
(1.School of Mechanical Engineering, Hangzhou Dianzi University, Hangzhou 310018, China;2.Shaoxing Special Equipment Inspection Institute, Shaoxing 312071, China)

Abstract: Aiming at applying big data analysis methods into elevator fault diagnosis and prediction, first, a large number of elevator inspection data was collected, and feature parameters related to car vibration were extracted from the inspection data. Then, the overall scheme was presented. By using supervised and unsupervised learning methods, data mining and analysis were conducted on the extracted feature parameters to find out the intrinsic relationship between the faults of elevator mechanical system and the car vibration signal. Finally, the elevator mechanical system faults could be diagnosed and predicted according to the magnitude of vibration characteristics and their changing trends. The results indicate that big data analysis method can be used to accurately diagnose and predict the failure of elevator mechanical system.

Key words: elevator; fault diagnosis and predication; big data; supervised learning; unsupervised learning