Research on Fault Prediction Based on Elevator Life Cycle Big Data

Abstract

Fault prediction and accident prevention are the main objectives of elevator safety. Applying the big data method to the mass data generated in the whole life cycle of elevator can realize elevator fault prediction in a broad sense. Based on data collection and preprocessing of manufacturing, installation, use, maintenance, inspection and other links during the lifecycle of elevator, the elevator database was built, and the light GBM (light gradient boosting machine) decision tree algorithm was used for feature extraction, data connection, training and prediction model setting up, which could realize the elevator historical fault query of a region and the statistics of regional fault distribution, fault type, system and component, thus the elevator fault prediction could be realized.