Intelligent Guidance Method for Elevator Emergency Treatment based on Automatic Recommendation and Fault Prediction

Abstract

In order to reduce the handling time of elevator failure and speed up the rescue of trapped personnel, an intelligent guidance method for elevator emergency treatment based on automatic recommendation of rescue units and prediction of fault causes is studied on the basis of elevator emergency treatment platform. The automatic recommendation module builds a multi-dimensional rescue unit capability evaluation index system, which establishes result recommendation methods such as recall, single-index recommendation, and comprehensive recommendation to achieve the optimal rescue unit recommendation for faulty elevators. The fault cause prediction module uses a variety of pre-trained word embedding models to vectorize fault text data on historical fault data sets, uses elevator fault text clustering algorithm based on attention mechanism and BI-LSTM model to obtain elevator fault labels, and uses the Boosting ensemble learning algorithm to construct an elevator fault prediction classification model for the marked elevator historical fault data set. The experimental results show that when the elevator fails, the automatic recommendation module can recommend the optimal rescue unit, and the fault prediction module can predict the cause of the elevator failure in real time, which quickly and accurately locates the fault area. For rescuers, it is convenient to deal with elevator failure in a targeted manner and greatly reduces the repair time. Therefore, this research is of great significance for speeding up rescue, improving emergency response capabilities, and ensuring the safe operation of elevators.