Elevator dynamic monitoring and early warning system based on machine learning algorithm

Abstract

AbstractWith the rapid progress of modern society, elevator, as a vertical means of transportation in modern high‐rise buildings, has become an indispensable part of People's Daily life. Because the elevator is a kind of mechanical equipment with a very high frequency of use, the peak hour will reach 250 times per hour, and because the elevator in the daily operation of the running parts work many times, the operating system is complex, good quality. Different passengers lead to frequent elevator failures. The dynamic monitoring and early warning system is a system which, according to the characteristics of the elevator, monitors the change trend of risk factors in the elevator by collecting relevant information, and evaluates the degree of deviation of various risk states from the early warning line, so as to send early warning signals to the decision‐making level and take pre‐control countermeasures in advance. Machine learning is a multi‐domain interdisciplinary discipline, which specialises in studying how computers simulate or realise human learning behaviours to acquire new knowledge or skills and reorganise existing knowledge structures to continuously improve their own performance. This study aims to monitor and warn the elevator dynamics. Machine learning algorithm is introduced in this work, and particle swarm optimisation algorithm is used to improve the model. The model is optimised, and the experimental comparison shows that the optimisation of model parameters can further improve the accuracy of elevator load prediction. Then, the deep network prediction model is constructed, and the number of network layers is determined through experiments. Firstly, a multi‐layer network is trained in the training model, and this process can be used to extract the indicators of inducing factors, so as to avoid the difficult and time‐consuming manual extraction of indicators, and improve the prediction accuracy of the system. After five experiments, the prediction errors of the system are 1.14%, 0.54%, 0.07%, 0.12%, and 0.68%. The experimental results show that machine learning algorithm and particle swarm optimisation algorithm play an active role in the research of elevator dynamic monitoring and warning system.