Applications Of Elevator Comprehensive Supervision And Quality Improvement Based On Data Mining Techniques

With the development of economy and technology,elevator has become a daily transportation tool.In recent years,the frequent occurrence of elevator failures has caused losses to people's lives and property,elevator safety has attracted the attention of the government.Through the analysis and mining of elevator related data,this thesis provides decision-making for elevator comprehensive supervision and quality improvement.This paper is divided into three parts which are the fault prediction model of elevator installation site and the algorithm of responsibility definition,the algorithm of elevator maintenance area division and maintenance site selection,elevator rescue mode design.For the first part,this paper uses XGBoost algorithm with elevator related data to build a model which realizes the prediction of whether the elevator installation sites malfunction in the next month.Data preprocessing,feature extraction,model training and model evaluation are included in the building process of this model.The experiment of the feasibility for the prediction model is conducted.This paper also proposes an algorithm for grading relevant factors of elevator failure based on Bayesian formula and the records of elevator fault.When the elevator fails,the responsibility score of each factor is determined by the algorithm which reminds responsible parties to enhance supervision with the highest responsibility score.For the second part,this paper proposes a maintenance area division algorithm based on grid clustering and a weighted time-distance maintenance site selection algorithm.The distribution of elevators and the failure rate of elevators is considered in this algorithm.The results of experiments are proved that the area division algorithm and site selection algorithm can significantly shorten the time from the maintenance site to reach the faulty elevator.For the last part,a dispatch rescue mode was proposed which in order to shorten the time to deal with elevator failure.When the elevator failure occurs,by analyzing the distance between the faulty elevator and the maintenance site and the experience of the maintenance personnel,the priority of the maintenance site is sorted.The simulation experiment shows that the dispatch mode greatly improves the efficiency of handling elevator failure.