A Research On Elevator Maintenance Decision Based On The Cox Model

With the rapid development of urbanization in China,there has been 40 million elevators in China until the end of 2015,and the number is increasing by fifteen to twenty percent every year.In recent years,there has been frequent elevator accidents that lead to casualties and huge economic loss,which expose the defects of current maintenance policy.This paper aim to find factors that influence elevator reliability and provide suggestions on elevator maintenance services.First,we connect four elevator characteristics including type of cities,type of buildings,age of elevator and type of elevator with elevator failure rate,and builds a Cox proportional hazard model,the empirical research demonstrates that the first three covariates are significant factors.Based on Cox model's good ability of failure prediction,the paper builds a preventive maintenance decision model with the goal of minimizing the total cost of repair and maintenance per unit time,and provides personalized maintenance intervals for different types of elevators according to their service environment and their own age.The result shows that as the environment is becoming worse and the elevator age is older,the corresponding maintenance intervals is shorter.Based on extended time-dependent Cox model,this paper also research on real-time health status of elevator.Based on the failure data from an elevator company,we conduct an empirical study and find that maintenance measures,maintenance interval,failure number in past 3 months and average temperature in the last week are significant factor to real-time reliability of elevator.Based on the coefficient of significant variables,in order to improve elevator reliability,we provide some suggestions on maintenance services for elevator company.