Research On The Safety Assessment Of Elevators Based On Machine Learning

With the increased number of elevators, elevator safety has become one of the main factors affecting economic development and social stability.Traditionally records and manual inspection methods were used to detect elevator safety. Many types of elevator operators and relative human factors cause difficult to guarantee quality. To improve the elevator safety assessment, the elevator safety evaluation method based on machine learning is proposed.In this paper,the main researches are as follows:1) Elevator safety checklist was developed. It is used to collect data related to elevator safety;2) Elevator safety assessment algorithm was designed.The algorithm consists of three steps: Firstly, elevator safety checklist analysis is used to analyse the elevator's initial data.Then,fuzzy mathematical principle is used to process data and the evaluation results feedback to the user in the score and grade forms.Finally,it combines machine learning with maintenance knowledge in evaluation of knowledge to provide the poor condition elevators effective and comprehensive suggestion measures and proposals;3) Elevator safety assessment software was developed.The system implements the basic equipment information management and the evaluation of the elevator's security situation;The system can identify more detailed and specific the potential risks of existing in elevator, which combines elevator safety checklist analysis, fuzzy mathematical theory and machine learning knowledge. Experiments proved that the method not only makes the results of evaluation of the elevator safety quantitative and accuracy, but also provides manager some reliable recommendations, and it has important practical significance in ensuring personal safety and property security.