Research On The Prediction Of The Brake Torque Of An Elevator Brake Based On Intelligent Algorithms

Elevator brake,as one of the main parts of the elevator,is directly affecting the life safety of passengers.Every year,the insufficient braking torque of the brake causes many casualties.The elevator brake early warning can be realized and the failure of the brake is avoid effectively.Elevator accident,this thesis takes a certain type of elevator brake as the research object.Through the analysis of the influence factors of the brake braking torque,intelligent algorithms such as BP neural network and migration algorithm are used to carry out the research on the prediction method of the brake braking torque.The specific content is as follows:First of all,the thesis of Chinese and foreign studies status of BP neural network are introduced,particle swarm algorithm and migration learning to achieve prediction.At the same time,it also analyzes the current status of elevator brake braking torque and points out the current situation.Existing shortcomings,and on this basis,the work content of the full text is proposed.Afterwards,the structure of the elevator brake are analyzed,the brake failure mode and the fault tree model are introduced.It is determined that the main mode of brake failure is insufficient braking torque.The main reasons that affect the braking torque include A variety of factors,such as brake shoe wear,pin jamming,and eccentric wear,finally determined the main influencing factors.After that,in order to realize the brake warning,the BP neural network is used as the braking torque prediction model to solve the problem that the physical model cannot predict.The PSO algorithm can improved the problem of the BP network easy to the smallest value,and the random weight method and improved convergence are proposed.The problem of the inertia weight setting of the PSO algorithm is solved.The experiment is designed and carried out.The improved PSO-BP algorithm proposed is analyzed and verified.In the next part,in order to realize the early warning of the elevator brake braking torque in the real environment,based on the neural network constructed in the third chapter.The migration algorithm is introduced,and the similarity improvement scheme is designed for the Tr Ada Boost algorithm,and an improvement is established.The MTLPSO-BP prediction algorithm solves the problem of predicting a small amount of parameter data of the brake in the real environment;by simulating different working environments of the brake.The brake test under different brake shoe wear is designed and carried out,and the relevant test data is collected.Further analysis and research are carried out on the proposed algorithm.In the end,summarizes the full text and prospects for the next research content.