Pattern Recognition And Scheduling Method For Multi-Car Elevator Group Control System

With the development of high-level and intelligent buildings,multi-car elevator has become a frontier issue in the field of vertical transportation.Multi-car elevator by adding a number of elevator cars in the elevator shaft,improved the lack of traditional elevator,improve the efficiency of the elevator,and to a certain extent,reduce the construction costs and elevator energy consumption.Based on the scheduling problem of multi-car elevator control system,this thesis analyzes the dispatch strategies,control strategies and collision avoidance strategies of the in the vertical transportation field,and proposes a mixed kernel traffic flow data recognition algorithm based on KPCA and SVM and a multi-car elevator systems dispatch strategies using dynamic zoning based on Fast R-CNN.The main research contents of this thesis are as follows:(1)Based on the analysis of the development trend at home and abroad and the problems that need to be solved,this thesis proposes a traffic flow data recognition method based on hybrid kernel and a multi-car elevator scheduling method based on machine vision and zone limitation.(2)It discusses the difference between multi-car elevator system and the traditional elevator system,introduces the characteristics and operation mode of double-car elevator,circulatory multi-car elevator,single-shaft multi-car elevator,bifurcate circulatory multi-car elevator and three dimensional circulatory multi-car elevator.(3)This thesis introduces the characteristics of the six traffic modes of the multi-car elevator system,namely,the free traffic mode,the up peak traffic mode,the down peak traffic mode,the four-layer traffic mode,the two-layer traffic mode and the random layer traffic mode.The characteristics of each traffic mode,multi-car elevator system and the meaning of each performance index are analyzed.(4)The multi-objective optimization mathematical model is studied,and the average waiting time of passengers,average transit time,elevator energy consumption and passenger congestion degree are determined.The four indexes are weighted by weighting method and established comprehensive evaluation function of multi-car elevator control system.(5)Taking the traffic data of the actual elevator as an example,the data are pretreated by KPCA,Through the pretreatment analysis,the four-layer traffic mode in the six traffic modes of the multi-car elevator system can be ignored,and the processed data input to the trained SVM model to identify.In order to prove the feasibility of the algorithm,a single shaft multi car lift as the object,the proposed method is validated by data processing,feasibility and generality,and the experimental results show that the proposed method has higher recognition accuracy.(6)Aiming at the problem of multi-car elevator scheduling,a dynamic zoning method based on Fast R-CNN is proposed.The Fast R-CNN model is used to detect the number of people in the front of the car and in front of the car.According to the test results and the transport efficiency of each car to send a reasonable order,and the running area of the car is re-divided according to the dispatching result to achieve a reasonable scheduling.The simulation results show that the method is effective and can improve the running efficiency and flexibility of the multi car elevator.