Research On Control System For Twin Elevator Based On Particle Swarm Optimization

With the rapid economic development and the improvement of city’smodernization level, more and more high-rise and super high-rise buildings standaround us as the urban landmarks, which leads that people require a higher servicequality of an elevator that is the indispensable vertical vehicle in a high-rise building.In recent years, the birth of a new member of elevator industry-twin elevator hascaused that experts and scholars in elevator industry and elevator manufacturerspaid close attention to it. By virtue of its structural features, twin elevator greatlyreduces the floor space of elevator wells in high-rise and super high-rise buildingsand improves the economical value for buildings. Furthermore, compared withtraditional elevator, twin elevator possesses a higher transportation capability. It hasbeen becoming the research emphases and difficulties of domestic and overseaselevator industry that how the more efficient optimal dispatching to the twinelevator system can be achieved, the elevator service quality and operatingefficiency can be enhanced and the overall energy consumption can be decreased.This paper applies the intelligent control algorithm to the control research on twinelevator group control system, which can lay a solid theoretical foundation forrealizing the practical engineering dispatching of twin elevator in the future.On the basis of the traffic flow data collected artificially in site, this workraises wavelet neural network traffic flow prediction method based on geneticalgorithm and multi-class traffic mode recognition method based on GA-SVM,which make elevator group control system achieve optimized dispatching betteraccording to traffic flow variation and traffic mode. Aiming at the lack of theexisting collection method for traffic flow data, this subject designs a device forcollecting the waiting crowd quantity by use of infrared body detectors withexperimental test accomplished. Test result reveals that this device can realize thedata collection for the waiting flow in each floor, which provide important real-timeparameters for finishing the optimal dispatching of elevator group control system.In the light of the structural features of twin elevator, this paper designs theoperation rules suited to twin elevator and confirms the optimal control targets oftwin elevator system. By means of the design and simulation on traffic flowgeneration module, the traffic flow data in different modes is obtained. The globalcontrol algorithm for twin elevator group is designed by use of particle swarmoptimization. In addition, in the Matlab programming environment the minimumwaiting time algorithm, adaptive mutation PSO algorithm and mixed PSO algorithmare compared by simulation. Simulation results show that two improved PSO algorithm possess a better control effect in the average waiting time, the incidenceof long waiting time and stopping number and the effectiveness of the algorithm isverified. The optimal dispatching to twin elevator group control system in differenttraffic modes is achieved.This paper makes an elementary theoretical research on twin elevator groupcontrol system and its control methods, which can provide a new idea and methodfor elevator group control technology, satisfy people’s growing needs to theperformance and comfort level of elevator, respond to the call of buildingenergy-efficiency policy energetically advocated in our country and also lay a goodtheoretical foundation for the research work and engineering application of twinelevator in the future.