Research On Intelligent Control System Of PC Component Production Line Based On Multi-Agent

Construction industrialization has become the mainstream of the development of the international construction industry.The industrialization of concrete prefabrication is an important part of the construction industrialization.At present,the level of intelligence and informatization of most domestic production lines of concrete prefabrication is not high.In response to the strategy of "Made in China 2025" and "Integration of Two Industries" promulgated by China,this paper takes the production line control system of precast concrete components(PC components)as the research object,and devotes itself to the intelligent control system of the production line of PC components based on Multi-Agent.Firstly,by investigating the research status of this kind of production line control system at home and abroad,multi-agent control theory is introduced,and the control index,control and collaboration requirements of intelligent control system for PC component production line are analyzed in depth.Solutions are designed,and the research contents and research methods of this paper are established.Secondly,the agent and multi-agent basic models of intelligent control system of PC component production line are established,including agent adaptive model,blackboard conversation model,bidding model based on contract network mechanism,etc.The simplified model of scheduling system of PC component production line is constructed,and the evaluation algorithms of production time,production energy consumption,production environment and production health are designed respectively.KNN(K-nearest neighbor algorithm)machine learning algorithm is used to evaluate the production and energy consumption,and variable weight comprehensive evaluation scheduling algorithm based on order stacking degree is studied.Thirdly,the hardware and software design of the key parts agent control system of PC component production line based on Multi-Agent is completed,including hardware selection,electric wiring diagram design of palletizer agent based on EPLAN P8,field bus communication network design,RFID hardware system design,etc.The data reading and writing program of the RFID system,intelligent kiln selection,intelligent positioning and intelligent interaction are also completed.The compilation of integrative palletizer agent control program and the verification of simulation experiment.Finally,the programming of KNN machine learning evaluation scheduling algorithm is completed,and more than 3000 actual production data of PC component production line of Jiangsu Tangchen Machinery Equipment Co.,Ltd.are used to test the KNN machine learning evaluation scheduling algorithm program based on production energy consumption evaluation,which verifies the accuracy of the algorithm program for energy consumption prediction;WINCC-Simulation simulation software is used to carry out variable weights.The simulation experiment of the scheduling algorithm is re-evaluated,and the functional relationship between the time-consuming weight coefficient,the energy-consuming weight coefficient and the order stacking degree is obtained.