Research On Coordination Of Rolling Process Based On Multi-Agent

Hot rolling is a complex industrial process,which has the characteristics of strong parameter coupling,high nonlinearity and poor anti-interference ability.With the continuous improvement of the control requirements of hot rolling process,traditional control strategy has become more and more difficult to meet the requirements of high precision control,and it is urgent to support the new theory.Therefore,in this paper,in view of the above problems in 2 250 mm rolling line of a steel plant,the multi-intelligent technology theory and multistage mode matching method are applied to the optimization of parameters of hot rolling process,so as to quickly adapt the best operating parameters in the superior operational pattern library(rolling schedules)under the current working conditions,so as to improve the product quality and further enhance the competitiveness of the enterprise.Based on the deep research of relevant theories,combined with the characteristics of parameters,complexity of industrial process,optimization and control status of hot rolling line and actual production status of steel plant,this paper proposes a multi-agent control idea using mode matching method,including: 1)the heating furnace and hot rolling processes are described as agents;2)evaluation model of working condition index is established;3)multistep matching of parameters based on fuzzy c-means clustering method and transformation between Euclidean distance and similarity 4)input output prediction model and parameter optimization algorithm,etc.,to realize the coordination of all the hot rolling links.The main research contents and achievements of this paper are as follows:(1)In view of the heating furnace and hot rolling processes,they are abstracted as different single agents,and the data flow,material flow and information flow and flow direction are determined.The agent of different processes and different levels are described to complete the overall collaborative framework;(2)In view of the problem that traversal search may lead to time-consuming and resource consuming due to the large and uneven distribution of data on rolling line,based on the superior operational mode database established by expert experience and historical data,the operation method of multi-level mode matching of hot rolling agent is proposed under the framework of overall coordination.After the parameters are determined and the data preprocessing is completed,the first-step pattern matching is completed by fuzzy cmeans clustering method.The similarity measurement method based on Euclidean distance is used to determine the similarity between the current working condition and the elements in the primary matching subset,and the secondary pattern matching is completed.The method of multi-stage pattern matching can improve the speed and efficiency of pattern matching,which is of great significance to the hot rolling line with strong instantaneity and fast response;(3)In view of the poor data in the superior operational pattern library which may appear in the early rolling stage or special steel and special process,which leads to the failure of multi-step matching method to find the optimal operation mode,this paper combines fuzzy theory with neural network and establishes the input and output prediction model based on fuzzy neural network,and based on the model,particle swarm optimization intelligent algorithm is adopted until it is obtained and obtained The output of the optimal operating parameters satisfying the requirements,and at the same time,the new optimal mode is updated to the superior operational pattern library to complete the expansion of the library.The simulation results are compared with the actual production process of the steel plant,which proves the feasibility of the matching and evolution strategy,which can be used to guide the industrial production and have a positive effect on the overall operation of hot rolling.