Research And Implementation For HotRolling Intelligent Scheduling Expert System In Anshan Iron&Steel 1700mm Line

The expert system is an important branch of artificial intelligence,is a system with abundant specialized knowledge and experience,through applying artificial intelligence technology and computer technology,it can imitate the reasoning process of human expert resolving particular problem.In recent years,expert system technology has gradually become mature,widely used in engineering,science,medicine,military and business,etc.,its achievements are quite substantial.Anshan Iron and Steel Group existing five hot-rolling lines all have relatively complete MES systems,they receive daily-plans which are issued by ERP,work out mill plans and send them to the process control systems,but lack mill unit-plan making link.The goal of hot-rolling intelligent production scheduling expert system goal is to work out unit-plan of hot-rolling process optimally,to maximize the weight of rolling within the roll change period and improve products quality.Based on Ansteel 1700mm hot-rolling lines,this paper design the theory and system,and consider thecommonality problems of all Hot rolled steel strip.In this paper,The author design,organize and implement already existing Ansteel heat rolling MES system,the mill-plan give priority to MMI mode,for the reason that the different level between not only different scheduling personnel but also the different status of the same scheduling personnel,it will cause greatly difference in quality and quantity even in the same product condition,this is the random problem and level fluctuation problem in manual scheduling.Based on years of analyzing the actual product performance dates and sample designing,the author has established 1700mm hot rolling intelligent scheduling expert system,and effectively combined with the existing information system,gradually modeling the mill unit-plan.In this paper,it take the total rolling weight of unit-plan as the maximum backpack capacity,it calculate the initial solution through using the knowledge-library of expert system,and judge the punishment of the process-solution through using rule-library of the expert system,the produced plan is taken as the sample of the expert system,it utilize artificial neural network technology to training the expert system to improve its running effect.Many times simulation experiments results show us that the conformance degree of heat rolling unit-plan sequence provided by expert system can reach more than 85%,on the basic of reducing scheduling personnel workload greatly,it also effectively avoid the random problem and the level fluctuation problem.