Modeling And Optimization Of Finish Rolling Schedule

Based on the technics models and producing requirements of strip steel rolling,this paper researches intelligent prediction models for crown and establishes optimization calculation model of finish rolling schedule utilizing multi-objective genetic algorithm.The purpose of this optimization model is to reduce the cost, increase product output and raise product quality. The decision-making variables and objective variables for optimization are decided according to technics mechanism and practical producing conditions. It is under the consideration of interactive relationships among various factors such as strip thickness, strip width, temperature, strip profile, efficiency and cost to realize the system's optimization. Based on lots of practical data the regression prediction models are established separately using artificial neural network and support vector machine and are applied in the optimization model of finish rolling schedule to predict crown. The optimization model adopts genetic algorithm (GA) as the basic optimizing method and converts the multi-objective problem to a single objective problem using the method of weighting aggregation. To speed up the optimizing process, several effective measures are taken to improve the performance of GA. The proper weights of multi-objective are fixed through many simulation experiments.The simulation software platform is developed by VC++.NET. Via the simulation based on practical data, the results show that the optimization model is feasible and available.