| Good rolling procedures can reduce the energy consumption and improve the quality of strip in the production of aluminum strip. Based on an aluminum strip finishing mill group as the research object, the rolling force prediction and rolling schedule optimization problem has carried on the simulation research.Mechanism model is widely used in the analysis of the process of rolling, this paper studies deeply the main mathematical models, such as temperature, tension, rolling force. Field data is more and more easy to get and machine learning technology become more mature in recent years, regression calculation method has become one of the main methods of rolling force model. In this paper, the support vector machine model was trained based on the data obtained from field. Penalty parameters and kernel function parameters were optimized by fruit flies optimization algorithm in the process of training. By adding an location discriminant factor, fruit flies algorithm has the optimization ability of the negative coordinate axis, and adaptive length strategy and grouping is added to the search strategy. The rolling force prediction accuracy of the model is good, which can meet the needs of the production site.Multiple objective functions need to be considered in the rolling schedule optimization, in order to solve the multi-objective optimization problem, the difference vector and selection method of best particle is improved in the differential evolution algorithm, differential evolutionary algorithm and distributed estimation algorithm were grouped by switching factor, convergence and distribution has improved significantly relative to the original algorithm. Rolling schedule of an aluminum strip finishing mill group was optimized by the algorithm, approximate Pareto frontier with good convergence and distribution are derived, tradeoff between each target is fully considered and choose convenience is provided to the decision makers. |
PDF Full Text Request