Research On Theory And Method Of Hot Rolling Production Operation Optimization

In order to satisfy the requirements on shape and other quality indicators, the task of the operation optimization problem for hot rolling is to set the optimal values for the reduction and exit thickness of each stand. The study on hot rolling production operation optimization is of great significance for the fine production process of hot rolling.A mathematical programming model is built for the operation optimization problem of hot rolling which is derived from the actual production. This model is transformed into the transcendental geometric programming model. Then, by using proper variable transform, the transcendental geometric programming model is turned into a convex programming model which can be solved directly using convex programming software. For the actual production of hot rolling, a dybrid adaptive genetic algorithm is proposed to approximately solve the operation optimization problem. For the robust operation optimization of hot rolling, dynamic programming is developed to solve it. The main research results of this dissertation are as follows:(1) A mathematic model is built for the operation optimization problem of hot rolling which is derived from the actual production. It is a difficulty to get global optimization for the operation optimization problem that has highly nonlinear characteristic. Considering the structural characteristics of the mathematic model, firstly the intermediate variables are simplified, and then the operation optimization model is turned into a transcendental geometric programming model by using mathematical conversion.(2) The transcendental geometric programming model which is transformed from the operation optimization model of hot rolling cannot get the global optimization directly. Based on the model structural, this paper gives some theoretical analysis and mathematical transformation that the transcendental geometric programming model can help the operation optimization model of the hot rolling transformet into a convex programming model. Convex programming software can be used to slove the convex programming model(3) For the actual production of hot rolling, an operation optimization model is establishted with the objective of the quality of strip. Based on the actual hot rolling process requirements and production requirements, the model takes the exit thickness of each stand as the decision variable, and considers the effect of the ramping and temperature for the shape quality as the requirements.(4) For the operation optimization mathematic programming model of the actual production of hot rolling, a hybrid adaptive genetic algorithm is established to solve the model.In order to improve the robustness of HSaGA, a variety of different crossover operators is used, and a new mutation operator selection strategy is built to improve the convergence rate and maintain the diversity of the search population. The computation results show that HSaGA is superior to other the popular evolutionary algorithms through the benchmark function test. And the computation results are better than the results of experience rules for solving the operation optimization.(5) For actual production of hot rolling, the situation that the thickness of the waiting slab deviates from the presetting thichness or the parameters of the model are difficult to measure accurately often happens. A robust optimization model for the operation optimization proble of hot rollin is built. A dynamic programming method to solve the robust optimization model is proposed. The experimental results show that the algorithm is effective for the robust operation optimization.