Research On The Method Of Optimization Control About Large-Scale Industrial Process Based On Model And Data

With the increasing complexity of production technology and the increasing demand for production, optimization control of large-scale industrial process has become an important research topic. While conducting the optimization control of large-scale industrial process, due to environmental changes, fluctuations in production conditions and other reasons, the mathematical description of the actual process changes slowly. The accurate mathematical model for this reason is difficult to abtain. So there are discrepancies between mathematical model and mathematical description of the actual process. Fuzzy model is an effective method to deal with this problem. At the same time, how to effectively use a large number of process data and optimization control experience has important theoretical and practical significance on optimization control of large-scale industrial process.In this paper, according to the fuzzy optimization problem which the optimization control of large-scale industrial process based on fuzzy is involved in, the differential evolution algorithm with simplex operator is introduced. The method that combines differential evolution algorithm with simplex method is in order to give full play to a goog local search ability of simplex method and good global search capability of differential evolution algorithm, thereby improving the overall performance of the optimization algorithm. The validity of the proposed method is verified from the simulation research on the example of large-scale industrial process based on fuzzy model which uses interartion balance method, interartion prediction method and mixed method. The large-scale industrial process is to achieve the optimization control through the optimization control method above. In the actual industrial process, after a period of running and early accumulation, there is often a large number of data, which contains much optimization control experience. In order to effectively use the information of the actual production process, the combination of model and data is used to solve the optimization control problem of large-scale industrial process. On the basis of model-based optimization control method, the case base of large-scale industrial process is established according to the historical optimization control data. The target case can get the decision attribute values through case retrieving. While the case matching fails, the model-based method is used to realize optimization control, at the same time, the results are saved as a new case. If there is no most similar case in the case base, the cases those have lower similarities are searched out as the initial values of the model-based optimization method to quicken the optimization speed. In addition, the compensation case base that based on experts compensation experience is established to realize decision compensation, and the simulation research proves the effectiveness.