Research On Iterative Optimization Control Methods For Batch Process Considering The Model Error

The batch process which has the characteristic of little batch, many varieties, serialization, the complexity of the synthetic steps, technology-intensive and so on, is used more and more widely. The enterprise of the batch production has put forward to higher requirement to the process optimization control the optimization of the production automation and the efficiency and low consumption. So studying the problem of the optimal control of the batch process has important theoretical and practical significance.The strategy of the batch-to-batch iterative optimization control based on the mind of the iterative learn control can make full use of the running time terminability and the operation of repeatability to deal with the problem of the optimization control which has nonlinear and strong coupling with higher uncertainty. But due to the uncertainty of the process model and the existing of the kinds of interfere, there maybe bring barrier to the realizing of the batch process batch-to-batch iterative optimization control based on model. The paper mainly studies how to effectively address the optimization control problem of the batch process when the process model has error, especially big error. That is the batch process optimization control considering the influence of the model error. The paper proposes Boosting integrated training algorithm which is used to treat small sample to build batch process model. And on the basis of the batch-to-batch iterative optimization control, the paper proposes the strategy the batch-to-batch iterative optimization control based on the integrated training algorithm. The strategy introduces the compound model error grade to the learning law of the iterative control, and makes the algorithm of the batch-to-batch iterative optimization find the most advantages of the validity of workload model. Furthermore, it can improve the influence of the model error to the optimization control. In the process of the batch-to-batch iterative, the weighs adjust adaptively based on the changes of the actual value, in the situation of the big model error, the process realize iterative optimization control self-correcting. The method is valid through the simulation of the alcohol fermentation process optimization. Though the batch-to-batch iterative optimization control can improve the influence of the process model error to the optimization control effectively, it can’t do anything to the perturbation happening within batch. Bring the ideal of the batch-to-batch to the current batch, the simulation verifies the method of the batch-to-batch and within batch integrated optimization control is valid.