Constrained Batch-to-Batch Optimal Control for Batch Process Based on Support Vector Regression Model with Batchwise Error Feedback

A batch-to-batch optimal control method is presented for batch processes under input and output constraints with batch wise error feedback. Generally it is very difficult to acquire an accurate mechanistic model for a batch process. Because support vector machine is powerful for the problems characterized by small samples, nonlinearity, high dimension and local minima, support vector regression model is developed for end-point optimal control of batch process. Because there exist model error and disturbances, an iterative (batch-to-batch) method is used to exploit the repetitive nature of batch processes to determine the optimal operating policy. To ensure the safe, smooth operations of batch process, certain constraints are taken into considered. Furthermore, batch wise error feedback is incorporated into the computation of the optimal operating policy to guarantee the convergence of the batch-to-batch optimal control. Numerical simulation shows that the method can improve the process performance through batch to batch under constraints.