Iterative learning control for batch processes based on support vector regression model with batchwise error feedback

An iterative learning control strategy based on support vector regression model with batchwise error feedback was applied for batch process control. Generally it is very difficult to acquire an accurate mechanistic model for a batch process. Therefore, empirical model based on process data attracts much attention. Because Support Vector Machine is powerful for the data-driven modeling problems characterized by small samples, nonlinearity, high dimension and local minima, support vector regression model was applied to the iterative learning control of batch process. Using the repetitive nature of the batch process, based on the model, iterative learning control can determine the optimal operating trajectory. To overcome influences of the modeling error and the disturbances, batchwise error feedback was incorporated into the computation of the optimal operating policy to guarantee the convergence of the iteration process. Simulation example demonstrated that the method can improve the process performance gradually as a batch process repeated, which made the product quality index converge to the desired value even model mismatches and disturbances exist.