摘要：Current Optimization and Model Predictive Control practices for batch processes are implemented using two models, one for determining the optimal trajectories and another identified around those trajectories for control purposes. Here we use the recently developed Dynamic Response Surface Modeling methodology from which the optimal trajectories and the local linear or nonlinear state-space models for control purposes are obtained. Because concentration measurements at each batch run are very infrequent, this might be the most attractive way to obtain a dynamic model for control purposes.
