Batch-to-batch optimization for economic performance improvement in batch processes by rational disturbances utilization

Batch processes usually adopt a two-layer architecture control strategy to reduce the calculation cost while ensuring economic performance. The upper layer in the strategy obtains the nominal optimal reference trajectory by solving an economic optimization problem, while the lower MPC one possesses the task of tracking this trajectory as closely as possible. The disadvantage of this two-layer strategy lies in the lower layer resisting all the disturbance for satisfactory tracking results, regardless of the fact that part of the disturbance will be beneficial to economic performance. In order to realize the economic promotion in batch processes, a batch-to-batch optimization method is proposed in this paper. As an alternative to the traditional lower-layer tracking MPC, the proposed algorithm uses Gaussian process regression to establish a more accurate linear time-varying model and learns the information from the previous batches to obtain a more precise state estimation, thus continuously moving up economics by solving the economic optimization problem. Convergence of the proposed algorithm is proved and simulation is performed to verify the effectiveness of the proposed algorithm on economic performance improvement in comparison with the traditional tracking one.(c) 2022 Institution of Chemical Engineers. Published by Elsevier Ltd. All rights reserved.