A Feedback-based Optimization Method for Uncertain Batch Processes

摘要

Optimal control of batch processes is often implemented in an open-loop manner with an online optimizer. In this paper, a feedback-based optimization method is proposed for batch processes which suffer from parametric uncertainties. Firstly, the optimality conditions and the expression for optimal input are derived based on the uncertain parametric model, then the output measurements and corresponding optimal inputs are collected via off-line simulation. Then, the explicit control law is obtained through regression, which is implemented in a feedback manner for optimization purpose. Case study on a fed-batch reactor indicates that the proposed approach can attain good optimizing performance in a wide range of uncertainties.