Neural network based on-line shrinking horizon re-optimization of fed-batch processes

  • Authors:

  • Zhihua Xiong;Jie Zhang;Xiong Wang;Yongmao Xu

  • Affiliations:

  • Department of Automation, Tsinghua University, Beiijng, China;School of Chemical Engineering and Advanced Materials, University of Newcastle, Newcastle upon Tyne, UK;Department of Automation, Tsinghua University, Beiijng, China;Department of Automation, Tsinghua University, Beiijng, China

  • Venue:
  • ISNN'05 Proceedings of the Second international conference on Advances in Neural Networks - Volume Part III
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Neural network is used to model fed-batch processes from process operational data. Due to model-plant mismatches and unknown disturbances, the off-line calculated control policy based on the neural network models may no longer be optimal when applied to the actual process. Thus the control policy should be re-optimized. Based on the mid-batch process measurements, on-line shrinking horizon optimization is carried out for the remaining batch period. The iterative dynamic programming algorithm based on neural network models is developed to solve the on-line optimization problem. The proposed scheme is illustrated on a simulated fed-batch chemical reactor.