PI control method based on virtual un-modeled dynamics compensation for a class of time-delay high-order linear systems

Abstract

In this paper, aiming at a class of time-delay high-order SISO linear systems in process industries, we propose a PI control strategy based on virtual un-modeled dynamics compensation, this method is based on the traditional PI control method, estimating system's virtual un-modeled dynamics by adopting data-driven control method, then design the compensator to eliminate the un-modeled dynamics' negative impact. The method combines the advantages of PI and data-driven control method, enlarges the PI control parameters scope for time-delay high-order systems, and improves the control system's dynamic and static performances. Finally, in order to verify the effectiveness of proposed method, a physical experiment of tank water level control is performed on the multi-function process control platform, and the experimental results indicate that this un-modeled dynamicscompensation based control strategy has higher control accuracy and practical value compared with traditional PI control strategy.