Genetic-Algorithm-Based Adaptive Control of Superheat Steam Temperature on a Power Plant Boiler

Abstract

Superheat steam temperature control is critical to the normal and optimal operation of a power plant. Usually, cascade proportional integral derivative (PID) control system is introduced to regulate the superheat temperature with the PID parameters fixed constant. However, Due to the gains and time constants of the controlled plant change significantly with disturbances, the control effects of the conventional PID are not ideal. In this paper, a new adaptive control strategy is proposed to improve conventional superheat steam temperature control. Based on cascade PID control system, recursive least squares (RLS) method is used to identify the plant parameters and then genetic algorithm (GA) is applied to adjust the parameters of the PID controllers. MATLAB simulation investigation indicates that, compared to the parameter fixed PID system, GA-self-tuned adaptive control system is able to reject endogenous and exogenous disturbances more effectively and rapidly, thus has better self-adaptation capability and robustness.