Design of improved generalized predictive controller in the critical thermal process of 600MW subcritical unit

Intelligent control technology is the key to construct Infrastructure facility of Smart Grid, especially promoting high security and stability of power quality of 600MW subcritical main unit. Main steam pressure control system as a typical thermal process always affects power quality, which presents the dynamic characteristics of large time delay, integral inertia and parameter perturbation. Firstly, Controlled Auto-Regressive Integrated Moving Average (CADIMA) model based on colored noise (CN) with Hurwitz polynomials (HP) is proposed to establish accurately predictive model of practical application scene. Then improved generalized predictive control (GPC) algorithm is employing to design controller of main steam pressure and testify that the introduced CN with HP never impact on the stability of controller. Finally verify the validity of improved GPC through the simulation experiments consisted of two kinds of operating modes: nominal model and perturbation model, which shows more excellent at control performances, comparing to traditional proportion integration differentiation (PID).