As we all know coordinated control system (CCS) has large delay, strong coupling and its precise mathematical model is hard to obtain. So the regular control strategy is often difficult to get satisfactory control quality. Generalized predictive control (GPC) is a kind of advanced control strategy, it uses objective optimization, rolling horizon and feedback compensation, for these reasons the control effect is good when this algorithm is applied in CCS. However the GPC algorithm has large calculated burden and its structure is complex. Reference [6] put forward an improved generalized predictive control algorithm which avoids calculating inverse matrixes and complex Diophantine equation. Moreover this improved algorithm also has strong limit to the control action. In this paper this improved GPC algorithm is applied in 300MW one through boiler unit and the control effect is compared with GPC. The result shows that this algorithm reduces the on-line calculated burden greatly and saves much optimization time. Besides when the set point is changed the output signals and input signals don't vary largely. The overshoot is restrained so that it helps extend the life of the equipment.
[1]
Du Xiao.
Fast Algorithm of Constrained Model Predictive Control
,
2001
.
[2]
Xinghuo Yu,et al.
New Coordinated Control Design for Thermal-Power-Generation Units
,
2010,
IEEE Transactions on Industrial Electronics.
[3]
Donghai Li,et al.
Multivariable Nonlinear Control Design for Boiler-Turbine Units
,
2006,
2006 6th World Congress on Intelligent Control and Automation.
[4]
Pang Zhong.
A Fast Algorithm for Generalized Predictive Control
,
2005
.
[5]
David W. Clarke,et al.
Generalized predictive control - Part I. The basic algorithm
,
1987,
Autom..