Feasible constrained nonlinear predictive control on power plant

Thermal power plant is required to ensure a fast load change without violating thermal constraints. While model predictive control has been widely used in power plant, incorporating of constraints is a major problem. Two alternative methods of exploiting the nonlinear predictive control are described. One is the input-output feedback linearization technique. The other is the neuro-fuzzy networks(NFNs). Steam-boiler generation control using the two nonlinear predictive methods show satisfactory results and improved performance compared with conventional predictive method. Comparing results considering both the integral absolute value and the relative optimization time needed for completing the simulation have also been addressed in detail.