Optimal Tracking Control of the Boiler-turbine System Based on Adaptive Dynamic Programming

To guarantee the efficient performance of the power plant, an adaptive tacking controller for the nonlinear boiler-turbine system based on offline policy iteration adaptive dynamic prorgamming (ADP) method is proposed in this paper. The optimal tracking controller is obtained through offline learning, which can maintain the characteristics of load changes in drum boiler-turbine type power plants. To implement the proposed method, neural networks (NNs) are used to construct the cost function and approximate optimal solution is achived. Then convergence of the method is analyzed. Simulation studies on the typical boiler-turbine system demonstrate that the proposed control strategy can achieve a satisfactory performance during a short period.