Study on Terminal Sliding Mode Control for Boiler-Turbine Unit Based on RBF-NN Inverse System

Aiming at the features of boiler-turbine unit, such as multiple inputs, multiple outputs, nonlinearity and strong coupling etc., an inverse system method is used to achieve feedback linearization and decoupling of the unit, during which the inverse system is identified using radial basis function (RBF) neural network, while the modeling errors reduced through online learning. A terminal sliding mode controller is designed for the unit to reach finite time convergence, and its global stability is analyzed and guaranteed using Lyapunov approach. Simulation results show that the control system works well within a wide range of operation conditions, which is superior to the system designed based on classical inverse system method.