Abstract: In addressing the nonlinear and uncertain characteristics of hydroelectric unit regulation systems (HURS), this paper presents a new design of a nonlinear robust H2/H∞ control strategy based on Particle Swarm Optimization and Gravitational Search Algorithm (PSOGSA). Firstly, a refined comprehensive model of a HURS is developed, with the turbine model described by BP neural networks. Then, the output feedback linearization theory and linear matrix inequalities were used to design a nonlinear robust H2/H∞ controller to ensure the system’s good robust H2/H∞ performance against disturbances. A PSOGSA algorithm is adopted to optimize the parameters of the weighted functions of the controller so as to achieve its optimal performance. Simulation results indicate this PSOGSA-based nonlinear robust H2/H∞ control strategy has advantages in improving the HURS stability and capability against disturbances, providing an effective approach for practical HURS design.

Key words: hydro-generator, regulating system, H2/H∞ control, linear matrix inequality (LMI)