Nonlinear Load Control Based On New Algorithm Of Chaos Hybrid Optimization

The control system of power plant is a complicated nolinear and multi-variable system. With the development of modern control theory, more and more new control methods have been applied in this control system. It is well known that control and decision-making problems can be regarded as the optimizations of structural and parameter. Therefore, developing a new optimization technology is the key step to solve the existing problems in load control system.Optimization algorithms always are involved in the course of solving engineering optimization problems. Due to the character of global optimization, chaotic optimization has become one of the forefront projects and important issues. Aiming to improve the traditional chaotic optimization, some typical map systems were studied. Further more, two new kinds of chaotic hybrid optimization algorithms were proposed and applied to nonlinear load control system. It is shown that chaotic optimization has great future in engineering applications. The main contributions of this thesis are listed as follows:1. Improvements in the optimization algorithm:Some new chaotic map systems were studied. It was showed that the algorithm based on Skew Tent map had better efficiency and accuracy than that based on Logistic through numerical analysis and simulation.Two new kinds of chaotic hybrid optimization algorithms were studied: Chaos and pattern search hybrid algorithm based on Skew Tent map; Chaos particle swarm optimization. The first one solved the problem of inadequate local searching capability. The last one solved the problem of multi-variable optimization in large space.2. Applications of optimization technology in the load control system:A new type of nonlinear predictive controller was presented combined with NN identification and chaos optimization for the control of nonlinear unit load control system. The receding horizon optimization of predictive control carried into execution depending on chaos and pattern search hybrid algorithm based on Skew Tent map. Good control indices were showed through simulation.Aiming at the nonlinear characters in the turbine regulating system and the problems of regular adjustment of the controller parameters, a novel RBF-PID control strategy based on chaos particle swarm optimization algorithm was proposed. Simulationg results showed that the control system performance is better than the conventional PID control. Futher more, it had ability of self-study and adaptability to the uncertainties.One method of controlling single-machine infinite-bus system was proposed based on neural network feedback linearization control theory. The network was trained by chaos particle swarm algorithm. The control was designed to aquire the value of turbine output torque which provided reference input value for the turbine system.