基于改进型GA优化FNNC的SG水位控制系统仿真研究

蒸汽发生器水位直接影响到整个核电站的安全及稳定运行,但蒸汽发生器本身由于所具有的高度复杂性、非线性性、时变性等特性,导致传统的串级PID控制等方法难以取得好的控制效果.本研究在串级控制的基础上,采用模糊神经网络来对蒸汽发生器水位进行控制,该控制算法能够充分发挥模糊控制及神经网络的优点.另外,为了减小模糊神经网络参数初值的选择对控制器的性能影响,将一种改进型遗传算法用于模糊神经网络控制器的参数优化.仿真结果表明,设计的控制方法无论是抗干扰能力还是在鲁棒性方面与传统的串级PID控制及常规的模糊神经网络控制相比较都有了很大的提高.

Simulation Study of a Steam Generator(SG) Water Level Control System Based on Improved Genetic Algorithm and Optimized Fuzzy Neural Network Control

The water level in a steam generator will directly influence the safe and stable operation of a whole nuclear power plant.However,the highly complexity,nonlinearity and time variation etc.characteristics of the steam generator itself cause the traditional cascade PID(proportional,integral and differential) control and other methods difficult to achieve a good control effectiveness.On the basis of the cascade control,the authors adopted the fuzzy neural network to control the water level of a steam generator.Such a control algorithm can give full play of the merits of the fuzzy control and neural network.In addition,to diminish the influence of the controller performance on the initial value selection of the fuzzy neural network,an improved genetic algorithm was used for parameter optimization of the fuzzy neural network controller.The simulation results show that both interference-resistant capacity and robustness of the control method thus designed are improved greatly when compared with those of the traditional cascade PID control and conventional fuzzy neural network control.