压水堆核电站稳压器压力控制系统仿真研究

Simulation of Pressurizer Pressure Control System of Pressurized Water Reactor Nuclear Power Station

研究PID控制器参数优化问题,针对稳压器压力控制系统具有复杂非线性、时变性特点,引起系统的输出品质特性较差,超调量大,调节时间长,上升时间长,控制精度差等。传统PID的控制参数难以精确整定,且依赖于对象的精确数学模型。为了提高PID控制精度,减小超调量、调节时间和上升时间,提出用单神经元的神经网络来优化PID控制器参数的方法。通过单神经元的自学习和自适应能力,获得最优控制性能的PID控制参数。仿真结果表明,单神经元神经网络的PID控制方法与传统的PID控制方法相比,系统响应速度更快,超调量更小,为优化控制系统提供了参考。

Study PID controller parameters optimization problem. The pressure control system of pressurizer has the characteristics of complex nonlinear and time-varying, leading to the poor outputs of the system, such as large overshoot, long setting time and low control accuracy. It is difficult to get precise parameters with traditional PID controller, and the PID control method is relied on the precise mathematical model badly. In order to improve the precision of PID control, decrease the overshoot and the setting time, and the rising time, a PID controller parameter optimization method was put forward based on single neuron neural network. Through the self-learning and the self-adap- tive ability of the single neuron, the optimal PID controller parameters were obtained. The computer simulation exper- iment demonstrates that the single neuron PID controller performs very well： the response is quicke ant the overshoot is minimal compared with the tradition PID regulator. And it provides some reference for optimization control system.