NSG水位神经自适应PID控制与仿真研究

核动力蒸汽发生器(NSG)是一个高度复杂的非线性时变系统。由于蒸汽发生器在瞬态、启动和低功率下的“收缩”与“膨胀”现象引起的逆动力学效应，使蒸汽发生器的水位控制变得复杂。本文针对传统的核动力蒸汽发生器水位PID控制方法存在的缺点，将神经网络方法与PID控制的结构结合起来，提出了核动力蒸汽发生器水位神经自适应PID控制方法。采用BP学习算法调整控制器神经网络的连接权值，实现了控制器参数的在线整定。仿真研究表明，所设计的控制器具有良好的控制性能，且结构简单，易于实现。

Research of Adaptive Neural Network PID Control and Simulation for Nuclear Steam Generator Water Level

The nuclear steam generator is a highly complex, non-linear, and time-vary system. Its water level control is complicated due to reverse thermal-dynamic effects known as "shrink" and "swell" effects, which occur during plant transients and are more prominent at start up and low power operation. The method of adaptive neural-PID control that combines neural network and PID control is put forward in allusion to shortcoming of conventional PID control for nuclear steam generator water level. The on-line adjusting of controller parameter is achieved by employing back propagation algorithm to update the weights of neural network. The performance of designed controller with simple structure and characteristic of implementing easily is demonstrated by computer simulation.