一种改进RBF网络与PID相协调的自适应控制

为充分利用人工神经网络处理非线性问题能力和PID控制器设计简单以及鲁棒性强的优点,设计了高性能控制器.通过对最小资源分配网络的裁剪策略进行改进,使改进的网络具有更加紧凑的结构;把改进的最小资源分配网络应用到神经网络自适应控制器中,同时结合神经网络自适应控制和PID控制,提出一种神经网络与PID相协调的自适应控制策略,并应用到对电站锅炉过热汽温的控制上.仿真结果表明,提出的控制算法与常规的控制算法相比具有更好的控制品质.

Self-adaptive control with coordinated PID and improved RBF network

A high performance controller is devised to make full use of neural network capacity in dealing with nonlinear problems by applying a PID controller's advantages of simple design and strong robustness.A PID neural network(PIDNN) control system consists of three hidden layers of proportional(P),integral(I),and derivative(D) neurons.A PIDNN's weights are adjusted by back-propagation algorithms.Through an improved pruning strategy for a minimal resource allocating network(MRAN),the resulting network is made more compact.The improved MRAN is applied to the neural adaptive control(NAC) strategy.Combining NAC with PID control,a self-adaptive control strategy is presented which coordinates the neural network and PID.This strategy was used to simulate control of the superheated steam temperature of a power station boiler.Results show the algorithm performs better than conventional control algorithms.