自回归神经网络在主汽温控制中的应用

Application of Self Recurrent Neural Network in Temperature Control of Boiler Main Steam

提出了自回归神经网络拓朴结构及其训练算法，把具有两个“自回归神经网络”的复合控制系统用于主汽温控制，一个作为辨识器，另一个作为控制器。仿真结果验证了该控制系统的优越性能，可知在过程控制领域中，神经网络在处理困难问题上，有良好的发展潜力。

The self recurrent neural network (SRNN) and its training algorithm are presented. A SRNN-based control system incorporating an identifier and a controller is applied to temperature control of boiler main steam. The good performance of the proposed control system is proved by simulation results. The neural network control approach is supposed to have the potential for handling difficult problems in process control systems.