BP神经网络在锅炉汽包水位自动调节系统中的应用

Application of BP neural network for the automatic control system of water level in boiler barrel

传统的工业锅炉汽包水位自动调节系统一般采用PID控制器,但在虚假水位问题中采用PID控制器有误差难以调节和自适应性差的特点.尝试在该问题中引入BP神经网络模型,改进了PID控制器的上述问题,系统的研究了虚假水位产生的原因以及隐层数、数据间隔和数据量对BP网络控制效果的影响,并有望将其发展成为解决锅炉汽包水位问题的一个更普遍而且具有自适应特征的方法.

Generally, a traditional system uses the PID controller for automatically governing the water level of boiler barrels. However, this method has two main problems: difficulties in adjusting the errors and the poor auto-adaptation.Thus, the target of this paper is to solve the two problems with the application of a new model of BP neural network. The result of this study justifies the feasibility of BP neural network by using the simulation of Matlab. Meanwhile, this paper also systematically analyses the influences of hidden layer, interval value and data number on this new method to solve more industrial problems related to the water level of boiler barrels.