基于改进算法的锅炉燃烧系统的经济性研究

The economy research of boiler combustion system based on the improved algorithm

针对锅炉燃烧系统的非线性、时变性和强耦合的特点,传统的控制方法的控制精度不高、自适应能力差等,提出了一种改进的模糊神经网络控制算法,对烟气含氧量进行控制。为克服常规算法的缺陷,将BP算法和粒子群PSO算法二者相结合,充分利用PSO算法的全局寻优能力和BP算法的局部搜索能力。另外引入了动态递归神经网络,对系统模型进行在线辨识,从而提高了网络的训练效率和控制器的控制效果,使系统达到经济燃烧。

For boiler combustion system of complex characteristics of nonlinear,time- varying and strong coupling,the traditional control method of control accuracy is not high,and poor adaptive ability,an improved fuzzy neural network control algorithm is proposed to control the flue gas oxygen content. To overcome the defects of conventional algorithm,combining the Back Propagation algorithm and particle swarm optimization,to make full use of the global optimization ability of PSO algorithm and BP algorithm＇s local search ability. In addition dynamic recurrent neural network is introduced to identify the system model online,which improve the training efficiency of the network and the control effect of the controller and make the system achieve economic combustion.