摘要
针对污水处理过程这一多变量、强耦合的复杂非线性系统,提出了一种基于差分进化算法的模糊神经网络控制方法,并应用于污水处理过程溶解氧浓度的控制。首先利用差分进化(DE)结合BP的混合算法对给定的模糊神经网络控制器结构参数进行离线优化,然后利用BP算法较强的局部搜索能力,对参数进一步在线调整。将所提出的控制器用于污水处理BSM1仿真平台的溶解氧浓度控制,控制性能优于常规的模糊控制器,仿真结果表明了该控制策略的有效性。
Wastewater Treatment Process （WWTP） is a multivariable and strong coupled nonlinear system. This paper proposes a fuzzy neural network controller based on Differential Evolution （DE） and Back Propagation （BP） algorithm, which is used to control the Dissolved Oxygen （DO） in WWTP. The fuzzy neural network is firstly off-line optimized by DE and BP algorithm, and then is further on-line adjusted by means of the local searching ability of BP algorithm. The proposed controller is used to control the DO of a benchmark WWTP-BSM1 （Bechmark Simulation Model No. 1）. The experiment results show that the present algorithm can attain better performance than the conventional fuzzy controller.
出处
《华东理工大学学报（自然科学版）》
CAS
CSCD
北大核心
2013年第1期55-60,共6页
Journal of East China University of Science and Technology
基金
国家自然科学基金(60974066)
