摘要
针对污水处理过程关键水质参数的测量问题,提出了一种基于改进混合粒子群算法的集成神经网络软测量方法.本方法利用出水关键水质参数(出水COD,出水BOD,出水TN,出水TP)之间的耦合关系,建立了基于集成前馈神经网络的软测量模型.同时利用改进的粒子群算法对集成神经网络进行训练,该训练算法具有较高的精度,能够建立有效的污水处理过程软测量模型.仿真结果表明,基于该算法的污水处理过程集成神经网络软测量模型能够准确地测量出水关键水质参数,测量精度较高.
A hierarchically neural network soft measurement method based on improved hybrid particle swarm opti- mization （IHPSO） is proposed for the key parameter measurement in a wastewater treatment process. Based on the correlationship among the key parameters, namely, chemical oxygen demand （COD）, biochemical ox- ygen demand （BOD）, total nitrogen （NT）, and total phosphorus （TP）, an soft measurement model of inter- gated feed-forward neural network is established. Meanwhile, a modified particle swarm optimization is ap- plied to training the new neural network, which has a higher precision and is more effective when a soft meas- urement model for the wastewater treatment process is constructed. The results of the simulation show that the integrated feed-forward neural network based on the IHPSO algorithm can accurately measure the key parame- ters in the wastewater treatment process with high measurement accurary.
出处
《信息与控制》
CSCD
北大核心
2014年第1期123-128,共6页
Information and Control
基金
国家自然科学基金资助项目(61034008
61203099)
北京市自然科学基金资助项目(4122006)
教育部博士点新教师基金资助项目(20121103120020)
关键词
污水处理过程
集成神经网络
改进粒子群算法
关键参数测量
wastewater treatment process
hierarchically neural network（HNN）
improved hybrid particleswarm optimization （IHPSO）
key parameter measurement
