摘要
针对污水处理过程中关键水质参数无法在线监测的问题,提出基于PCA GABP神经网络的污水水质软测量方法。该方法由两部分组成:主元分析PCA和GABP神经网络。其中,GABP算法采用局部改进遗传算法优化神经网络权值,并采用自适应学习速率动量梯度下降算法对神经网络进行训练,建立软测量模型。仿真结果表明该软测量模型稳定性好、精度高,可用于污水处理厂对BOD进行在线预测。
To the problem that on-line information of some essential wastewater parameters is inaccessible in monitoring and controlling wastewater treatment processes. A soft-measuring technique applied to wastewater quality measurement is put forward based on PCA genetic neural network. It is composed of two elements: principle components analysis (PCA), and genetic neural network. This model can be applied to on-line predict wastewater BOD. Neural network is trained by improved BP algorithm, moreover, applying genetic algorithm to optimize the weights. The simulation results show that the soft-measuring model has good stability and high precision and can be applied to on-line predict wastewater BOD.
出处
《控制工程》
CSCD
2004年第3期212-215,共4页
Control Engineering of China
基金
国家自然科学基金资助项目(50274003和60304012)
北京市科技新星计划资助项目(H020821210120)