摘要
采用序批式污泥法处理污水的过程存在一定的非线性、时变性、随机性和不确定性,为此提出了一种基于核主元分析和小波神经网络模型的污水处理参数软测量技术。在保证水质信息量损失较小的情况下,使用核主元分析法对输入变量进行降维。将小波神经网络软测量模型和在线测量仪表相结合,对氧化还原电位、溶解氧、pH值及COD等参数控制信息进行实时检测;PLC控制器输出控制信号,控制整个系统设备的运行。仿真结果表明,和传统方法相比,该技术动态性能好、误差少,具有很好的鲁棒性和稳定性。
The sewage treatment process using sequencing batch sludge method features many disadvantages,such as nonlinearity,time variation,randomness,and uncertainty. Thus the soft measurement technology based on kernel principal component analysis and wavelet neural network model for sewage treatment parameters is proposed. To keep minimum loss for water quality information,the dimension of input variable is reduced by using kernel principal component analysis method. The wavelet neural network software measurement model is combined with online measuring instrument to detect parameter control information,including redox potential,dissolved oxygen,pH,COD,etc.,in real time; the operation of overall control system is controlled by the PLC controller. The result of simulation indicates that comparing with traditional method,this technology features better dynamic performance,smaller error,and possesses excellent robustness and stability.
出处
《自动化仪表》
CAS
北大核心
2014年第9期61-64,共4页
Process Automation Instrumentation
基金
国家自然科学基金资助项目(编号:51365010)
广西科技计划基金资助项目(编号:2013DB41017)
关键词
序批式污泥法
小波神经网络
核主元分析
PLC控制器
学习算法
Sequencing batch sludge method Wavelet neural network Kernel principal component analysis
PLC controller Learning algorithm
