污水处理过程的集合型监测方法研究

Research on Ensemble Monitoring Approach for Wastewater Treatment Process

针对具有非线性、多变量特性的污水处理过程监测准确性有待提高问题,提出一种集合型EMDIFCM-KPCA过程监测方法。该方法首先利用经验模态分解方法对数据进行预处理,然后用改进的模糊C均值聚类对数据进行聚类,最后采用核主成分分析方法对污水处理过程中的异常状态进行监测。水处理基准模型(BSM1)过程监测实验结果表明,将EMD-IFCM-KPCA集合型方法应用于污水处理过程,监测准确性优于传统KPCA,IFCM-KPCA等方法。

To improve the monitoring accuracy of the wastewater treatment process containing the nonlinear and multivariate characteristics,an ensemble monitoring method of EMD-IFCM-KPCA is presented.Firstly,empirical mode decomposition（EMD） method is used for data preprocessing.Secondly,the preprocessed data are clustered by the improved fuzzy C-means（IFCM） clustering.Finally,kernel principal component analysis（KPCA） method is used to monitor the wastwater treatment process.The monitoring experimental results of benchmark model（BSMI） of water treatment process show that the monitoring accuracy of EMD-IFCM-KPCA method is superior to the conventional KPCA,IFCM-KPCA methods in the process of the wastewater treatment.