The Research On The Fault Disgnosis Methods For The Wastewater Treatment

Along with the increase of the amount of wastewater,the scale of the wastewater treatment process is increasing.It is inevitable to have a fault in wastewater treatment process,and the fault type is diversified.So it is very necessary to make an accurate and effective fault detection and diagnosis for the whole wastewater treatment process.Due to the previous fault detection methods of the wastewater treatment process are based on the single model process,the fault detection methods for multi-model process are rarely used in the wastewater treatment process.Thus,a kind of an ensemble fault detection method is proposed in this thesis,namely the EMD-IFCM-KPCA method.In order to solve the problem of rapidity,an ensemble fault diagnosis method is also proposed in this thesis,namely the IJITL-IEMD-RLSSVM method.Both of methods are applied to fault detection and diagnosis for wastewater treatment process.Experimental simulation validates the effectiveness of the proposed methods.The ensemble EMD-IFCM-KPCA fault detection method is proposed.Firstly,Empirical Mode Decomposition(EMD)is used to preprocess data.Secondly,data are clustered by the Improved Fuzzy C-means cluster(IFCM)method.Lastly,Kernel Principal Component Analysis(KPCA)and T2,SPE control chart are used to conduct fault detection.The ensemble IJITL-IEMD-RLSSVM fault diagnosis method is also proposed.Firstly,the history data is classified by Improved Just in Time Learing(IJITL)method.And the history model data is looked for and output which is similar to the diagnosis data.Secondly,Improved Empirical Mode Decomposition(IEMD)method is used to preprocess the diagnosis data and output history model data.After that the Recurrence Least Squares Support Vector Machine(RLSSVM)method is used to classifiy the fault.Experimental simulation of the wastewater treatment process is conducted to validate the efficiency of the proposed methods.Firstly,the multi-model ensemble EMD-IFCM-KPCA fault detection method is compared with the other two methods which named the KPCA method and IFCM-KPCA method to verify the validity and the accuracy of the proposed method.Secondly,the multi-modal ensemble IJITL-IEMD-RLSSVM fault diagnosis method is compared with the other two methods which named IJITL-SVM method and IJITL-IEMD-SVM method to verify the rapidity and accuracy of the proposed method.Theoretical research and experimental results show that the ensemble EMD-IFCM-KPCA fault detection method and the ensemble IJITL-IEMD-RLLSVM fault diagnosis method have an effective result in fault detection and diagnosis for the wastewater treatment process.What's more,the proposed methods are superior to the traditional methods in accuracy and rapidity.