Study On The Expert System Of Abnormal Data Diagnosis For BAF Process In Municipal Wastewater Treatment Plant

Municipal Wastewater Treatment Plant(MWWTP) is used to making thedomestic sewage and industrial wastewater to discharge to reach standard, throughthe sewage treatment system, then protecting the natural water. The monitoring andmanagement to the operation of MWWTP is the key to guarantee it in normaloperation. In china, the responsible department of sewage treatment industry haveestablished the monitoring system and management assessment measures of themunicipal drainage and sewage treatment system, due to the lack of a clearquantitative analysis and intelligent judging system of the water quality andquantity, which influenced the ability of judging and anlysising of the governmentdepartment when MWWTP in problem. Therefore, it will be an urgent problem atpresent that how to make the water quality monitoring and fault diagnosistechnology of the municipal wastewater treatment plants scientific, automation andintelligence.This paper regards Biological Aerated Filter (BAF) water treatment process asthe research object, studying the establishment and implementation of the expertsystem of abnormal data diagnosis for BAF process in MWWTP. According to thestatistic characteristics of the monitoring data, this paper uses the hierarchicalclustering analysis to classify the data, and then explain the results through thewater temperature; Uses the [0,1] normalization method to do the data pretreatment;Based on support vector machine (SVM) theory, to construct the abnormal datadetection model of the urban sewage treatment plants; Using the Grid SearchMethod and the Particle Swarm Optimization Method respectively to select theoptimal parameters and do the analysis and comparison, finally test the model.Based on the knowledge of the Expert System Theory, this paper establishedthe BAF process fault diagnosis expert system. Through the fault tree clear up andorganize the water treatment technology knowledge and expert experiencediagnosis knowledge, and adopt the production rules to express the diagnosisknowledge, thus to build up a database of expert system; The breadth first, andpositive inference control strategy, and brings Fuzzy Theory in expert system, bydetermining variable membership functions and the belief propagation of credibility,construct the expert system inference machine.This paper, do the functional analysis,module design and system data flow-oriented design of the municipal wastewater treatment plants BAF abnormal datadiagnosis expert system,, develop in the workbench of c#, and applicates the practical cases. The results show that the system can correctly judge the waterabnormal data of sewage treatment plant, and analysis several main reason causesof the fault based on it, finally give out the corresponding solutions. This systemcan guarantee the stable and efficient operation of the urban drainage and sewagetreatment system, can significantly improve the urban drainage and wastewatertreatment plant efficiency, and provide powerful technology guarantee and technicalsupport for a government regulation.