Study On The Expert System Of Fault Diagnosis For A~2/O Process In Municipal Wastewater Treatment Plant

Urban sewage treatment system is an important prerequisite to protect urbanwater environment safety, whose good or bad working condition will directly affectthe safety of urban water environment. So the monitoring and management forurban sewage treatment plants running is very necessary and critical. At present,operating regulation system for the urban sewage treatment plant has beenpreliminary formed, however, due to a lack of clear abnormal judging system andintelligent fault diagnosis system, there are often a lot of problems, such as reporteddata lacking scientific rigor, poor drainage monitoring ability, not timely faultdiagnosis and so on. These problems not only restricts the ability of managementto master plant’s operation condition and analysis the fault, but also hindered thedevelopment of urban sewage treatment.Therefore, from getting reliable data andstrengthening urban sewage treatment plant operation regulation capacityperspective, current problems to be urgently solved is how to realize scientific andintelligent discrimination of the abnormal data and fault diagnosis of the process.In view of man-made wrong operation, artificially inflate, low experimentalinstrument accuracy, the abnormal data caused by monitoring equipment failure andnot standard phenomenon of the city sewage treatment plant water in the operationprocess, according to the theory related to statistics, it can implement the abnormaldata detection of the urban sewage treatment plants, which can effectivelydiscriminate data whether they are from online monitoring devices or from theartificial reports and thenwipe off singular data in order to judge whether waterquality is standard or not. Meanwhile, it also can distinguish whether there was asignificant difference between artificial report and monitor data to verify whetherthere is any abnormal data of artificial report. When water quality exceeds bid,abnormal data detection module will serve as a pretreatment measurement toprovide service support for the subsequent fault diagnosis.The A²/O process in one sewage treatment plant is served as research objects.In the light of wastewater treatment system involved with complex biologicalreaction mechanisms and larger uncertainty, daily fault diagnosis of the improvedA²/O process was achieved based on bayesian network in municipal sewagetreatment plant.Comprehensively considering the impacts of influent quality, flowquantity, F/M, DO, ORP, SVI and C/N on effluent quality （such as SS, COD, NH₃-Nand TN）, a bayesian network modeling was constructed by field expert knowledgeand historical training data, which can be applied to real diagnosis process. The actual application results show that the bayesian network modeling is characteristicof precise diagnosis and high reliablity. It can be a new effective method to solvethe uncertainty problem of fault diagnosis process in current sewage treatment plant.Lastly, we build the A²/O process fault diagnosis expert system of the urban sewagetreatment plants based on bayesian network.In this paper, we introduced the requirements analysis, database design anddetailed system function design of the fault diagnosis expert system of the urbansewage treatment plants.Based on Java development platform, we use B/Sarchitecture and classic SSH framework to build the city sewage treatment plantfault diagnosis expert system application software, which is combined with acertain city sewage treatment plant with A²/O process. Results show that the systemcan find out the cause of the problem from the abnormal data and givecorresponding solutions during the process of reasoning.The implementation of thesystem will provide the supervision department with strong science and technologyguarantee and technical support. Besides, the system has important significance inimproving the level of urban sewage treatment plant operation and managementand make the operation of the sewage treatment stable, normal and efficiently.