DIAGNOSIS SYSTEM FOR BIOLOGICAL WASTEWATER TREATMENT PROCESS USING BAYESIAN NETWORKS

A system for diagnosing a wastewater treatment process using a Bayesian network is provided to operate a wastewater treatment facility continuously and stably by training the Bayesian network with data accumulated in each processing facility, and predicting and diagnosing a water quality of effluent with the trained Bayesian network. A measurer(100) collects/transmits daily effluent data output by analyzing a sample of each effluent in everyday and real-time measurement of a sensor installed each reacting tank. A database(200) stores the daily effluent data and the real-time measurement, and stores the predicted effluent data inferred by inputting the real-time measurement to a Bayesian network when the daily effluent data is not generated. An effluent water quality determiner/predictor(310,320) determines abnormality based on a legal reference value and change of composition of the effluent while determining the abnormality of the composition of the effluent by receiving the daily or expected effluent data from the database. A process diagnosis unit(400) diagnoses a current state of each reaction tank and the abnormality of a water treatment process by inputting abnormality determination and the real-time measurement data to the Bayesian network(10).