USING NEURAL NETWORK MODELS TO PREDICT WASTEWATER FLOWS

Wastewater treatment plants usually have a satisfactory performance under normal dry weather flow conditions, however, hydraulic load variations constitute a large portion of the operating life of treatment facilities, especially those located in older cities that have all or partial combined sewer systems, and most of the observed problems in complying with permit requirements occur during those load transients that result in upsets to the different physical and biological processes at the treatment facility. The ability to predict the hydraulic load to a treatment facility is very beneficial for the proper operation of the facility during storm events. Most of the hydrologic and hydraulic models describing sewage collection systems utilize the mechanistic modeling approach that require detailed knowledge of the system and usually rely on a large number of parameters, some of which are uncertain or difficult to determine. Another modeling approach that has proven to be predictive and adaptive in engineering applications is Artificial Neural Networks (ANNs). Presented in this paper, an ANN model that is used to make short-term predictions of wastewater inflow rate that enters the Gold Bar Wastewater Treatment Plant (GBWWTP), the largest plant in the Edmonton area (Alberta, Canada). The potential of using the model as part of a real-time process control system is also discussed.