Simulation of the mean performance of municipal waste treatment plants

Abstract

Studies of the Kitchener and Humber, Ontario, municipal waste treatment plants suggest that it is possible to represent the mean performance of a treatment plant by simulation using steady state models, despite the transient operation of such plants and data deficiencies. A conventional model for the activated sludge unit and a recently developed settler model furnished a satisfactory simulation of the Kitchener Plant provided the model parameters were evaluated using plant data. Good agreement was obtained between predicted and measured mean performance in 1966 and 1967 employing parameters obtained from 1968 data. However, only simulation of means for periods longer than 4 weeks seems reliable. A group of models developed at the Robert A. Taft Center for Water Pollution Research in Cincinnati, Ohio, were found to be unsatisfactory for the simulation of the Kitchener and Humber plants, probably because of inappropriate model parameters. It was found that parameters which gave good results in a model for the Kitchener primary settler were unsatisfactory when used with the model for the same unit in the Humber plant. Thus, it seems that parameters in simulation models must be drawn from data obtained for the plant being simulated.Variations in the raw sewage flowing to the Kitchener plant are significant and combine with plant operating practices to cause the plant to function in an unsteady state. This aspect has been neglected in past research on treatment plants and now warrants serious study.