A pattern-based adaptive method for the analysis and prediction of time-series in sewage treatment

Abstract

Urban sewage treatment plants are characterized by a enormous energy consumption, but studies in this field show that a significant potential for reducing this consumption exists by using appropriate control and optimization concepts [1]. Therefore, a possible approach is the use of predictive methods. To apply predictive methods, a load forecast of the sewage treatment plants is necessary. In this paper we will present an approach to analyze and predict the loads for sewage treatment plants. Thereby we demonstrate that the times-series are strongly pattern-based; hence representative patterns will be used for prediction. Furthermore we will introduce an adaption algorithm to handle timevariances in the regarded signal. An ex-post evaluation of the results will conclude this paper.