Forecasting raw-water quality parameters for the North Saskatchewan River by neural network modeling

Abstract

In water treatment processes, raw-water colour is a key parameter for process control and monitoring. Therefore, the ability to predict the raw-water colour is desired to aid in the optimization of the treatment process. However, due to the high variance and the inherent non-linear relationship of the raw-water colour time series, it is difficult to produce a reliable model with conventional modeling approaches. In this paper, the artificial neural network (ANN) modeling technique is used to establish a model for forecasting the raw-water colouring in a large river. A general ANN modeling scheme is also recommended for the rest of the raw-water parameters. The modeling process typically includes four stages: source data analysis, system priming, system fine-tuning and model evaluation. Some optimization issues involved in the modeling phases and the potential applications of ANN in the water treatment industry are also discussed. Results indicate that the ANN modeling scheme shows much promise for water quality modeling and process control in water treatment.