Performance Assessment for Short-Term Water Demand Forecasting Models on Distinctive Water Uses in Korea

It is crucial to forecast the water demand accurately for supplying efficiently and stably in a supply system. In particular, forecasting short-term helps saving energy reducing operating costs. With introduction of Smart Water Grid (SWG) system, amount consumption obtained real-time through smart meter, which can be used demand. The models widely include Autoregressive Integrated Moving Average, Radial Basis Function-Artificial Neural Network, Quantitative Multi-Model Predictor Plus, Long Short-Term Memory. However, there lack research on assessing performance SWG demonstration plant. Therefore, this study, was forecasted each model using data collected from assessed. Research Group installed meter block 112 located YeongJong Island, Incheon, actual plant were adopted. assessed by Residual, Root Mean Square Error, Normalized Nash–Sutcliffe Efficiency, Pearson Correlation Coefficient as indices. As result forecasting, it difficult only time consumption. have limitations reflecting characteristics consumers, system managed more precisely if other factors (weather, customer behavior, etc.) influencing are applied.