Development and verification of usage prediction service for intelligent smart water service

This paper proposes a real-time consumption prediction model that reflects the characteristics of consumers and time-series characteristics so that water supply can be effectively supplied by using a time-series prediction model. In order to reflect the characteristics of consumers, consumers were classified by industry, and each consumer's real-time usage was used to reflect time-series characteristics. There are two major water supply real-time usage prediction models: a short-term forecasting model that predicts daily usage and a long-term forecasting model that predicts weekly usage. Through such research and demonstration, it is expected that water supply and production plans based on predictions will be established to provide high-quality water supply and to enable efficient operation and management of water supply facilities.