A Water Demand Forecasting Model using BPNN for Residential Building

This paper presents a residential water demand forecasting model using a back propagation neural network (BPNN) in the context of residential buildings in Korea. The water demand of a building demonstrates a highly complex and non-linear phenomenon reflecting such features as geographic and climatic and special types of buildings. We describe the impact of several potential determinant factors affecting water use in residential buildings in four different provinces in Korea. Empirical data sets consisting of water consumption retrieved from multiple residential buildings in Korea were evaluated to verify the performance evaluation. Our results show that the proposed model can successfully predict estimated outputs through the BPNN. The model we propose can used in decision making for the residential water management policy in Korea through the optimal estimation of residential water consumption.