Author(s): Hugo Jacque; Behzad Mozafari; Recep Dereli; Sarah Cotterill

Linked Author(s):

Keywords: Stochastic models Non-residential Time-series clustering Water-using products End-use

Abstract: Drinking water demand models are essential tools for researchers and practitioners in water quantity and quality modelling. However, there is a lack of models specifically designed for non-residential buildings, which significantly contribute to urban water consumption. This study extends the SIMDEUM model, a stochastic end-use model previously validated for offices, hotels, and nursing homes, to include educational and food service buildings. Key objectives include developing water use probability distribution functions (PDFs) and implementing end-use types specific to these building types. The PDFs are derived from analyses of hourly occupancy data, using time-series clustering to identify distinct occupancy patterns. New end-use types, such as urinals, ice machines, and food steamers, have been integrated into the model. To validate the extended SIMDEUM model, interviews and water audits were conducted in two university campus buildings, and smart water meter data was retrieved. Preliminary validation results show that the model accurately predicts daily water demand patterns for educational buildings. In conclusion, the extended SIMDEUM model appears promising to describe the water demand in non-residential buildings, thereby enhancing its applicability to a broader range of building types.

DOI:

Year: 2025