High Resolution Residential Domestic Hot Water Consumption Profiles Using Data Mining Clustering Techniques on Time of Use Data

Domestic Hot Water (DHW) consumption profiles are essential to compute the associated energy usage. However, the estimation of DHW consumption is challenging when appropriate on-site measured data is not available. In this case, average daily DHW consumption estimations can be obtained as a function of the number of occupants. The simulation of high temporal resolution DHW consumption profiles is more challenging because these profiles are also strongly correlated to the daily activities of occupants. In the current paper, a bottom-up methodology to obtain DHW consumption at high temporal resolution is presented. The novelty introduced by the methodology is the correlation of DHW consumption profiles to occupant activities performed by population subgroups characterised by similar occupancy behavior. The population subgroups are identified using data mining clustering techniques. The accuracy of the model is higher than 98% when the average DHW volume consumption is compared with available data.