Load and energy profile development using occupancy and interactions/activities in the residential building

Energy usage in households varies from unit to unit thus resulting in non-linearity in usage patterns at a particular time of use periods. This study presents Artificificial Neural Network approach based on characteristic variables such as activities and interactions taking place in the building, income level and occupancy presence which impact on energy usage and has the platform of addressing and solving the complexity phenomenon, non-linearity issues, the dynamic nature and randomness of data connected to energy usage The ANN-based model interpreted non-linearity connected to energy usage by learning historical patterns and computed its outputs in reference to connected characteristics variables. The performance of the fully-trained predictor is assessed by employing two performance indicators namely: coefficient of correlation (r) and root mean square percentage error (RMSPE). And the results showed good prediction accuracy during standard and peak periods for energy demand. This study is important for reducing energy consumption and contributing to peak demand thus utility operational expenditure and consumers, themselves will be free from high electricity costs.