Residential Energy Consumption Modelling Using Socio-Economic and Geospatial Information

The increasing adoption of distributed energy resources (DERs), including photovoltaic (PV) systems, electric vehicles (EVs), heat pumps, battery storage systems, and night storage heating, is transforming residential energy consumption patterns and challenging the applicability of traditional standard load profiles for distribution grid studies. This thesis presents a data-driven framework for generating synthetic residential net load profiles by integrating synthetic household energy demand with technology-specific models for distributed generation and flexible energy loads using geospatial, socio-economic, meteorological, and measured grid data. The generated profiles are aggregated at the transformer and substation levels and validated against measured residual load data from three residential distribution grid areas in Germany. Model performance is evaluated by using quantitative metrics to assess distribution similarity, temporal characteristics, and the representation of extreme load conditions. The results demonstrate that the proposed framework successfully reproduces the temporal and seasonal characteristics of measured residual loads while capturing the combined impacts of distributed generation and residential energy technologies. Although some deviations remain during periods of high renewable generation and peak demand, the synthetic profiles show good agreement with measured data across areas without external grid influence, indicating the strength and transferability of the proposed methodology. The developed framework provides a scalable approach for residential energy modelling and supports distribution grid planning, flexibility assessments, and future energy system analyses in increasingly electrified residential environments.