Leveraging Census and Survey Data for Dynamic Simulation of Energy and Water Use in Households

This thesis presents an advanced agent-based model that accurately mimics household energy and water consumption patterns to enable future studies in the energy transition context. Leveraging data from census and national energy use surveys, the model generates a synthetic population that mirrors the statistical characteristics of real demographics. Each household within this population is assigned a specific set of appliances, informed by real-world usage data, to simulate daily and weekly consumption behaviours. The model captures the nuanced variations in energy and water use across different times of day and types of days, i.e. weekdays and weekends, at the level of the single appliance. The proposed approach surpasses existing modelling techniques by incorporating real, heterogeneous load profiles that reflect the detailed variability and specificity required for accurate simulation. It facilitates in-depth analysis across various temporal and spatial scales from individual households to broader communities. The model's validity and precision were rigorously tested by simulating an entire population and analysing the resulting behaviours. Comparisons were drawn against both the model that served as the foundation for this study and actual consumption data from Italy, showcasing the model's ability to generate meaningful, scalable, and precise forecasts of residential energy and water usage. This work lays the groundwork for informed policymaking and conservation efforts in energy and water usage.