Data-driven modelling of energy demand response behaviour based on a large-scale residential trial

Recent years have seen an increasing interest in Demand Response (DR), as a means to satisfy the growing flexibility needs of modern power grids. This increased is required due proportion intermittent renewable energy generation into mix, and complexity demand profiles from electrification transport networks. Currently, less than 2% global potential for demand-side currently utilised, but more widespread adoption residential consumers resources can lead substantially higher utilisation potential. In order achieve this target, acquiring better understanding how DR participants respond events essential – recent advances novel machine learning statistical AI provide promising tools address challenge. study provides in-depth analysis customers responded incentive-based DR, utilising household-related data large-scale, real-world trial: Smart Grid, City (SGSC) project. Using number different approaches, we model relationship between household’s response features. Moreover, examine effects households’ features on behaviour, highlight key insights which raise questions about reported level consumers’ engagement schemes, motivation customers’ level. Finally, explore temporal structure although found no supporting evidence responders over time available trial, proposed methodologies could be used longer-term longitudinal studies. Our concludes with broader discussion our findings paths future research emerging area.