Recent research paper outlines how end-use-specific load profiles can be generated from smart-meter data

12.08.2015

The eLOAD model uses end-use-specific hourly load profiles to perform long-term load forecasts. A recent research paper demonstrates how such end-use-specific load profiles can be generated using load data records from smart meter surveys.

The paper combines half-hourly load records with survey data from 4,200 households participating in a representative smart meter project in Ireland to econometrically estimate end-use-specific load profiles, controlling for demographic and building characteristics. For each of nine “typical days”, representing combinations of different seasons and days of the week, we estimate 48 reduced form electricity demand equations. It was found that distinguishing load profiles between workdays and weekends proved essential, in particular for electric stoves, dishwashers and TVs. We calculated that the generated load profiles not only prove useful for long-term load projections, but also for a holistic understanding of the current system load curve. We were able to explain up to 40 % of the total system load by calibrating our estimated load profiles for five household end-uses with the actual load curve of Ireland in 2011. Our estimates indicate that lighting and electric heating account for about one third of the winter evening peak load in the Irish power system, but that these load profiles vary substantially between seasons. Based on our assessment, we conclude that energy efficiency policies should focus on lighting and thermal insulation in order to reduce the Irish system load peak. We find that energy-efficient lighting and a widespread technology switch from electric heating to heat pumps could reduce the winter evening peak load by 17 %. Finally, policies promoting load shifting should address electric heating before targeting other end uses like driers or dishwashers. The paper was presented at the 2015 eceee summer study.

Download paper here: Unravelling load patterns ofresidential end-uses from smart meter data