Anders Reichert Liisberg: With the increasing installation of smart meters for measuring electricity consumption and with the possibility to access these data [Engrosmodellen], there is a huge potential to give energy advice directly to the consumers based on in depth analysis of their own data.
In cooperation with SEAS-NVE the focus in this project is to develop models describing the underlying occupant behaviour based on time series of electricity consumption with hourly time resolution. This is used for direct energy advice to residences and small and medium sized enterprises (SME's).
For this purpose the research activities will include:
- Cluster analysis and classification of consumer patterns, both for residences and SME's.
- Building disaggregating models to extract hidden knowledge describing the underlying occupant behaviour, from the electricity consumption.
- To investigate which models (parametric, semi-parametric or non-parametric models) have the best forecasting performance based on consumer patterns.
- To describe the characteristics of the electricity consumption by taking into account the parameters development over time.
- To investigate if the advice actually have the desired effect in minimizing the electricity consumption.
PhD project title: Datadriven models for energy advicing leading to behavioural changes in SMEs and residences
Effective start/end date 15/05/2016 → 14/05/2019
Supervisors: Jan Kloppenborg Møller, Peder Bacher, Henrik Madsen, Section for Dynamical System