bENBIS: Energy Demand Forecasting Workshop
During this workshop participants will be introduced to a wide variety of methods that can be used for forecasting energy demand. It will cover methodological as well as practical aspects that are relevant for a broad audience from academia and industry. The workshop will start with a broad overview of forecasting methods for nonstationary time series, and will apply those in the field of energy demand. After this keynote lecture, three dedicated talks will guide the participants through several initiatives and case studies, with attention to practical constraints and the (dis-)advantages of the methods used.
- 13h00 - 13h30: Welcome drink
- 13h30 - 15h00: Speaker 1 - J.M. Poggi (Univ. Paris Descartes & Univ. Paris-Sud Orsay, LMO)
- 15h00 - 15h30: Coffee break
- 15h30 - 16h00: Speaker 2 - G. Deconinck (KU Leuven, EnergyVille)
- 16h00 - 16h30: Speaker 3 - D. Metten (EDF Luminus)
- 16h30 - 17h00: Speaker 4 - C. Ritter (Ritter and Danielson Consulting & Université Catholique de Louvain)
Date: January 18, 2018
Location: Kasteelpark Arenberg 1, 3001 Heverlee, aula 01.07 map photo
Public car park: P.01 Celestijnenlaan 3001 Heverlee (a code will be send after registration) info
More info: Bart De Ketelaere, E. Buyse
Registration is free, but mandatory.
You can submit your registration here!
Titles & Abstracts:
Jean-Michel Poggi (Univ. Paris Descartes and Univ. Paris-Sud Orsay, LMO).
The talk starts with the industrial motivation of this work about nonparametric forecasting of electricity demand from the point of view the context this collaboration between academia and EDF and presenting methods previously considered in this context. Of course, the methods presented in this talk are quite general but the given application context allows to better highlight pros and cons. We then present two methods for detecting patterns and clusters in high dimensional time-dependent functional data. Our methods are based on wavelet-based similarity measures, since wavelets are well suited for identifying highly discriminant time-scale features. The multiresolution aspect of the wavelet transform provides a time-scale decomposition of the signals allowing to visualize and to cluster the functional data into homogeneous groups. For each input function, through its empirical orthogonal wavelet transform the first method uses the distribution of energy across scales to generate a representation that can be sufficient to make the signals well distinguishable. Our new similarity measure combined with a feature selection technique is then used within classical clustering algorithms to effectively differentiate among high dimensional populations. The second method uses similarity measures between the whole time-scale representations that are based on wavelet-coherence tools. The clustering is then performed using a k-centroid algorithm starting from these similarities. Finally, the practical performance of these methods is illustrated through the daily profiles of the French electricity power demand involved in nonparametric forecasting as well as individual consumers clustering involved in the forecasting by disaggregation of the electricity consumption.
Geert Deconinck (Afdeling ESAT - ELECTA, Elektrische Energie en Computerarchitecturen, KU Leuven)
Smart grids require lots of data to characterise the flexibility in electricity demand. Demand response (e.g. shifting appliance use) allows to integrate more renewables into the grid, or to have lower energy costs, without impacting user comfort. Within EnergyVille, KU Leuven's research centre on sustainable energy for an urban environment, a cloud-based data platform has been developed to capture and process such data - together with data from energy markets and users - in order to use it as a basis for scalable demand response applications. This presentation will shed a light on such applications and on the underlying platform.
David Metten (EDF Luminus)
Production of solar panels on the roofs of residential customers has a significant impact on the energy consumption profiles of these customers. Presentation will give a view on the problems we faced, analysis we’ve done and solutions we came up with to challenge this profile behavior.
Christian Ritter (Ritter and Danielson Consulting and Université Catholique de Louvain)
Everybody consumes energy in their own way; yet, on average there are just a few characteristic patterns. This talk is about these patterns, what they look like and why, and about what some of this means for the Belgian energy markets. It is an entry level presentation focusing on graphics.