Section: New Results

Electricity load forecasting and clustering

Participants : Jean-Michel Poggi, Benjamin Auder, Benjamin Goehry.

B. Auder, J-M. Poggi (with J. Cugliari, Y. Goude) are interested in hierarchical time-series for bottom-up forecasting. The idea is to disaggregate the signal in such a way that the sum of disaggregated forecasts improves the direct prediction. The 3-steps strategy defines numerous super-consumers by curve clustering, builds a hierarchy of partitions and selects the best one minimizing a forecast criterion. Using a nonparametric model to handle forecasting, and wavelets to define various notions of similarity between load curves, this disaggregation strategy applied to French individual consumers leads to a gain of 16% in forecast accuracy. Then the upscaling capacity of this strategy facing massive data is explored and different proposals using R are experimented. The proposed solutions to make the algorithm scalable combines data storage, parallel computing and double clustering step to define the super-consumers. This has been published in the journal Energies.

Benjamin Goehry is completing a thesis co-supervised by P. Massart and J-M. Poggi, aiming at extending this scheme by introducing the use of random forests as time series forecasting models adapted to each cluster.

J.-M. Poggi (with J. Cugliari) published in Wiley StatsRef-Statistics Reference Online, a paper entitled Electricity demand forecasting. the focus is on short-term demand forecasting at some aggregate level (e.g., zone or nationwide demands) from data with at least hourly sampled data. The main salient features of the load curve are first highlighted. Some of the common covariates used in the prediction task are also discussed. Then, some basic or now classical methodological approaches for electricity demand forecasting are detailed.