Probabilistic Electric Load Forecasting Model for the Uruguayan Interconnected Electrical System

E. Cornalino and R. Chaer



The aim of this research is to improve the capacity to represent and forecast the electric demand for next week’s scheduling. Currently the demand forecast used for this purpose is deterministic, which is not representative of reality, even if an ideal temperature forecast was available. The current context of the Uruguayan electrical system has high probability of exportable surplus energy. For this reason, improvements to the procedure used to calculate systems supply costs and the quantity of exportable energy are welcome, in order to maximize the benefit we can get from resources. The methodology applied is based on previous developments for simulation of stochastic variables within the SimSEE platform [2]. It combines daily step CEGH model [3] with a k-means clustering
method [4].
Obtained results were satisfactory both from the point of view of the representation of the temporal behavior of the power demand, and from the point of view of the error obtained in the predictions.
What is more, this improvements helps to reduce risks involved when making energy commitments with neighbouring countries.

Published in: Renewable Energy & Power Quality Journal (RE&PQJ, Nº. 16)
Pages: 181-186 Date of Publication: 2018/04/20
ISSN: 2172-038X Date of Current Version:2018/03/23
REF: 255-18 Issue Date: April 2018
DOI:10.24084/repqj16.255 Publisher: EA4EPQ

Authors and affiliations

E. Cornalino(1) and R. Chaer(1-2)
1. Administración del Mercado Eléctrico (ADME), Uruguay
2. Instituto de Ingeniería Eléctrica, Facultad de Ingeniería, Universidad de la República, Uruguay

Key words

Probabilistic load forecasting, Simulation of stochastic variables, Dispersion, Decision at risk.


[1] T. Hong, S. Fan. “Probabilistic electric load forecasting_ A tutorial review”, International Journal of Forecasting 32 (2016) 914–938
[2] R. Chaer et al. Memoria Final Proyecto ANII-FSE2009-18 pág. 45 / 176
[3] W.R. Gilks, S. Richardson, D. Spiegelhalter. “Markov Chain Monte Carlo in Practice”. Chapman & Hall. 1996.
[4] R Chaer, “Fundamentos de modelo CEGH de procesos estocásticos multivariable”. Technical Report, IIE-Fing, Udelar -2011, Montevideo.
[5] A.K. Jain, “Data clustering 50 years beyond K-means”, in Pattern Recognition Letters 31 (2010) 651–666.
[6] Steven Stoft. “Power System Economics: Designing Markets for Electricity” IEEE Press & WILEY-INTERSCIENCE