Análise dos fatores que influenciam o desempenho de um modelo de previsão de demanda de energia

Leonardo Santos Amaral; Gustavo Medeiros de Araújo; Ricardo Alexandre Reinaldo de Moraes

doi:10.47909/anis.978-9916-9760-3-6.111

Authors

Leonardo Santos Amaral Universidade Estadual de Montes Claros - UNIMONTES, Av. Prof. Rui Braga, S/N - Vila Mauriceia, Montes Claros - MG, Brasi
Gustavo Medeiros de Araújo Universidade Federal de Santa Catarina UFSC, R. Eng. Agronômico Andrei Cristian Ferreira, s/n -Trindade, Florianópolis - SC, Brasil
Ricardo Alexandre Reinaldo de Moraes Universidade Federal de Santa Catarina UFSC, R. Eng. Agronômico Andrei Cristian Ferreira, s/n - Trindade, Florianópolis - SC, Brasi

DOI:

https://doi.org/10.47909/anis.978-9916-9760-3-6.111

Keywords:

Machine Learning, Artificial neural network, DemandForecast

Abstract

The forecast of energy demand is one of the essential variables for the operation, planning, and estimation of tariffs for electric energy networks. There is currently a significant investment in modernizing the seven domains of an Intelligent Electric Grid (Smart-Grids). The operation of these various domains requires greater precision in forecasting demand for use and energy generation. Several techniques have been applied for this purpose, the most promising being machine learning, which in recent years sss received special attention. The focus of this work is to analyze a univariate demand forecasting method that employs deep artificial neural networks. The load time series history is used as input, added to the seasonality representation to predict the future load values of the electrical system. It will be shown that several factors influence the answer, and one of them is the choice of the way to decompose the problem and, therefore, the applied model.

Downloads

Download data is not yet available.

References

Buckland, M. K. (1991). Information as thing. Jornal of the American Society for Information Science, v45, n5, p351-360

Carvalho, J. F. (2014).Energia e sociedade. Estudos Avançados, v. 28, n. 82, p. 25-39.Disponível em: https://www.scielo.br/j/ea/a/GYqnmvDvtGtr8N5zdsYdj5j/?format=pdf&lang=ptAcessado em: 05 set de 2020.

Dudek, G. (2020).‘‘Multilayer perceptron for short-term load forecasting: from global to local approach,’’NeuralComput. Appl., vol. 32, no. 8,pp. 3695–3707, Apr.doi: https://doi.org/10.1007/s00521-019-04130-y

Fallah, S. N. Ganjkhani, M. Shamshirband, S. and Chau, K.-W. (2019)‘‘Computational intelligence on short-term load forecasting: A methodological overview’’ Energies, vol. 12, no. 3, p. 393, Jan.doi:10.3390/en12030393.

Fang, X. Misha, S. Xue, G. and Yang, D.(2012). “Smart grid - the new andimproved power grid: A survey”IEEE Commun. Surveys Tutorials.Doi: 10.1109/SURV.2011.101911.00087.

Chollet, F. (2015). “Keras: The python deep learning API”. Disponível em https://keras.io/. Acessado em 30 de maio 2021

Raza, Khosravi A. (2015). “A review on artificial intelligence based load demandforecasting techniques for smart grid and buildings”. Renew Sustain EnergyRev50:1352–72. Doi: http://dx.doi.org/10.1016/j.rser.2015.04.06