Home // ENERGY 2023, The Thirteenth International Conference on Smart Grids, Green Communications and IT Energy-aware Technologies // View article

Performance Analysis of Single and Multi-step Short-term Load Forecasts Using Multilayer Perceptron

Authors:
Athanasios Ioannis Arvanitidis
Dimitrios Kontogiannis
Georgios Vontzos
Vasileios Laitsos
Dimitrios Bargiotas
Miltiadis Alamaniotis

Keywords: multi-layer perceptron, univariate prediction, multivariate prediction, short-term load forecasting

Abstract:
Load forecasting is one of the most critical factors in modern power systems since it is the cornerstone for efficient monitoring, resource management and decision making. Therefore, there is an accrescent need for accurate and fast electrical load predictions. Many scientific approaches have been carried out in the field of load forecasting. In particular, the field of Machine Learning has attracted great research interest, due to the ability to adapt to time-series through forecasting tasks on multiple prediction horizons, a research area that presents challenges to several traditional methods. For this purpose, this research offers a thorough comparative study of several structural morphologies of Multi-Layer Perceptrons, in order to investigate electrical load forecasting accuracy for one, twelve and twenty-four time-steps ahead. Based on data from the Greek Power System for the years 2017 to 2019, the three proposed neural networks' structural morphologies are assessed in terms of precision through the Mean Absolute Error, Mean Squared Error, and Mean Absolute Percent Error of the predicted outcomes.

Pages: 63 to 67

Copyright: Copyright (c) IARIA, 2023

Publication date: March 13, 2023

Published in: conference

ISSN: 2308-412X

ISBN: 978-1-68558-054-4

Location: Barcelona, Spain

Dates: from March 13, 2023 to March 17, 2023