Midterm Load Forecasting: A Multistep Approach Based on Phase Space Reconstruction and Support Vector Machine
| Autor: | Gen Li, Yunhua Li, Farzad Roozitalab |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
021103 operations research
Electrical load Computer Networks and Communications Process (engineering) business.industry Computer science 0211 other engineering and technologies 02 engineering and technology Machine learning computer.software_genre Field (computer science) Computer Science Applications Data modeling Support vector machine Data point Control and Systems Engineering Robustness (computer science) Artificial intelligence Electrical and Electronic Engineering business Divergence (statistics) computer Information Systems |
| Zdroj: | IEEE Systems Journal. 14:4967-4977 |
| ISSN: | 2373-7816 1932-8184 |
| DOI: | 10.1109/jsyst.2019.2962971 |
| Popis: | Electrical load forecasting is a vital process for balancing the electricity supply and demand sections. In order to make a precise prediction, many intelligent methods aiming at forecasting future load of different horizons have been proposed in recent years. Regardless of the great progress in this field, yet two major problems widely exist in the field of midterm load forecasting, which are pseudo midterm forecast and divergence of the forecasting error. This article proposes a multistep forecasting approach, mainly based on phase space reconstruction and support vector machine (SVM) methods, to solve these two problems. The relations between the forecasting step in the future and the historical data points that need to be fed into the SVM model are derived, which endue the model with the real ability to forecast adjustable steps of future load without the divergence of error and, therefore, have a notable engineering application significance. The proposed methodology is implemented on the European Network on Intelligent Technologies dataset. Compared with previous methods, the results show that the multistep implemented model is of more accurate prediction and stronger robustness. Moreover, the method is easy to implement and able to be combined with other intelligent methods to get better performance. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|