Conference paper Open Access
PROGNOZA NET POZICIJA PRIMENOM NEURALNIH MREŽA
Prešić, Dušan; Tasić, Andrej
FORECAST OF NET POSITINS USING NEURAL NETWORKS
Abstract – Common Grid Model Alignment (CGMA) process is a set of procedures by which the initial estimates of net positions are revised such that the resulting set of net positions on the Pan-European level is balanced. The process is applied for those time horizons for which market schedules of net positions are not available – from two days ahead up to one year ahead. Beside the task to control CGMA process, Alignment Agents could be nominated by their TSOs regarding net position forecast. It is rather intuitive that initial estimates of net positions are more accurate if the forecast is performed at the regional level.
This paper describes theoretical foundations of neural networks, as well as its application on net position forecast. Also, based on collected set of input data, calculations for testing period of one year are performed, where hourly values of net positions are forecasted for following biding zones – Serbia, Bosnia and Herzegovina, Montenegro and Bulgaria. Forecast was performed for time horizon of two days ahead. At the end, several indicators that describe quality of the forecast on the yearly level are shown.
| Name | Size | |
|---|---|---|
|
PROGNOZA NET POZICIJA PRIMENOM NEURALNIH MREŽA.pdf
md5:19fef150a0388641a16f41142d6435d5 |
510.7 kB | Download |
| All versions | This version | |
|---|---|---|
| Views | 90 | 90 |
| Downloads | 51 | 51 |
| Data volume | 26.0 MB | 26.0 MB |
| Unique views | 77 | 77 |
| Unique downloads | 45 | 45 |
- Publication date:
- December 5, 2019
- DOI:
-
- Keyword(s):
- Net position Forecast Neural Network Regional Security Coordinator CGMA process
- Grants:
-
European Commission:
- CROSSBOW - CROSS BOrder management of variable renewable energies and storage units enabling a transnational Wholesale market (773430)
- Meeting:
- 34. Session of CIGRE Serbia, Vrnjačka banja, Serbia, 02-06 June 2019 (Session C2)
- License (for files):
- Creative Commons Attribution 4.0 International