Summary
Red Eléctrica, the sole transmission system operator (TSO) for the Spanish electricity system, is responsible for ensuring the stable operation of the electrical system. To this end it is mandatory to study in detail angle, voltage and frequency stability for small and large disturbances [6]. In the case of large disturbance studies, the system's behaviour is analysed in the event of the loss of one element, commonly referred to as N-1 [1]. These studies are not only carried out in real time, but also in proposed future grid developments and in capacity access calculation (CA) [4] to allow connection of new generation or demand to the grid, among other objectives.
Read more Read lessTherefore, a reduced number of real scenarios and contingencies are usually selected to simulate, usually the most representative or most unfavourable ones. Even so, the computational workload can be high, and the results obtained can be excessively conservative.
Machine learning and artificial intelligence (AI) techniques significantly improve the time required to complete these studies [7,8,9,10]. AI has been used in the literature to select representative real scenarios, to characterize scenarios, to carry out security analysis, as well as to perform dynamic or static simulations to better understand aspects affecting stability, among other aspects. This paper details the use of the Artificial Intelligence Dynamic Assessment
(AIDA) tool, developed by Red Eléctrica, to ensure dynamic stability in grid development proposals. The tool was applied in 2030 horizon scenarios for the Balearic Islands electricity system [4]. AIDA allows significant number of dynamic simulations, as well as statistical and/or artificial intelligence (AI) analysis of voltage, frequency, and angle magnitudes during the simulations. Additionally, different AI techniques, like Deep Multi-Layer Perceptron
(DeepMLP), Density-Based Spatial Clustering of Applications with Noise (DBSCAN), Light
Gradient-Boosting Machine (LightBM) and graph neural network (GNN) are used to select a reduced number of representative scenarios for the stability study. AIDA can be used to ensure capacity allocation under stability security conditions to compare with traditional methods.
Furthermore, AIDA enables a significantly more efficient use of computational resources, using techniques such as multiprocessing, parallelization and intelligent task automatization.
Additional informations
| Publication type | Session Materials |
|---|---|
| Reference | C1_11496_2026 |
| Publication year | |
| Publisher | CIGRE |
| Country | Spain |
| Study committees | |
| File size | 882 KB |
| Price for non member | 30 € |
| Price for member | 30 € |
Authors
LORENZO CABRERA Eduardo - Red Eléctrica, Spain; GALLEGO FERNÁNDEZ Javier - Red Eléctrica, Spain; MORÁN RÍO Diana Patricia - Red Eléctrica, Spain; MATEO SÁNCHEZ Lucía - Red Eléctrica, Spain
Keywords
Machine learning, Power System stability, Artificial Intelligence, GNN, LGBM
