2025

Energies

Rocznik: 2025 | Tom: Vol. 18 | Numer: Nr 16

artykuł naukowy

angielski

EN The operational roles of artificial intelligence in energy security remain inconsistently defined across the scientific literature. To address this gap, the present review examines 165 peer-reviewed abstracts published between 2021 and 2025 using a triangulated methodology that combines trigram frequency analysis, manual qualitative coding, and semantic clustering with sentence embeddings. Eight core roles were identified: forecasting and prediction, optimisation of energy systems, renewable energy integration, monitoring and anomaly detection, grid management and stability, energy market operations/trading, cybersecurity, and infrastructure and resource planning. According to the results, the most frequently identified roles, based on the average distribution across all three methods, are forecasting and prediction, optimisation of energy systems, and energy market operations/trading. Roles such as cybersecurity and infrastructure and resource planning appear less frequently and are primarily detected through manual interpretation and semantic clustering. Trigram analysis alone failed to capture these functions due to terminological ambiguity or diffuse expression. However, correlation coefficients indicate high concordance between manual and semantic methods (Spearman’s ρ = 0.91), confirming the robustness of the classification. A structured typology of AI roles supports the development of more coherent analytical frameworks in energy research. Future research incorporating full texts, policy taxonomies, and real-world use cases may help integrate AI more effectively into energy security planning and decision support environments.

11.08.2025

1 - 23

10.3390/en18164275

https://www.mdpi.com/1996-1073/18/16/4275

Bibliografia, netografia na stronach 18-23.

otwarte czasopismo

ostateczna wersja opublikowana

11.08.2025

w momencie opublikowania

Pobierz plik

publiczny

140

3,2 [Lista 2024]