Keywords
How to Cite
Abstract
This paper addresses the challenge of constructing a feature space for robust segmentation of load profiles in Arctic power systems, where extreme climatic non-stationarity and polar cycles invalidate assumptions of classical seasonality. Using data from three facilities collected between 2023 and 2025, a 36-dimensional feature vector was developed. Cluster analysis revealed 4–6 physically interpretable operating regimes, with Silhouette coefficients ranging from 0.68 to 0.74. Intra-cluster variance decreased by 40–60% compared to a single aggregated dataset, providing an empirical foundation for localized load modeling.
References
1. Hong T., Fan S. Probabilistic electric load forecasting: A tutorial review // International Journal of Forecasting. – 2016. – Vol. 32, № 3. – P. 914–938.
2- Andersson L., Alvehag K., Söder L. Reliability considerations for microgrids in remote areas // IEEE Transactions on Smart Grid. – 2018. – Vol. 9, № 4. – P. 3137–3145.
3- Gama J., Žliobaitė I., Bifet A., et al. A survey on concept drift adaptation // ACM Computing Surveys. – 2014. – Vol. 46, № 4. – Article 44.
4- Taylor J.W. Short-term electricity demand forecasting using double seasonal exponential smoothing // Journal of the Operational Research Society. – 2003. – Vol. 54, № 8. – P. 799–805.
5- Arrieta A.B., Díaz-Rodríguez N., del Ser J., et al. Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges // Information Fusion. – 2020. – Vol. 58. – P. 82–115.
This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright (c) 2026 Нагибин, Н.А. (Автор)