{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,24]],"date-time":"2026-02-24T10:05:02Z","timestamp":1771927502190,"version":"3.50.1"},"reference-count":56,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,1,27]],"date-time":"2026-01-27T00:00:00Z","timestamp":1769472000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>High-voltage substations form the backbone of critical electrical infrastructure, making predictive maintenance essential for ensuring grid resilience and operational reliability. Federated learning (FL) presents an innovative strategy for predictive maintenance, allowing multiple utility providers to improve model performance jointly while maintaining data confidentiality. Rather than transmitting raw records, each electrical utility performs local model updates and shares only the refined parameters, thereby safeguarding sensitive information and capitalizing on the heterogeneity of equipment conditions across sites. This study develops a set of privacy-preserving FL frameworks to enhance preventive maintenance of substation circuit breakers, large power transformers, and emergency generators. It rigorously tackles the issue of data heterogeneity arising from variations in distribution patterns across utilities, an inherent challenge that hampers effective collaborative model development. Four FL strategies\u2014Federated Averaging (FedAvg and FedAvgM), Federated Proximal (FedProx), and Federated Batch Normalization (FedBN), are evaluated for robustness against distributional shifts. Model performance in this study is evaluated using the F-score, which for the non-IID case ranges from 0.60 to 0.88 depending on the number of clients, the federated learning algorithm used, and the non-IID partitioning strategy employed. Also, a first-of-a kind Federated Information Criterion (FIC) is proposed in this manuscript as an extension of the classical information criterion. The results demonstrate that FedBN is best suited in mitigating cross-utility heterogeneity, yielding highest F-score of 0.88 and a moderately low FIC score of 4.35. Such tailored FL methods significantly improve predictive accuracy, enabling scalable and privacy-preserving deployment of FL in critical power system applications.<\/jats:p>","DOI":"10.3389\/frai.2025.1697175","type":"journal-article","created":{"date-parts":[[2026,1,27]],"date-time":"2026-01-27T06:39:29Z","timestamp":1769495969000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Federated learning for critical electrical infrastructure\u2014handling data heterogeneity for predictive maintenance of substation equipment"],"prefix":"10.3389","volume":"8","author":[{"given":"Soham","family":"Ghosh","sequence":"first","affiliation":[{"name":"Department of Electrical Engineering, Black and Veatch","place":["Overland Park, KS, United States"]}]},{"given":"Gaurav","family":"Mittal","sequence":"additional","affiliation":[{"name":"Department of Enterprise Solutions, Black and Veatch","place":["Overland Park, KS, United States"]}]}],"member":"1965","published-online":{"date-parts":[[2026,1,27]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"7331","DOI":"10.3390\/s23177331","article-title":"Federated learning for predictive maintenance and anomaly detection using time series data distribution shifts in manufacturing processes","volume":"23","author":"Ahn","year":"2023","journal-title":"Sensors"},{"key":"ref2","author":"Akbari","year":"2008"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"6241","DOI":"10.3390\/en15176241","article-title":"FedDP: a privacy-protecting theft detection scheme in smart grids using federated learning","volume":"15","author":"Ashraf","year":"2022","journal-title":"Energies"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"7840","DOI":"10.3390\/s23187840","article-title":"Low-latency collaborative predictive maintenance: over-the-air federated learning in noisy industrial environments","volume":"23","author":"Bemani","year":"2023","journal-title":"Sensors"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"1512","DOI":"10.3390\/s21041512","article-title":"An ensemble learning solution for predictive maintenance of wind turbines main bearing","volume":"21","author":"Beretta","year":"2021","journal-title":"Sensors"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"120367","DOI":"10.1109\/ACCESS.2021.3108839","article-title":"Privacy-aware resource sharing in cross-device federated model training for collaborative predictive maintenance","volume":"9","author":"Bharti","year":"2021","journal-title":"IEEE Access"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"9048","DOI":"10.1016\/j.comnet.2022.109048","article-title":"Fusion of federated learning and industrial internet of things: a survey","volume":"212","author":"Boobalan","year":"2022","journal-title":"Comput. Netw."},{"key":"ref8","doi-asserted-by":"publisher","first-page":"26","DOI":"10.1515\/sbeef-2017-0018","article-title":"The major predictive maintenance actions of the electric equipments in the industrial facilities","volume":"18","author":"Cazacu","year":"2018","journal-title":"Sci. Bulletin Elect. Eng. Faculty"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"7339","DOI":"10.1109\/JIOT.2023.3325822","article-title":"Federated learning for healthcare applications","volume":"11","author":"Chaddad","year":"2024","journal-title":"IEEE Internet Things J."},{"key":"ref10","doi-asserted-by":"publisher","first-page":"583","DOI":"10.1016\/B978-0-444-51862-0.50018-6","article-title":"AIC, BIC and recent advances in model selection","volume":"7","author":"Chakrabarti","year":"2011","journal-title":"Philosophy Statist."},{"key":"ref11","doi-asserted-by":"publisher","first-page":"2859","DOI":"10.3390\/en13112859","article-title":"Combined duval pentagons: a simplified approach","volume":"13","author":"Cheim","year":"2020","journal-title":"Energies"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1520\/SSMS20200029","article-title":"Federated learning as a privacy-providing machine learning for defect predictions in smart manufacturing","volume":"5","author":"da Silveira Dib","year":"2021","journal-title":"Smart Sust. Manufact. Syst."},{"key":"ref13","author":"Dasari","year":"2021"},{"key":"ref14","first-page":"5315","article-title":"Flamby: datasets and benchmarks for cross-silo federated learning in realistic healthcare settings","volume":"35","author":"du Ogier Terrail","year":"2022","journal-title":"Adv. Neural Inf. Proces. Syst."},{"key":"ref15","doi-asserted-by":"publisher","first-page":"1113\u20131122","DOI":"10.1007\/s40031-021-00599-1","article-title":"Ensemble machine learning methods for better dynamic assessment of transformer status","volume":"102","author":"Ghosh","year":"2021","journal-title":"J. Inst. Eng. (India) Ser. B"},{"key":"ref16","author":"Giuseppi","year":"2022"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"1037","DOI":"10.1109\/TPWRD.2002.803840","article-title":"Voltage sag and recovery time in repetitive events","volume":"17","author":"Gomex","year":"2002","journal-title":"IEEE Trans. Power Deliv."},{"key":"ref18","author":"He","year":"2021"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"15","DOI":"10.1109\/67.648490","article-title":"Electric power quality: a tutorial introduction","volume":"11","author":"Heydt","year":"1998","journal-title":"IEEE Comp. Appl. Power"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"87","DOI":"10.1016\/j.ijepes.2012.03.018","article-title":"Influence of voltage dips on industrial equipment: analysis and assessment","volume":"41","author":"Honrubia-Escribano","year":"2012","journal-title":"Int. J. Electr. Power Energy Syst."},{"key":"ref21","doi-asserted-by":"publisher","first-page":"6832","DOI":"10.3390\/app11156832","article-title":"Improved ensemble-learning algorithm for predictive maintenance in the manufacturing process","volume":"11","author":"Hung","year":"2021","journal-title":"Appl. Sci."},{"key":"ref22","first-page":"867","article-title":"Substations","volume-title":"Springer handbook of power systems","author":"Krieg","year":"2019"},{"key":"ref23","doi-asserted-by":"publisher","first-page":"1803","DOI":"10.1109\/JPROC.2019.2919306","article-title":"Review and perspectives on data sharing and privacy in expanding electricity access","volume":"107","author":"Lee","year":"2019","journal-title":"Proc. IEEE"},{"key":"ref24","author":"Li","year":"2022"},{"key":"ref25","doi-asserted-by":"publisher","first-page":"106854","DOI":"10.1016\/j.cie.2020.106854","article-title":"A review of applications in federated learning","volume":"149","author":"Li","year":"2020","journal-title":"Comput. Ind. Eng."},{"key":"ref26","author":"Li","year":"2019"},{"key":"ref27","author":"Li","year":"2020"},{"key":"ref28","author":"Li","year":"2021"},{"key":"ref29","volume-title":"On the dirichlet distribution, Department of Mathematics and Statistics","author":"Lin","year":"2016"},{"key":"ref30","author":"Madonsela","year":"2018"},{"key":"ref31","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3625558","article-title":"Heterogeneous federated learning: state-of-the-art and research challenges","volume":"56","author":"Mang","year":"2023","journal-title":"ACM Comput. Surv."},{"key":"ref32","doi-asserted-by":"crossref","first-page":"133","DOI":"10.1016\/0304-4076(94)01612-4","article-title":"A generalization of the beta distribution with applications","volume":"66","author":"McDonald","year":"1995","journal-title":"J. Econ."},{"key":"ref33","author":"McMahan","year":"2017"},{"key":"ref34","doi-asserted-by":"publisher","first-page":"36","DOI":"10.1109\/MPOT.2011.940233","article-title":"Failures, monitoring and new trends of power transformers","volume":"30","author":"Metwally","year":"2011","journal-title":"IEEE Potentials"},{"key":"ref35","volume-title":"An introduction to predictive maintenance","author":"Mobley","year":"2002"},{"key":"ref36","doi-asserted-by":"publisher","first-page":"23","DOI":"10.3390\/s23135970","article-title":"From corrective to predictive maintenance\u2014a review of maintenance approaches for the power industry","volume":"23","author":"Mol\u0119da","year":"2023","journal-title":"Sensors"},{"key":"ref37","doi-asserted-by":"publisher","first-page":"220","DOI":"10.1016\/j.applthermaleng.2013.07.028","article-title":"Application of infrared thermography for predictive\/preventive maintenance of thermal defect in electrical equipment","volume":"61","author":"Nazmul Huda","year":"2013","journal-title":"Appl. Therm. Eng."},{"key":"ref38","author":"Nguyen","year":"2022"},{"key":"ref39","author":"Ram\u00edrez","year":"2023"},{"key":"ref40","doi-asserted-by":"publisher","first-page":"5257","DOI":"10.1007\/s13042-024-02234-z","article-title":"Comparative analysis of open-source federated learning frameworks - a literature-based survey and review","volume":"15","author":"Riedel","year":"2024","journal-title":"Int. J. Mach. Learn. Cybern."},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-020-00323-1","article-title":"The future of digital health with federated learning","volume":"3","author":"Rieke","year":"2020","journal-title":"NPJ Digit. Med."},{"key":"ref42","author":"Saidani","year":"2023"},{"key":"ref43","author":"Shi","year":"2023"},{"key":"ref44","doi-asserted-by":"publisher","first-page":"132","DOI":"10.1109\/MPE.2022.3153780","article-title":"Advancements in clean air insulation technologies for switchgear and circuit breakers","volume":"20","author":"Siddiqui","year":"2022","journal-title":"IEEE Power Energy Mag."},{"key":"ref45","doi-asserted-by":"publisher","first-page":"5605","DOI":"10.1109\/TII.2020.3034674","article-title":"Adaptive federated learning and digital twin for industrial internet of things","volume":"17","author":"Sun","year":"2021","journal-title":"IEEE Trans. Ind. Inform."},{"key":"ref46","doi-asserted-by":"publisher","first-page":"15164","DOI":"10.1609\/aaai.v38i13.29439","article-title":"On the role of server momentum in federated learning","volume":"38","author":"Sun","year":"2024","journal-title":"Proc. AAAI Conf. Artif. Intell."},{"key":"ref47","article-title":"A comprehensive experimental comparison between federated and centralized learning","author":"Swier","year":"2025","journal-title":"Database"},{"key":"ref48","doi-asserted-by":"publisher","first-page":"15420","DOI":"10.1609\/aaai.v38i14.29467","article-title":"Data disparity and temporal unavailability aware asynchronous federated learning for predictive maintenance on transportation fleets","volume":"38","author":"Wahl","year":"2024","journal-title":"Proc. AAAI Conf. Artif. Intell."},{"key":"ref49","doi-asserted-by":"publisher","first-page":"3637","DOI":"10.1109\/TSG.2021.3066577","article-title":"Electricity consumer characteristics identification: a federated learning approach","volume":"12","author":"Wang","year":"2021","journal-title":"IEEE Trans. Smart Grid"},{"key":"ref50","doi-asserted-by":"publisher","first-page":"6069","DOI":"10.1109\/JIOT.2021.3110784","article-title":"FedDetect: a novel privacy-preserving federated learning framework for energy theft detection in smart grid","volume":"9","author":"Wen","year":"2022","journal-title":"IEEE Internet Things J."},{"key":"ref51","first-page":"111106","article-title":"FedLLM-bench: realistic benchmarks for federated learning of large language models","volume":"37","author":"Ye","year":"2024","journal-title":"Adv. Neural Inf. Proces. Syst."},{"key":"ref52","author":"Zhao","year":"2021"},{"key":"ref53","doi-asserted-by":"publisher","first-page":"98","DOI":"10.1109\/TII.2022.3172901","article-title":"Cloud-edge collaboration-based local voltage control for DGs with privacy preservation","volume":"19","author":"Zhao","year":"2023","journal-title":"IEEE Trans. Ind. Inform."},{"key":"ref54","doi-asserted-by":"publisher","first-page":"3823","DOI":"10.1109\/TWC.2023.3311824","article-title":"Federated learning in heterogeneous networks with unreliable communication","volume":"23","author":"Zheng","year":"2024","journal-title":"IEEE Trans. Wirel. Commun."},{"key":"ref55","doi-asserted-by":"crossref","first-page":"012007","DOI":"10.1088\/1742-6596\/1113\/1\/012007","article-title":"The principle of least effort and Zipf distribution","volume":"1113","author":"Zhu","year":"2018","journal-title":"J. Phys. Conf. Ser."},{"key":"ref56","first-page":"111","volume-title":"Pysyft: a library for easy federated learning studies in computational intelligence","author":"Ziller","year":"2021"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1697175\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,24]],"date-time":"2026-02-24T09:12:44Z","timestamp":1771924364000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1697175\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,27]]},"references-count":56,"alternative-id":["10.3389\/frai.2025.1697175"],"URL":"https:\/\/doi.org\/10.3389\/frai.2025.1697175","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,1,27]]},"article-number":"1697175"}}