2026· E3S Web of Conferences· Vol 729, pp. 07005· 0 citations
Abstract
While Paper Insulated Lead Covered (PILC) cables are a legacy technology, they remain a critical fixture in medium-voltage power grids. However, as these cables are continuously used, they present growing risks to grid stability and public safety due to electrical, thermal, and environmental factors. As replacing entire networks is cost-prohibitive, utility providers need a smarter way to predict when and where a cable might fail. This paper presents an explainable, data-driven framework for multi-class health assessment of 20 kV PILC cables, aligning with the industry 4.0 standards of autonomous asset management. Using a real-world dataset of 999 inspection records from European utilities, four supervised learning models; Random Forest, AdaBoost, XGBoost, and a Multilayer Perceptron deep neural network were trained. The models categorize cable health into five standard IEC/IEEE health bands, achieving high classification accuracies between 97.0% and 99.5%. To ensure transparency in the decision-making process, SHapley Additive exPlanations (SHAP) were employed, identifying Partial Discharge (PD) and Thermal Difference (TD) Stability as the primary predictors of insulation degradation. This framework provides an interpretable tool for power utilities to transition from reactive to predictive maintenance scheduling.
Lithium-ion batteries act as core energy suppliers for Automated Guided Vehicles. Once the power supply system fails, normal industrial production workflows will be disrupted severely. In practical field applications, it is impossible to directly measure the real health status of such batteries. This work develops a Bidirectional Long Short-Term Memory network whose hyperparameters are tuned by the Whale Optimization Algorithm, aiming to achieve accurate State-of- Health estimation for lithium-ion batteries. To begin with, we screen a set of health indicators closely linked to battery capacity and carry out targeted analysis for these indicators. After feature extraction, two statistical approaches, namely Pearson correlation analysis and Maximal Information Coefficient, are adopted to conduct quantitative evaluation. The analytical results prove that the selected indicators have strong correlations with battery capacity. On this basis, WOA is applied to refine the hyperparameters of the BiLSTM network. A series of comparative tests against mainstream baseline models verify that the proposed hybrid approach delivers higher precision and more stable estimation performance.
Zhen Ni, Ziyi Zhu, Kainan Zhang et al.· International Conference on...· 0 citations
The convergence of civil infrastructure and electrical power systems within smart city frameworks necessitates robust, cross-domain monitoring strategies. While machine learning (ML) has shown promise in isolated Structural Health Monitoring (SHM) and Predictive Maintenance (PdM), comparative evaluations across both domains remain fragmented. This study presents a comprehensive comparative analysis of four prominent ML algorithms Random Forest (RF), Support Vector Machines (SVM), Long Short-Term Memory (LSTM) networks, and XGBoost applied to multimodal sensor data. We utilized a synthesized dataset comprising vibration signatures from civil structures (bridge decks) and thermal-electrical load profiles from substation transformers. Our findings indicate that while LSTM networks excel in capturing temporal dependencies in electrical load forecasting (achieving an F1-score of 0.94), tree-based ensemble methods, specifically XGBoost, demonstrate superior efficacy in classifying structural damage from high-dimensional vibration features (accuracy of 96.2%). Furthermore, RF offered the most computationally efficient inference, making it highly suitable for edge-deployment in resource-constrained IoT nodes. This paper provides a practical decision-making framework for civil and electrical engineers selecting ML architectures for integrated smart infrastructure monitoring.
M. el-sseid, L. B. Ben Dalla, Tasnem ELsseid et al.· Al-Farooq Journal of Science...· 0 citations
This article explores machine learning techniques (MLTs) as a modern alternative to enhance the interpretation of DGA data for early-stage fault detection in service transformers, and demonstrates that random forest and gradient boosting outperform others, achieving up to 98% accuracy.
Rupali Balabantaraya, A. Chatterjee, A. Sahoo et al.· Electrica· 0 citations
Machine learning (ML) techniques have been widely applied to fault detection and diagnosis in Electric Submersible Pumps (ESPs), often reporting high predictive accuracy. However, high performance does not necessarily imply that learned decision boundaries reflect physically meaningful fault mechanisms. This study distinguishes epistemic interpretability, associated with model transparency, from physical interpretability, defined here as the stability of diagnostic decisions across distinct physical units, with operating-regime stability discussed as a broader requirement that cannot be directly isolated from the reference feature file used in this study. Current ML practices in ESP fault diagnosis are examined through a structured literature review and an empirical analysis of a public multi-pump vibration dataset. Representative supervised models are evaluated under sample-wise and cross-unit (pump-wise) validation. Performance is assessed using accuracy, macro-F1, confusion matrices, Principal Component Analysis-based class-space geometry, centroid distances, stability metrics, and bootstrap-based confidence intervals and empirical significance tests for cross-unit degradation. Results show that sample-wise validation can overestimate robustness, whereas cross-unit evaluation reveals performance degradation in most nonlinear and ensemble configurations, with statistical support in several high-capacity models. Class imbalance mitigation reduces majority-class bias but does not eliminate structural misclassification patterns. Representation-level analysis shows that fault condition is the stronger organizing factor in the feature space, while pump identity contributes a measurable but non-dominant fraction of feature-space variance. These findings indicate that predictive accuracy alone is insufficient to characterize diagnostic quality in ESP applications. Explicit cross-unit validation and stability-oriented evaluation are required to assess diagnostic robustness under conditions closer to deployment-relevant generalization.
Miguel A. de C. Michalski, Felipe L. Valentim, Gilberto F. M. De Souza· IEEE Access· 0 citations
Studies of machine-learning-based power-system protection increasingly report near-perfect scores, yet the meaning of those scores depends strongly on the evaluation setting. Protection task, physical scope, measurements, timing, targets, preprocessing, and validation often vary jointly and remain incompletely specified. This paper proposes a standardization-oriented framework that treats evaluation design as part of the scientific contribution. It defines seven required study dimensions: protection objective, physical scope, observability, timing and decision windows, targets and sample validity, validation protocol, and evaluation outputs. The framework is instantiated in a bounded case study on the public PROTECT-90 electromagnetic-transient benchmark, comprising 9022 simulated episodes from a 90 kV double-line topology, for onset-conditioned fault classification and localization. Under centralized sensing, simulation-metadata-aligned 20 ms windows, and episode-grouped validation, a multi-layer perceptron (MLP) achieved a five-fold mean macro-averaged F1 score of 0.991 +/- 0.001 for classification and a localization mean absolute error of 10.20 +/- 0.25% of line length (mean +/- std across episode-grouped folds). Extending the decision horizon to 50 ms preserved this task-dependent performance asymmetry, while reduced observability approximately doubled the MLP localization error but had little effect on classification. A synchronized two-ended conventional locator outperformed the learning locators under its richer clean information set, and measurement degradation showed that clean predictive performance did not determine robustness. The framework turns evaluation assumptions into explicit, reproducible evidence and provides a basis for more comparable, auditable evaluation and future certification-oriented assessment of machine-learning protection functions.
J. Oelhaf, Georg Kordowich, P. Pérez-Toro et al.· 0 citations
Electrical induction generators are vulnerable to winding faults that can degrade operational reliability and lead to unplanned maintenance. This study proposes a multiple-convolutional neural network (Multi-CNN) extreme ensemble learning framework for intelligent fault diagnosis and maintenance decision support using three-phase current and voltage measurements. Experimental recordings representing healthy operation, inter-turn faults, and inter-winding faults were segmented into non-overlapping 200-sample windows. Hjorth activity, mobility, and complexity were calculated for the three-phase current signals and the three-phase voltage signals, producing 18 features for each of 900 instances. Four convolutional neural network architectures were trained, and their class-probability outputs were combined through an extreme learning machine. Stratified blocked five-fold cross-validation was used to evaluate the models while preserving the chronological structure of the data. The proposed ensemble achieved 98.111% accuracy, 98.146% precision, 98.111% recall, 98.108% F1-score, and 97.167% Matthews correlation coefficient, correctly classifying 883 of 900 out-of-fold instances. It also attained a macro-averaged area under the receiver operating characteristic curve of 0.995. These results demonstrate that Hjorth-based electrical-signal characterization and Multi-CNN ensemble fusion can provide accurate and computationally efficient support for fault identification and predictive maintenance decisions in electrical induction generators.
M. Ahmed, Ahmed Mohammed Mohsin Alzubaidi, Z. Khan et al.· Applied System Innovation· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
