A frequency response analysis (FRA) is a well-known technique for evaluating the mechanical stability of a power transformer’s active part components. FRA’s measuring practices have been industrialised and are codified in IEEE and IEC standards. However, because there is no valid coding in the standard, the interpretation of FRA data is still far from being a widely acknowledged and authoritative approach. This study proposes an innovative fault segmentation and localisation technique based on FRA data. The algorithm is based on regression analysis to estimate the repeatability and relationship between the FRA fingerprint and the latest measured data. Initially, the measuring frequency is discretised into three regions to narrow the location of the fault; the regression model of the fingerprint and current FRA data are then evaluated. As a benchmark, two statistical indicators are the employed benchmark against the proposed method. Finally, the proposed scheme identifies and characterises various transformer conditions, such as healthy windings, axial and radial winding deformations, core deformation and electrical faults. The database used in this study consists of FRA measurements from 70 mineral-oil-immersed power transformers of different designs, ratings and manufacturers that were physically inspected for various faults and comparable frequency regions. The results achieved corroborate the efficacy of the proposed regression analysis fault recognition algorithm (RAFRA) model for transformer fault diagnosis using FRA. Further recommendations are made to address the reproducibility concerns induced by multiple FRA testing conditions.

Metal oxide arresters (MOAs) are protective devices widely used in power transmission and distribution systems to clamp transient overvoltages caused by lightning strikes and switching events, thereby protecting electrical equipment and enhancing power system reliability. However, failure of an MOA could significantly compromise the insulation coordination and overvoltage protection of power systems, thus exposing critical assets to potentially damaging voltage stresses. Accurate and early assessment of the health condition of MOAs is essential for maintaining power system reliability and minimising the risk of catastrophic insulation failure. Owing to several limitations associated with model-based condition-monitoring techniques of MOAs, this paper investigates whether combining the temporal learning capability of Long Short-Term Memory (LSTM) with the discriminative strength of Support Vector Machine (SVM) can provide a more effective and computationally feasible approach for MOA health assessment. Therefore, a dataset comprising 5157 observations extracted from historical inspection records of 20 kV distribution surge arresters was used to train predictive models for assessing the degradation condition of MOAs. The historical inspection measurements reflect the operational condition and degradation characteristics of MOAs. The results demonstrate that the hybrid LSTM–SVM model achieved the highest predictive performance, with an $R^2$ of 0.916 and an F1-score of 0.9711, outperforming the standalone LSTM and SVM models. However, this improvement is achieved at the expense of increased computational requirements, with the hybrid model requiring 367.8 s of training time and 548 MB of peak memory, which are higher than that observed for LSTM and SVM. Despite its higher training cost, the LSTM–SVM model achieved an inference time of 1.45 ms/sample, which is lower than that of LSTM, although higher than that of SVM. Statistical validation of the LSTM–SVM model based on the Diebold–Mariano (DM) test yields a p-value < 0.001 and negative DM statistic, providing statistical evidence of improved predictive performance. These findings suggest that the LSTM–SVM model could enable earlier identification of deteriorating MOAs, thereby supporting timely maintenance interventions and reducing the likelihood of unexpected failures and associated power supply disruptions.

This paper investigates whether event-level pitch-bearing fatigue damage can be estimated directly from turbine measurements, and whether these mechanical damage metrics leave measurable fingerprints in the generator DC-link voltage and current. To achieve this, a case study was performed using SCADA and structural load data from the 45 kW Chalmers (Björkö) research turbine. This data was segmented into 223 park-run-park pitch events. For each event, blade-root flapwise and edgewise bending moments were converted into radial and axial loads at the pitch bearing; an equivalent dynamic bearing load $P_{\mathrm{eq}}\left(t\right)$ was reconstructed using SKF and DG03 formulations; and rainflow counting with an S–N curve and Palmgren–Miner’s rule was used to compute event-level damage indices compatible with the International Standard Organization basic rating life concepts. In parallel, DC-link voltage and current were summarized into time-domain features, combined with operating-condition descriptors, and clustered using PCA-based k-means. The resulting clusters captured distinct electrical regimes that, across several event batches, corresponded to different levels of accumulated fatigue damage: regimes with sustained high DC-link voltage and longer duration tended to exhibit higher mean damage indices than lower, steadier DC regimes, indicating an electromechanical link. The results show that physics-based lifetime estimation and unsupervised analysis of existing electrical traces can be combined into a hybrid workflow for pitch-bearing condition assessment without additional sensors.