The most effective method for insulation assessment in electrical power apparatus is Partial Discharge (PD) detection. During measurements the interference from background environment hampers the PD signal and reduces its measurement accuracy. This paper discuss on the implementation of a Hankel matrix based Fast Singular Value Decomposition (H-FSVD) technique for removing noise from the PD signals. The data is first represented into Hankel Matrix (HM) structure, with appropriate sampling then using Lanczo process the Hankel matrix size is reduced and through Singular Spectral Analysis thresholds are fixed for noise detection and removal. This algorithm has been examined on simulated as well as PD signals measured on two different laboratory environments from transformers with real and simulated noise.The experiment is part of a series of experiments to detect PD patterns related to realistic transformer defects. The denoising performance of H-FSVD is compared with the wavelet based denoising methods, Empirical Mode Decomposition method and normal SVD. Numerical results show that H-FSVD efficiently removes the noise with less computation time, even for large size data.

The insulation systems of equipment, cables, and electrical machines subjected to high voltage are continuously exposed to multiple aging mechanisms of electrical, mechanical, thermal, and environmental type. Normally, a failure in the material under these conditions does not occur immediately, but over time, different degradation processes emerge and evolve, progressively deteriorating the insulation, until breakdown occurs, and consequently, the asset finishes its operation with a catastrophic failure. In this context, partial discharge (PD) measurement can be considered as one of the best indicators when diagnosing the status of much electrical equipment in operation. However, the simultaneous presence of noise sources or multiple PD sources can generate important difficulties in identifying the type or types of sources measured. These practical limitations can be solved if, prior to the identification, a separation process is carried out, which allows classifying the different sources acting over the equipment being monitored. Once the sources separation is executed, the subsequent identification and diagnosis process can be carried out more easily. In this article, we present a novel separation technique based on the temporal and spectral behavior of the PD signals. For this technique, the separation of sources is carried out through three different parameters. Two of them are based on the peakedness of PD signals and electrical noise, and the third parameter is associated with their spectral content. Based on these parameters, a 3-D separation map is established, representing in clusters each source captured by the sensors.

The electricity sector heavily relies on oil-filled electricity equipment, particularly power transformers, which are critical and costly components in power generation and transmission systems. However, concerns arise as many of these assets have surpassed their useful life or are nearing the end of it. For this reason, the monitoring and diagnosis of failures in liquid insulation systems plays an important role when it comes to extending the life of these pieces of equipment. In its current state, the measurement systems used to capture and quantify the evolution of gases have limitations hindering their widespread use in routine measurements. These limitations have forced a large part of the research and development efforts to focus on developing and proposing new forms of measurement that can be applied without any type of technical-economic restriction and still provide a much more accurate diagnosis of the failure. This paper introduces a novel system that utilizes an electronic nose equipped with 8 MOS-type gas sensors to measure dissolved gases in liquid insulation systems. The obtained results validate the system’s exceptional performance in differentiating mineral oil samples based on the type and concentration of the predominant gases. This innovative approach shows great promise for routine monitoring and diagnosis, offering an efficient and cost-effective solution. Additionally, it exhibits significant potential for widespread implementation and provides a reliable means of assessing the condition of liquid insulation systems in various electrical assets.

Localization of partial discharge (PD) is a reliable and necessary technique for early prediction of impending failure in transformer windings. Different PD sources have diverse impacts on insulation, and PDs measured at different locations exhibit unique energy characteristics. An approach based on the energy characteristics of the PD signal is proposed here to locate PD inside the winding. An electrical detection method has been considered with PD pulses measured at the top and bottom of the winding through high-frequency current transformers (HFCTs). To reveal multiscale intrinsic characteristics of the PD signal, authors extracted intrinsic mode function (IMF) using ensemble empirical mode decomposition (EEMD). Next, the instantaneous energy distribution-permutation entropy (PE-IED) values of the first several IMFs were calculated to capture the energy of the PD. Then, dimensionality reduction was ensured by the fast principal component analysis that uses the fixed-point algorithm method. Finally, adaptive density-based clustering using the nearest neighbor graph (ADBSCAN-NNG) approach is proposed to identify the local high-density energy components. The proposed method is verified using the winding of a vacuum cast coil transformer and the localization results reveal the feasibility of the proposed method in locating PDs over existing localization algorithms. The reduced complexity of the proposed algorithm ensures the accurate onsite PD localization in transformer windings.

Detection and measurement of partial discharge (PD) phenomena combined with the separation and identification of PD sources is the way to achieve effective insulation integrity assessment. However, during measurement, PD signals are coupled with interferences (discrete spectral, pulsive, and white noises). Recovering PD signals from such interferences would improve PD source separation (thus identification), but still remains a challenging task. Several denoising methods have been proposed to suppress interferences. However, using a universal method to achieve interference removal is probably impossible, as the characteristics of the interferences are distinct. This study proposes a novel low‐rank H‐Matrix‐based singular value decomposition (SVD) filter (H‐SVD) that removes different types of interferences. Denoising is done by projecting the measured pulse in a lower dimensional signal space. To assess the effectiveness of the proposed method, H‐SVD filter is first applied to simulated PD data and later on real‐time PD data with the introduction of three different types of synthetic noises. The results of the evaluation metrics confirm that H‐SVD has significant performance improvements compared to existing state‐of‐the‐art PD denoising methods.

Precise partial discharge (PD) detection is a key factor in anticipating insulation failures. The continuous efforts of researchers have led to the design of a variety of algorithms focusing on PD pattern classification. However, the trade-off between features taken up for classification and the detection rate continues to pose considerable challenges in terms of feature selection from acquired data, increased computing time, and so on. In this article, a Hypergraph (HG) based improved Random Forest (RF) algorithm by employing the Recursive Feature Elimination (RFE) algorithm (HG-RF-RFE), has been developed for PD source classification. HG representation of data is considered for obtaining statistical features, which turn out to be a subset of a set of all hyper edges called Hyper statistical features (Helly, Non-Helly, and Isolated hyper edges). HG-RF-RFE takes hyper statistical features and hyper edges as features for classification. The algorithm’s efficiency is tested against noise-free PD data obtained from SASTRA High Voltage Laboratory, and large-sized noisy PD data obtained from High-Voltage Research and Test Laboratory at Universidad Técnica Federico Santa Maria (LIDAT). The robustness of the proposed algorithm is tested with both time and phase domain PD features using the Mathews Correlation Coefficient (MCC), harmonic mean-based feature Score (F1 Score) as evaluation metrics, and by k-fold validation technique. The proposed HG-RF-RFE achieved 98.8% accuracy with minimal features and significantly reduces computation time without compromising accuracy. It is worth mentioning that the HG-RF-RFE technique is superior to many state of the art algorithms in terms of feature elimination and classification accuracy.