A Classification Method for Select Defects in Power Transformers Based on the Acoustic Signals

Abstract

Effective, accurate and adequately early detection of any potential defects in power transformers is still a challenging issue. As the acoustic method is known as one of the noninvasive and nondestructive testing methods, this paper proposes a new approach of the classification method for defect identification in power transformers based on the acoustic measurements. Typical application of acoustic emission (AE) method is the detection of partial discharges (PD); however, during PD measurements other defects may also be identified in the transformer. In this research, a database of various signal sources recorded during acoustic PD measurements in real-life power transformers over several years was gathered. Furthermore, all of the signals are divided into two groups (PD/other) and in the second step into eight classes of various defects. Based on these, selected classification models including machine learning algorithms were applied to training and validation. Energy patterns based on the discrete wavelet transform (DWT) were used as model inputs. As a result, the presented method allows one to identify with high accuracy, not only the selected kind of PD (1st step), but other kinds of faults or anomalies within the transformer being tested (2nd step). The proposed two-step classification method may be applied as a supplementary part of a technical condition assessment system or decision support system for management of power transformers.

Distribution Transformer Parameters Detection Based on Low-Frequency Noise, Machine Learning Methods, and Evolutionary Algorithm

“…The research presented in [ 12 , 13 , 14 , 15 ] proposes methods that precisely detect transformer flaws and discharge with an accuracy above 90%; nevertheless, the sensors have to be placed in close vicinity of a transformer. The proposed method, using low-frequency sound, allows the detection of basic parameters of various transformers from a distance of 50 m with an accuracy of at least 80%.…”

“…Many cases of power transformer breakdowns are related to insulation system failures, which might have been caused by the high activity of partial discharges [14]. Kunicki et al [15] proposed a method for detecting defects of power transformers. This method is based on machine learning classification of selected faults.…”

“…The classification is based on various machine learning algorithms (ML), which finds a pattern in data based on expert feedback. Therefore, in [12][13][14][15], the supervised learning [16] was used to train a model for various applications. However, in this research, the analyzed data were gathered from a significant distance, thus unsupervised learning was used [17] to find anomaly in the background noise, the source of which was unknown.…”

Ferroelectret-based Hydrophone Employed in Oil Identification—A Machine Learning Approach

“…As previously mentioned, some research involving acoustic and ultrasonic analysis in the detection of oil in transformers has been done in the last few decades [10][11][12]15,16].…”

Low Working Temperature of ZnO-MoS2 Nanocomposites for Delaying Aging with Good Acetylene Gas-Sensing Properties

“…It is critical to ensure the safety and reliability of power transformers during the operation for industrial production [ 2 , 3 ]. At present, oil-immersed power transformers are mainly popular in China [ 4 , 5 ]. Hydrocarbons in the insulating oil will decompose following thermal or electrical failure, causing the C–H bond and C–C bond breaking and forming gas molecules [ 6 , 7 , 8 , 9 ].…”