The genesis of the research work presented in this paper constitutes the issue of the effective and efficient recognition of single-source one-time partial discharge forms that can occur in insulation systems of power transformers. The paper presents research results referring to the use of single-direction artificial neural networks for recognizing basic partial discharge forms that can occur in paper-oil insulation impaired by aging processes. The research work results presented show the recognition effectiveness of basic partial discharge forms depending on the descriptor of the analysis of the acoustic emission signal analysis. The detailed cognitive aim was selection of input parameters and an artificial neural network which would be the best, considering recognition effectiveness and processing time, and which could be used as a classifier in an expert diagnostic system making identification of partial discharges measured by using the acoustic method possible.

Application possibilities of artificial neural networks for recognizing partial discharges measured by the acoustic emission method

This paper reviews the main features of acoustic emission technique. First, the characteristics of acoustic signals and their physics are addressed. Then, the structure of acoustic sensors employed for capturing signals is studied. In the next step, the acoustic method of PD measurement is compared with the standard electrical method. Afterward, the applications of AET in PD measurement in different equipment are summarized. All acoustic method and combined acoustic-electrical method for PD localization in transformers are discussed. Moreover, the applications of AET in the monitoring of GIS systems are explained, along with the acoustic behaviors of moving particles in these systems. Finally, the general considerations in the signal processing of the acoustic signals are reviewed.

Applications of the Acoustic Method in Partial Discharge Measurement: A Review

Nowadays, the acoustic detection is widely used for defect diagnosis of gas insulated substations (GIS) in normal operation and factory tests. In this paper in order to develop a data analyzer for acoustic detection system to make an assistant diagnosis, the characteristic of acoustic signals generated by different artificial defects such as protrusions, floating shield, void in spacer and bouncing particles are investigated. Some meaningful parameters behind the detected acoustic signals are extracted and discussed, which are used to distinguish background noise, partial discharge (PD) phenomena or bouncing particles. Based on those works, a comprehensive identification method realized by processing the acoustic pulse sequences qi, ( Δ ti, qi) and (ti, qi) is introduced, which gives a recognition result with noise, PD type or bouncing particles. For the sequence (ti, qi), the backpropagation artificial neural network optimized by genetic algorithm (GA-BPANN) is used as a classifier based on the fingerprint consisting of 24 operators, which are derivate from typical 2D histograms of phase-resolved partial discharge (PRPD). And with considering the trigger source may having a phase difference from the working voltage, identification with phase compensation (IPC) is used as a try to deal with the challenge. Experimental results show that the comprehensive identification method is practical and effective.

Investigation of a comprehensive identification method used in acoustic detection system for GIS

Partial discharge (PD) measurements can evaluate integrity of transformers’ insulation systems. Current research focuses on multiple PD sources separation to identify the types of insulation defects that may coexist in a transformer. This paper proposes a time-frequency (TF) sparsity map for revealing and separating different PD sources. TF sparsity map is developed based on decomposing signals into time and frequency domains at multiresolutions. Two decomposition methods, conventional wavelet transform-based signal decomposition and novel mathematical morphology (MM)-based signal decomposition are implemented in this paper. After sparsity values are calculated from the decomposed signals in time and frequency domains, sparsity trends are determined to provide unique representation of PD sources. By taking roughness of the trends, an accurate separation of multiple PD sources is obtained on a TF map. A density-based clustering is then evoked to form clusters related to different PD sources. The proposed method has been verified by signals acquired from multiple PD source models and substation transformers. Results show that an accurate representation of PD pulses in the presence of multiple PD sources and subsequently separation of PD sources can be achieved. Comparisons of wavelet transform and MM-based signal decomposition methods on TF sparsity maps construction and multiple PD sources separation are also provided.

Time-frequency sparsity map on automatic partial discharge sources separation for power transformer condition assessment

Partial discharge (PD) in transformer oil is one of the major causes of electrical insulation failures in oil transformers. An accurate location of PD provides important information required for detection and elimination of insulation defects. To improve accuracy of PD location, a new location method is proposed in this paper based on using circular array of ultrasonic sensors (CAUS) and implemented as follows. First, the fast independent component analysis (FastICA) algorithm is used to de-noise PD signal obtained by CAUS. Second, the total least square (TLS) algorithm is used to transform the wide-band signal obtained by CAUS to the narrowband one. Third, the mode excitation algorithm and the fast direction of arrival (FastDOA) algorithm are combined together to find three different directions of arrival (DOA) of the signal from PD to CAUS. Next, using the azimuth and pitch angles of these three DOAs, the PD location is determined in the center of a sphere whose surface tangents to these three DOAs in three different planes. Finally, the improved genetic algorithm is used to obtain accurate coordinates of the PD source. An experimental system of PD-location has been developed, which consists of (i) CAUS made of 8 sensors evenly placed along a circle, and a single sensor placed in the center of the circle, (ii) classic three-capacitance PD model, (iii) 16-channel synchronous data collector, (iv) oil tank, and (v) computer. The preliminary experimental results show that the accuracy of PD location after denoising is better than that before de-noising.

Location of partial discharge in transformer oil using circular array of ultrasonic sensors

This study proposes a noncontact type of acoustic measurement system and applies wavelet transform to noise suppression in order to raise the correct partial discharge (PD) signal identification rate. We first investigate an acoustic measurement method in the laboratory and apply the wavelet transform to suppress noise. A wavelet mother function most similar to the acoustic PD signals is chosen to set the filtering threshold value for the wavelet transform. Finally, the proposed acoustic measurement system is applied on line to epoxy-resin transformers, power distributors and the like. The superior measurement results we obtained will be able to help maintenance personnel eliminate field noise to correctly identify the PD fault types of power equipment.

Study of Partial Discharge Measurement in Power Equipment Using Acoustic Technique and Wavelet Transform

This paper first investigates the correlation between acoustic signals and electric power-source signal angles using ultrasonic detection technique. This paper then compares said acoustic signals' three-dimensional patterns with their polar-coordinate patterns and, with angle-zones established, provides a new analyzing and identification method for partial discharge (PD) fault types. This method is applicable to field inspection and measurement and can help identify epoxy-resin transformers' PD fault types. We first produce specific PD fault types in the laboratory and receive PD signals using ultrasonic technique. Then, we adopt polar-coordinate patterns to analyze their characteristics, deduce said characteristics into zone-identification patterns, and apply the D parameters to identify fault types. After utilizing this analyzing method to field epoxy-resin transformers, we learn that its diagnosis ability is fine and effective.

Yu-Hsun Lin

Papers

Study of Partial Discharge Measurement in Power Equipment Using Acoustic Technique and Wavelet Transform

New diagnosis approach to epoxy resin transformer partial discharge using acoustic technology