A non-iterative partial discharge source location method for transformers employing acoustic emission techniques

Partial discharge (PD) causes premature insulation failure of transformers. It is essential to detect PD to avoid unwanted failure of transformers. Only detection of PD is not sufficient for a transformer of huge size, unless it is possible to locate. Acoustic partial discharge measurement is advantageous for PD source location. There are different algorithms for PD source location. These are iterative and require large number of acoustic emission (AE) sensors. This paper presents a non-iterative source location algorithm employing four AE sensors. This algorithm is applied to experimental data. Proposed algorithm is also applied to published data and compared with existing iterative methods. Main error for source location is due to arrival time calculation. In order to reduce the error in AE signal arrival time calculation, different arrival time calculation methods are discussed and a comprehensive method is proposed. By applying these methods, arrival time is calculated from measured signal.     

Proposed method of partial discharge allocation with acoustic emission sensors within power transformers

Employing acoustic emission sensors for detection of partial discharge, PD, introduces many advantages. Besides easy installation and replacement, they are non-invasive and immune to electromagnetic noise and interference and their sensitivity does not vary with object capacitance. For PD allocating utilizing AE sensors, distance calculations are based on the arrival time of acoustic waves to the sensors. Considering structure-borne waves of higher speed, the peaks of some of indirect path AE signals with significant contribution are mistakenly considered as peaks of direct path AE signals. Furthermore, the acoustic signals are propagating through certain parts of the transformer, such as the windings, and this complicates the partial discharge detection and allocation. These would imply an incorrect distance between the source and sensor. A method based on a heuristic algorithm has been proposed which considering all possible indirect paths with the relevant propagation times and all the barriers on the travel path of acoustic signal, calculates the more precise arrival times to sensors. A test chamber has been utilized and artificial PD signals are produced at various points. Output results of algorithm have been compared with results of classic method. It has been shown that proposed method significantly reduces the positioning errors.     

Product analysis of partial discharge damage to oil-impregnated insulation paper

Surface products of oil-impregnated insulation paper during the damage process caused by partial discharge (PD), as well as gas within the cavity, were studied. An optical microscope and a scanning electron microscope (SEM) were used to investigate surface morphology, while an infrared spectroscopy (IR) and an X-ray photoelectron spectroscopy (XPS) were used to study surface products and their components. The volume variation in cavity gas was also analyzed. Furthermore, gas constituents and their relevant contents were studied using a gas chromatography-mass spectrometer (GC-MS). The study results reveal the following: during the PD damage process, the total gas volume and the content of electronegative gasses alternately decline and increase, while discharge types alternate between pulse type and pseudo-glow type (or glow type); “surface droplets” and “crystalline solids” appear on the insulation surface one after another; surface droplets mainly consist of (CO)-group-containing compounds, whereas crystalline solids are mainly carboxylic acids, with carboxyl groups also found in cellulose chains; and the discharge type related to the oxidization of decomposition products is the main factor that determines the state (liquid or solid) of the surface products.     

Localization of acoustic emission sources in tensile and ct specimens using a broadband acquisition technique

The acoustic emission sources in a conventional cylindrical tensile test sample of short transversely-cut carbon manganese steel are localized. There is not always a good correlation between the localization of the first signals and the zone which eventually fractures. During the Lüder's plateau, the ae signals are emitted in the deformation band and, in the hardening range, there is no significant ae in the gauge length of the sample.

In ct samples precracked by fatigue, the signals are due to the growth of the plastic zone around the crack tip, and the plastic zone size, measured by source localization, agrees with those provided by models derived from fracture mechanics.     

Location of acoustic emission sources generated by air flow

The location of continuous acoustic emission sources is a difficult problem of non-destructive testing. This article describes one-dimensional location of continuous acoustic emission sources by using an intelligent locator. The intelligent locator solves a location problem based on learning from examples. To verify whether continuous acoustic emission caused by leakage air flow can be located accurately by the intelligent locator, an experiment on a thin aluminum band was performed. Results show that it is possible to determine an accurate location by using a combination of a cross-correlation function with an appropriate bandpass filter. By using this combination, discrete and continuous acoustic emission sources can be located by using discrete acoustic emission sources for locator learning.     

Effect of partial insulation winding scheme on discharge characteristics of GdBCO coils

In this study, sudden discharge tests were performed on partially-insulated (PI) GdBCO coated conductor single-pancake coils to investigate the effects of a PI winding scheme on the discharge characteristics. Three sets of PI coils were constructed with insulation at 3 different intervals: every 2nd, 4th, and 8th turn. Each set included coils with 4 different insulation-angles: 90°, 180°, 270°, and 360°. Discharge delay of the PI coils was alleviated by increasing the insulation-interval or insulation-angle due to increased insulated area, which partially hinders radial current flow. The results of the sudden discharge test were in fairly good agreement with simulations of the PI coils using an equivalent circuit model consisting of a single RL circuit in parallel. The characteristic resistance of the PI coils was reasonably estimated through the numerical approach with contact surface resistance and compensation for the current-bottleneck effect, which enables the PI coils to be utilized in practical superconducting magnet applications.     

Numerical simulation of acoustic emission using acoustic contact elements

A new method is described for nonconformal finite-element simulation of the region of interaction between an acoustic fluid and deformed solid bodies. The method is based on the use of special contact finite elements simulating the interpolation coupling between the acoustic pressure in the fluid and displacements of a two-phase surface. The test results of the method demonstrate a significant acceleration of grid convergence of computations. The numerical method is verified by the problem of predicting noise emission by a thin-wall inhomogeneous shell.     

双声源位置和强度识别的三维声强方法

目前在远场识别声源空间位置和强度缺乏行之有效的方法。针对此问题,提出采用四传声器进行三维声强测量,从而构建出声强、声源坐标和声功率的非车线性方程组,求解方程得出声源空间坐标和强度的方法。以3个三维声强探头对两个同频率单极子声源的识别为例,分别利用数值仿真和半消声室内的实验进行方法验证,并对声源的识别空间分辨率做了测试,得出角度识别最大误差为3.83°,为真实值的8.5%,距离识别最大误差0.1 m,为真实距离的10%。结果表明采用该方法空间坐标和声功率识别均具有很高的准确度,双声源的空间位置分辨力也优于远场声全息方法。     

Results of the wavelet analysis of the acoustic emission signals generated by partial discharges in insulation oil of various temperature and flow speed

The paper presents measurement results of the acoustic emission (AE) signals generated by partial discharges (PDs) at insulation oil temperature and flow speed changes. For the AE signals registered, a wavelet analysis based on the use of the continuous (CWT) and discrete (DWT) wavelet transforms was carried out. CWT results are shown as scaling graphs, and DWT results in the form of time runs determined at seven decomposition levels and bar diagrams visualizing the amount of energy transferred by the particular details.     

Measurements and analysis of the acoustic signals produced by partial discharges in insulation oil

This paper focuses on the frequency analysis of acoustic signals produced by partial discharges (PDs) in insulation oil and the positioning of the PD occurrence for application in the diagnosis of oil-insulated transformers. Three types of electrode systems; the needle–plane, the plane–plane, and the wire–wire structures were assembled to simulate the partial discharge in insulation oil. A low-noise amplifier and a de-coupler were designed to detect the acoustic signal with high-sensitivity. The frequency ranges of the acoustic signal were 60–270 kHz in the needle–plane electrode system, 45–250 kHz in the plane–plane electrode system, and 50–180 kHz in the wire–wire electrode system. Their peak frequencies were 145 kHz, 118 kHz and 121 kHz, respectively.The position of the PD occurrence was calculated from the time difference of arrival (TOA) using three acoustic emission (AE) sensors. The position was found within a 1% error in the experimental set-up.     

A lasing wavelength stabilized simultaneous multipoint acoustic sensing system using pressure-coupled fiber Bragg gratings

A fiber Bragg grating (FBG) sensor head, using a pressure coupling mechanism, was designed for broadband frequency response and structural strain-free characteristic. The pressure-coupled sensor heads were connected to a simultaneous multipoint acoustic sensing system based on a tunable laser. An intelligent lasing wavelength stabilization algorithm capable of identifying the direction of spectrum movement, the wavelength shifting speed, and a fiber bending event was developed so that the simultaneous multipoint acoustic sensing system could be used in environments with rapid temperature variations. The lasing wavelength feedback control algorithm updated the lasing wavelength into the steep slope of the FBG spectrum even under conditions of rapid temperature change. The averaging lasing wavelength updating time was only 21 s because the system can decide a minimal size in scan window by finding the FBG spectrum shifting speed and direction in real time. The system was able to update the lasing wavelength which missed the steep slope of the FBG spectrum under maximum temperature variation rates 0.3014 and −0.3246 °C/s. The proposed system detected simultaneous impact waves at multiple points under conditions of rapid temperature change and change in dynamic strain.     

Time-delay estimation of acoustic emission signals using ICA

Acoustic emission (AE) analysis is used for characterization and location of developing defects in materials. AE sources often generate a mixture of various statistically independent signals. One difficult problem of AE analysis is the separation and characterization of signal components when the signals from various sources and the way in which the signals were mixed are unknown. Recently, blind source separation by independent component analysis (ICA) has been used to solve these problems. The main purpose of this paper is to demonstrate the applicability of ICA to time-delay (T-D) estimation of two independent continuous AE sources on an aluminum beam. It is shown that it is possible to estimate T-Ds by ICA, and thus to locate two independent simultaneously emitted sources.     

铝合金材料腐蚀损伤的声发射评价

为了研究如何利用声发射(AE)技术监测飞机主结构件的腐蚀损伤从而为确定飞机日历寿命提供科学依据。首先介绍了航空铝合金材料在加速腐蚀过程中的AE信号特征,说明如何利用AE技术来监测腐蚀损伤,进而讨论如何对材料的安全性能和剩余寿命进行评估。试验表明, AE能比超声和涡流等常规无损检测方法更早地发现铝合金材料的腐蚀损伤。通过研究加速腐蚀过程中腐蚀损伤程度及腐蚀AE信号随加速腐蚀时间的变化规律,获得了腐蚀损伤与AE参数之间的变化关系。说明AE技术可用于探测早期腐蚀、研究腐蚀发展规律及监测和评估腐蚀损伤方面。     

Beamforming array techniques for acoustic emission monitoring of large concrete structures

This paper introduces a novel method of acoustic emission (AE) analysis which is particularly suited for field applications on large plate-like reinforced concrete structures, such as walls and bridge decks. Similar to phased-array signal processing techniques developed for other non-destructive evaluation methods, this technique adapts beamforming tools developed for passive sonar and seismological applications for use in AE source localization and signal discrimination analyses. Instead of relying on the relatively weak P-wave, this method uses the energy-rich Rayleigh wave and requires only a small array of 4–8 sensors. Tests on an in-service reinforced concrete structure demonstrate that the azimuth of an artificial AE source can be determined via this method for sources located up to 3.8 m from the sensor array, even when the P-wave is undetectable. The beamforming array geometry also allows additional signal processing tools to be implemented, such as the VESPA process (VElocity SPectral Analysis), whereby the arrivals of different wave phases are identified by their apparent velocity of propagation. Beamforming AE can reduce sampling rate and time synchronization requirements between spatially distant sensors which in turn facilitates the use of wireless sensor networks for this application.     

Measurement of sound insulation of acoustic louvres by an impulse method

This paper presents results of a study of the sound attenuation of acoustic louvres. At the core of the study is an alternative method of measuring sound insulation, impulse response analysis, which circumvents the limitations imposed by standard and proposed standard methods. Using the impulse method, the sound transmission coefficient is measured at different angles of incidence and the angular dependency of transmission loss obtained. In the low frequency range, the transmission is governed by a mass layer effect. The value of transmission loss is independent of angle of incidence. For the mid and high frequencies, diffraction, interference and absorption determine louvre performance and an angular dependency is observed. The transmission at the angle of incidence, corresponding to a line-of-sight through the louvre blades, is the dominant contribution to the angle average value and a single measurement at the pitch of the louvre approximates the overall transmission loss. For the case considered, the geometry of the blades has little influence on the transmission at low frequencies and the mass of the blades has little influence at higher frequencies. In a companion paper, the impulse data are used to predict the insertion loss provided by the louvre when installed in a plant room.