Wavelet based de-noising for on-site partial discharge measurement signal

Performance comparison of some weak signal detection techniques is introduced. This comparison is very necessary since different applications require different operating conditions such as signal to noise ratio SNR, bandwidth, coherency, processing time and complexity. Three methods for detecting weak signals are considered. These are based on chaos theory, wavelet transform, and stocastic resonance. A detection algorithm based on a rectangular region in phase space plane is suggested in chaos method. The stocastic resonance method is considered in this research, as it is used for signal detection in underwater at a certain frequency. Simulation results obtained from MATLAB programs verify the studied methods giving an estimation of probability of detection and probability of false alarm versus SNR.

Section: Detection By Duffing-holmes's Oscillatormentioning

confidence: 99%

Performance comparison of some weak signal detection techniques

Mohammed

Mahmood

2022

“…The task of noise suppression within the PD data obtained can be carried out with various de-noising techniques. The techniques are known as Fast Fourier Transform (FFT) based de-noising, wavelet de-noising and low pass filtering [24][25][26][27].…”

Section: Partial Discharge Extractionmentioning

confidence: 99%

Partial discharge detection device using ultrasonic sensor on medium voltage XLPE cable

Radzi

Nik

Rameli

et al. 2020

There are many methods that have been studied by earlier researchers in order to detect the acoustic properties of Partial Discharge (PD) emitted by PD sources. One of the methods known as ultrasonic sensing on medium voltage Cross-Linked Polyethylene (XLPE) cable was adopted to detect partial discharges on commercial applications, usually by using Acoustic Emission (AE) sensors. This paper presents the processes of designing a PD sensor to detect the acoustic properties of the partial discharges on medium voltage XLPE cable. This PD sensor method works by detecting the partial discharges occurred at the joint of the cable which can act as an early warning device to help minimize the repair and maintenance costs of degrading cable. Result of the experiment shows the complete design of the prototype device, the device after fabrication and the functionality of the device. This design of the prototype can be beneficial for future uses in designing cost efficient and smaller sized PD detection devices. By positioning the sensor in horizontal position directly to the source of PD on the cable, the sensor will be able to detect acoustic properties of PD, by picking up the frequencies beyond 40 kHz. By varying the voltage applied values, a design of experiment (DOE) is carried out accordingly. Result of the experiment shows that the prototype device is functioning as expected, and hence this finding will be very useful to the consumers of power industries as the sensor device can serve as an alternative device to the commercialized PD sensing devices which are bulky and expensive.

“…It is localized in both frequency domain and time domain and this technique can analyze the non-stationary signals [16]. In it, the signal is divided into orthogonal set of frequencies and flexibility increases in time-frequency analysis [17][18][19][20][21]. The Wavelet transform is classified into two types in terms of their operation, continuous wavelet (CWT) and discrete wavelet transforms (DWT)…”

Section: Wavelet Transformmentioning

confidence: 99%

Analysis of power system parameters for islanding detection using wavelet transform

Raza¹,

Amjad

Umer

et al. 2020

<p><span>This paper classifies the basic power system parameters based on sensitivity and performance capability. This is the basic step in selecting the suitable parameter for intelligent islanding detection technique. The behaviour of basic power system parameters is analyzed using wavelet transform under all possible islanding and non-islanding conditions. In wavelet transform, the mother wavelet plays a significant role in the extraction of parameter features. The db4 mother wavelet is selected after an extensive analysis on different mother wavelets. As a result, it has been found that the extracted feature of reactive power shows the highest capability to distinguish islanding from non-islanding events.</span></p>