Partial Discharge Detection in High-Voltage Cables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for monitoring partial discharges in high voltage cables face challenges in distinguishing signals from electric noise, locating the source of discharges along long cables, and identifying the severity of defects, particularly in environments with high levels of background noise and complex signal patterns.
Innovation Solution
The method employs Wavelet transform and statistical analysis to differentiate partial discharge signals from noise, uses synchronized captures from multiple sensors to locate discharges regardless of background noise, and applies pattern recognition through a neural network to identify defect types, enabling effective discrimination, localization, and diagnosis of partial discharge sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If filtering techniques are used to remove electric background noise, then noise reduction is achieved, but partial discharge signals in the filtered frequency range are also attenuated or removed
Solution Approach 1:
The patent segments the frequency spectrum into multiple bands and applies different processing techniques to each band. Instead of applying a single filter to the entire spectrum, the system divides the frequency range and processes each segment separately, allowing preservation of PD signals while removing noise in specific frequency ranges.
Solution Approach 2:
The patent changes the parameter of frequency selection dynamically. Instead of using fixed filtering, the system identifies frequency bands where PD signals are present and adjusts the filtering parameters accordingly, allowing adaptive noise reduction that preserves relevant signal components.
2Object-affected harmful factors
If measurement frequency band is chosen to minimize noise amplitude, then noise reduction is achieved, but PD signal amplitude may also be weak in that band
Solution Approach 1:
The patent transitions from single-frequency analysis to multi-frequency spectrum analysis. By examining the frequency domain representation of signals, the system can identify PD signals across multiple frequency bands and select optimal bands for measurement, adding the frequency dimension to the analysis.
Solution Approach 2:
The system dynamically selects measurement frequency bands based on real-time spectral analysis. When noise is high in certain bands, the system switches to other bands where both noise and PD signals are favorable, making the measurement parameters adaptive rather than fixed.
3Measurement precision
If signal acquisition level is reduced to capture PD signals, then PD detection sensitivity is improved, but noise signal content considerably increases
Solution Approach 1:
The patent extracts PD signals from the mixed signal containing both PD and noise components through spectral analysis. By transforming the time-domain signal to frequency domain, the system can separate and extract PD signal components from noise components, then reconstruct the cleaned signal.
Solution Approach 2:
The Fast Fourier Transform (FFT) acts as an intermediary that converts the mixed time-domain signal into frequency-domain representation. This transformation allows the system to identify and separate PD signals from noise in the frequency domain before converting back to time domain for analysis.
4Measurement precision
If multiple sensors are used to locate PD sources, then localization accuracy is improved, but system complexity increases
Solution Approach 1:
The patent segments the cable into multiple sections and places sensors at strategic locations corresponding to potential defect zones such as joints and terminations. This segmented approach allows localization of PD sources to specific cable sections without requiring continuous sensor coverage.
Solution Approach 2:
The system uses time-domain analysis of signal arrival at multiple sensors to calculate the spatial location of PD sources. By measuring the time difference of arrival at different sensor positions, the system can triangulate the source location, adding the time dimension to spatial localization.
Data Source
Figure 1~2b
Figure 3a~3b
Figure 4
AI summary
The invention relates to a method specially designed for detecting events associated with partial discharges (PDs) in high voltage cables, comprising the identification of the location and the evaluation of the amplitude and rate of repetition per period of the grid voltage, with the possibility of identifying different sources producing PD signals as a function of the location thereof and recognising the type of defect associated with PDs in the same location, including the measurement of the generated electric signals and the discrimination thereof in relation to the background noise. The invention also relates to a system for carrying out said method, comprising means for discriminating the noise in relation to the transient waveform of the PD, determining the parameters associated therewith, determining the map of sources of PDs along the length of the cable, graphically representing said sources, and identifying the patterns of the sources of PDs separated as a function of the location thereof along the length of the cable.