Partial Discharge Progression Detection Using Temporal Differentials
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing partial discharge measurement methods fail to accurately determine the degree of insulation degradation in underground power transmission cables due to the loss of temporal information and increased neural network complexity, leading to potential misdiagnosis and reduced accuracy.
Innovation Solution
A partial discharge determination apparatus and method that incorporates temporal information by generating differential data from phase-resolved partial discharge patterns, using a neural network to classify the degree of partial discharge progression, including a clamp type high-frequency current transformer, a partial discharge determination apparatus, and a cable degradation monitoring system to analyze partial discharge pulses and determine the degree of insulation degradation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the phase angle and charge quantity are collectively patterned as occurrence frequency for multiple cycles, then the pattern can be simplified for analysis, but temporal information of the partial discharge is lost leading to erroneous determination
Solution Approach 1:
The patent segments the phase angle range into multiple discrete bins (e.g., 0-10 degrees, 10-20 degrees, etc.) and separately segments the charge quantity into multiple levels. This segmentation allows temporal information to be preserved in the time series data while maintaining analytical simplicity through standardized pattern representation.
Solution Approach 2:
The patent transforms the temporal information into a new dimension by creating time series data that tracks the standardized pattern across multiple cycles. Instead of losing temporal information in a single aggregated frequency count, the invention adds a time dimension to the pattern analysis, allowing both simplicity and temporal resolution to coexist.
2Measurement precision
If the neural network processes detailed time series data with temporal information, then determination accuracy improves, but the neural network complexity and computational load increase
Solution Approach 1:
The patent performs preliminary standardization of the phase angle and charge quantity data before feeding it to the neural network. By pre-processing the data into standardized patterns with defined bins and levels, the system reduces the complexity of the input data structure while preserving essential temporal information, thereby simplifying the neural network's processing requirements.
Solution Approach 2:
The patent changes the parameters of the input data by standardizing the phase angle into discrete bins and charge quantity into discrete levels across time cycles. This parameter transformation reduces the continuous data space into a manageable discrete structure that maintains accuracy information while reducing computational complexity for the neural network.
3Quantity of substance
If more cycles are used for pattern analysis, then more temporal information becomes available, but the loss of temporal causality in collective patterning increases
Solution Approach 1:
The patent maintains continuity of useful action by processing time series data that continuously tracks the standardized phase angle and charge quantity patterns across multiple cycles. This continuous time series representation preserves temporal causality while utilizing data from multiple cycles, allowing the system to benefit from increased data quantity without losing the sequential relationships between cycles.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Accurately determines the degree of partial discharge progression in underground power transmission cables, improving diagnosis accuracy by incorporating temporal information and enhancing the neural network's ability to classify insulation degradation states.
Implementation Method 1
a clamp type high-frequency current transformer
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A distribution pattern of a combination of a charge quantity and an occurrence phase angle of each of the partial discharges occurring in one or a plurality of cycle periods of an applied voltage of the power transmission cable is generated, differential data including a difference between the numbers of occurrences of the partial discharge for each combination of the charge quantity and the occurrence phase angle in two or more latest distribution patterns is generated, and the degree of progress of the partial discharge is determined based on data of the latest distribution patterns and the differential data.