Microgrid Weak-Signal Fault Identification Using DWT and VMD
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional fault protection schemes in inverter-based microgrids are unreliable for detecting weak-signal faults such as high impedance faults and inverter-related faults, which can pose safety threats and are not adequately addressed by existing methods like wavelet analysis, power line communication, and machine learning approaches.
Innovation Solution
A combined method using discrete wavelet transform (DWT) for denoising, variational mode decomposition (VMD) for fault detection, correlation matrix with K-nearest neighbors (KNN) for localization, and logic circuit for classification, leveraging intelligent electronic devices (IEDs) to identify and isolate weak-signal faults in islanded microgrids.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional overcurrent relays are used for fault detection, then the protection scheme is simple and easy to implement, but the relays become unreliable for weak-signal faults and cannot detect high impedance faults or inverter-related faults
Solution Approach 1:
The patent combines multiple signal processing techniques (wavelet transform for denoising, VMD for feature extraction, correlation matrix for localization, and logic circuit for classification) into an integrated fault identification system. This merging of techniques enables reliable detection of weak-signal faults while maintaining a systematic approach to fault protection in inverter-based microgrids.
2Reliability
If wavelet analysis is used for weak-signal fault detection, then fault detection capability is improved, but the selection of proper mother wavelet function becomes critical and complex
Solution Approach 1:
The patent introduces variational mode decomposition (VMD) as an intermediary technique that processes the denoised signals from wavelet transform. VMD automatically extracts intrinsic mode functions without requiring manual selection of wavelet parameters, thereby maintaining the fault detection capability while eliminating the complexity of mother wavelet function selection.
3Measurement precision
If power line communication devices are used for fault location, then fault location accuracy is improved, but the availability of PLC devices is limited and increases system complexity
Solution Approach 1:
The patent enables the existing intelligent electronic devices (IEDs) to perform fault location functions by computing correlation matrices from their own voltage and current measurements. The IEDs self-determine fault locations using signal correlation analysis without requiring additional power line communication devices, thereby achieving accurate fault location while avoiding increased hardware complexity.
4Measurement precision
If machine learning approaches are used for fault identification, then fault detection accuracy is improved, but computational complexity increases significantly
Solution Approach 1:
The patent segments the fault identification process into distinct functional stages: denoising (wavelet transform), feature extraction (VMD), localization (correlation matrix), and classification (logic circuit). Each stage processes data in a simplified manner appropriate to its function, avoiding the need for complex end-to-end machine learning models while achieving high identification accuracy through systematic decomposition of the problem.
Data Source
AI summary
Disclosed is a method and system for identifying an existence, location and type of a weak-signal fault in an islanded inverter-based microgrid. The weak-signal fault includes a high impedance fault, an inverter DC-side short-circuit fault, and an inverter tripping fault, and usually fails to be detected by conventional relay methods due to small magnitude of fault current. Upon received voltage and current measurements from intelligent electronic devices installed in the microgrid, the variation mode decomposition algorithm is firstly applied to detect the existence of fault based on denoised time series of measurements using discrete wavelet transform algorithm. After detecting the presence of fault, the correlation-based matrix is applied to locate the suspicious fault locations, and then K-nearest neighbors model is utilized to identify the faulty branch among those locations using dynamic time warping algorithm to measure the distance between neighbors. Following fault localization, fault classification is done by observing sequence components and phasor measurements and feeding the observational inputs to a fault classification logic circuit model.