Motor Eccentricity Prediction Using Topological Current Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting motor eccentricity faults, such as vibration analysis and motor current signature analysis, face challenges like noise interference, sensitivity variations, and the need for physical models and lengthy signal processing, making them inefficient and inaccurate for identifying fault-related features in stator current signals.
Innovation Solution
The use of topological data analysis (TDA) to extract fault-related features from time-domain motor current signals, representing them in persistence diagrams and vectorized Betti sequences, and applying machine learning models for predicting eccentricity fault levels without relying on physical models or extensive signal processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vibration analysis is used to detect eccentricity faults, then fault detection capability is improved, but noise interference from other sources increases measurement difficulty
Solution Approach 1:
The patent segments the complex vibration signal into specific frequency components using Fourier transform, isolating the characteristic frequencies related to eccentricity faults from the noisy background. This allows precise measurement of fault-related vibrations while filtering out unrelated noise sources.
Solution Approach 2:
The patent introduces signal processing techniques as an intermediary between the raw vibration signal and fault detection. The processing pipeline including Fourier transform and spectral analysis acts as a mediator that extracts meaningful fault information while suppressing noise interference.
2Measurement precision
If sensor location is optimized for vibration analysis, then measurement precision is improved, but device complexity increases due to multiple sensors
Solution Approach 1:
The patent develops a universal fault detection method based on current signal analysis that can be applied to various motor types and configurations without requiring specific sensor placements. The method uses standard current sensors already present in the system, making it universally applicable across different motor designs.
Solution Approach 2:
The patent utilizes the existing current sensors and control system of the motor drive as self-service resources for fault detection. No additional dedicated sensors or complex measurement systems are required, as the control system's current measurement capabilities are repurposed for eccentricity detection.
3Measurement precision
If physics-based models are used for fault detection, then measurement precision is improved, but loss of time increases due to lengthy signal processing
Solution Approach 1:
The patent extracts the essential fault detection functionality from complex physics-based models, retaining only the critical spectral analysis components needed for eccentricity detection. This extraction eliminates unnecessary computational steps while preserving the core measurement precision.
Solution Approach 2:
The patent applies partial signal processing by focusing only on the specific frequency bands and spectral features relevant to eccentricity faults, rather than performing complete and exhaustive signal analysis. This selective approach reduces processing time while maintaining adequate detection accuracy.
4Measurement precision
If conventional signal processing is used, then fault detection capability is improved, but device complexity increases due to requirement of physical models
Solution Approach 1:
The patent replaces physics-based mechanical models with data-driven signal processing approaches. Instead of requiring detailed knowledge of motor physics and complex analytical models, the method uses empirical spectral analysis of current signals to detect faults, simplifying the theoretical framework while maintaining detection capability.
Data Source
AI summary
A system and method for motor eccentricity fault detection is disclosed. The method includes extraction of fault-related features through topological data analysis (TDA) for motor current signals and apply them to motor eccentricity fault detection. The method further includes the procedure of obtaining topological features from time-domain data and representing them in persistence diagrams and vectorized Betti sequences. The method further includes the extraction of fault-related features from the obtained topological features of the data, which can be distinctively associated with not only fault type but also fault severity level. Further, the method includes use of machine learning models to extract fault-related features from TDA, for the prediction of motor eccentricity fault level, even for data from new eccentricity levels that are not seen in the training data.