Induction Motor Eccentricity Estimation Using Sparse Regression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for eccentricity detection in induction motors are inadequate for accurately estimating the severity of eccentricity, which is crucial for effective maintenance and control, due to uncertainties in measurement noise and relationships, leading to inaccurate or impractical severity level estimation.
Innovation Solution
A sparsity-driven regression model using machine learning with physics-informed constraints to estimate eccentricity severity by learning weights from sensor data, reducing noise accumulation and preserving statistically stable relationships between measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional measurement-based models are used to estimate eccentricity severity, then the model can capture dynamic relationships between measurements and eccentricity levels, but the uncertainty in measurement noise and relationships leads to inaccurate severity estimation
Solution Approach 1:
The patent introduces a sparsity-driven regression model as an intermediary between traditional measurement-based models and eccentricity severity estimation. This model acts as a mediator that filters out noise and uncertainty from multiple measurements while preserving the essential relationship between operating parameters and eccentricity levels, thereby improving estimation accuracy without sacrificing reliability under varying conditions
Solution Approach 2:
The patent transforms the estimation problem by changing the parameters from direct measurement values to sparsity-driven regression coefficients. By representing the system state through a sparse combination of basis functions with learned coefficients, the model captures dynamic relationships while being robust to measurement noise and uncertainties in operating conditions
2Loss of information
If multiple measurements of operating parameters are combined to estimate eccentricity severity, then more information is available for estimation, but noise from multiple measurements accumulates and reduces signal-to-noise ratio
Solution Approach 1:
The patent extracts only the essential information from multiple measurements by using a sparsity-driven approach. The regression model identifies and extracts the most relevant features and relationships while discarding noisy components, thereby utilizing maximum useful information while minimizing noise accumulation in the severity estimation
Solution Approach 2:
The patent segments the estimation process into distinct components: basis functions representing different operating conditions, sparsity coefficients weighting their contributions, and a regression framework combining them. This segmentation allows selective combination of measurements based on their actual informational value rather than simple aggregation, preventing noise accumulation
3Adaptability or versatility
If a data-driven model is used to capture relationships between measurements and eccentricity levels, then the model can adapt to different operating conditions, but the unknown relationships between measurements under different loads reduce estimation accuracy
Solution Approach 1:
The patent implements a dynamic sparsity-driven regression model where the sparsity coefficients are not fixed but adapt to different operating conditions. The model dynamically adjusts the weighting of different basis functions based on the current operating state, allowing it to capture unknown relationships between measurements and eccentricity levels under varying loads while maintaining high estimation accuracy
Solution Approach 2:
The patent allows the model parameters (sparsity coefficients) to change dynamically with operating conditions. By representing the system through a flexible basis function expansion with adaptive coefficients, the model can adapt to different load conditions and capture unknown relationships without sacrificing precision, as the parameters are learned from data rather than assumed fixed
Data Source
AI summary
A fault detection system of eccentricity severity of an induction machine including a rotor and stator is provided. The fault detection system includes a sensor interface configured to acquire sensor signals from sensors arranged at predetermined positions of the induction machine, wherein the sensor signals are indicative of an eccentricity level of a rotor of the induction machine, a memory coupled with a processor. The memory stores training data sets and instructions implementing a learning-based fault detection method for the induction machine. The instructions include steps of generating an eccentricity feature matrix from the sensor signals, where in the sensor signals include load torque, rotor speed, vibration acceleration of the rotor, vibration speed of the rotor, and current spectral of the stator or the induction machine, determining an eccentricity level of the induction machine based on the eccentricity feature matrix using the learning-based fault detection method, wherein the learning-based fault detection method is configured to find the eccentricity level from learning-based eccentricity feature matrix data sets.