Electrical Machine Fault Diagnostics Under Dynamic Load Oscillations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electrical signature analysis methods for detecting broken rotor bar, eccentricity, and bearing failures in motors fail when dynamic load oscillations occur, as they assume constant amplitude ranges and do not account for phase shifts between fault and load oscillation components.
Innovation Solution
The proposed systems and methods utilize a hybrid decision-making process that monitors the ratio and angle differences of magnitude features, independent of amplitude, and employ various algorithms such as real and reactive power calculations, voltage space vector orientation, rotor flux orientation, and machine learning techniques to detect faults under dynamic load conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing electrical signature analysis methods are used for fault detection, then fault detection can be performed under normal conditions, but detection reliability deteriorates when dynamic load oscillations occur
Solution Approach 1:
The patent transforms the fault detection approach by changing from monitoring absolute magnitude values to monitoring the ratio of magnitudes and angle differences. This parameter transformation makes the detection method invariant to load oscillation amplitude changes, thereby maintaining reliability under dynamic load conditions while adapting to varying operational states
Solution Approach 2:
The patent introduces intermediate parameters (magnitude ratios and angle differences) that serve as mediators between the raw electrical signals and fault detection decisions. These intermediate parameters filter out the harmful effect of load oscillations while preserving fault information, enabling reliable detection across different load conditions
2Measurement precision
If magnitude-based fault detection methods are used, then simple detection can be performed, but detection accuracy deteriorates when load oscillation frequencies match fault frequencies
Solution Approach 1:
The patent moves the detection from a one-dimensional magnitude-based approach to a two-dimensional approach incorporating both magnitude ratios and angle differences. This dimensional expansion allows the method to distinguish between load oscillation effects and actual fault signatures, improving measurement precision without requiring complex additional hardware
3Reliability
If constant amplitude assumptions are made in fault detection, then simplified analysis can be performed, but detection reliability deteriorates under varying load conditions
Solution Approach 1:
The patent transitions from static amplitude-based detection to a dynamic approach using angle differences that naturally adapt to varying load conditions. The angle difference parameter remains characteristic of faults regardless of load amplitude changes, providing both reliability and adaptability to dynamic operational states
Data Source
AI summary
Systems and methods are disclosed for improved fault diagnostics of electrical machines under dynamic load oscillations. The systems and methods may rely on one or more different algorithms for performing such fault diagnostics. One example, algorithm may involve determining a ratio of an instantaneous real power and a reactive power of the motor.