Motor Winding Fault Detection via Sequence Component Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Winding faults in multi-phase motors, such as phase-to-ground (PG), phase-to-phase (PP), and inter-turn (IT) faults, lead to unbalanced motor performance and potential damage, necessitating effective detection and corrective measures.
Innovation Solution
A method and system for detecting winding faults by comparing sequence components of instantaneous currents in multiple sets of windings, using current sensors and a fault detector to identify differences exceeding predetermined values, and generating alarms or corrective actions based on fault severity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If winding faults are detected using traditional methods, then fault detection capability is limited, but system complexity increases with advanced detection methods
Solution Approach 1:
The patent segments the current signal into symmetrical and asymmetrical components using Fourier series decomposition. This allows independent analysis of different fault characteristics, improving detection accuracy while keeping the system structure relatively simple by processing existing current measurements through mathematical transformation rather than adding complex hardware
Solution Approach 2:
The patent introduces sequence components (positive, negative, zero sequence) as intermediary variables that mediate between the raw current measurements and fault detection. These components serve as intermediate representations that simplify the detection process by transforming complex multi-phase current analysis into comparative analysis of standardized components
2Adaptability or versatility
If multiple types of winding faults are distinguished, then diagnostic capability improves, but detection method complexity increases
Solution Approach 1:
The patent applies local quality analysis by examining specific characteristics of sequence components (magnitude relationships, phase angles) that are locally indicative of specific fault types. Different fault types produce distinct local patterns in the sequence component relationships, allowing differentiation without requiring a completely different detection system for each fault type
Solution Approach 2:
The patent utilizes parameter changes in the sequence components (positive sequence current magnitude, negative sequence current magnitude, zero sequence current magnitude) to identify different fault types. By monitoring how these parameters change and compare against each other, the system can distinguish between phase-to-ground faults, phase-to-phase faults, and inter-turn faults using the same detection framework
3Reliability
If fault detection is performed continuously, then reliability improves, but energy consumption increases
Solution Approach 1:
The patent implements periodic fault detection by continuously monitoring current sequence components during motor operation. The detection process operates periodically at each control cycle, analyzing current measurements to determine if fault conditions exist. This continuous periodic monitoring maintains high reliability while managing energy consumption by performing calculations only when needed rather than maintaining constant active detection states
Data Source
AI summary
Winding fault detection system and method is provided. A first instantaneous current in first set of windings of the multi-wound electrical device is received. A second instantaneous current in second set of windings of the multi-wound electrical device is received. First sequence components are determined from the first instantaneous current. Second sequence components are determined from the second instantaneous current. The first sequence components are compared with second sequence components. A winding fault in the multi-wound electrical device is determined when a difference in the first sequence components and second sequence components is greater than a predetermined value.


