Stator Winding Weld Inspection Using Neural Network Vision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual and traditional image processing methods are ineffective in assessing weld quality in electric motor stator windings due to surface appearance variations, leading to high false detection rates and inefficiencies in inspecting large numbers of welds during assembly line manufacturing.
Innovation Solution
A method using a camera, such as a 2D area or line scan camera, to acquire images of welds between adjacent electrical wires and analyze them with a neural network to determine defects like voids, cracks, and contamination, distinguishing between surface and defect discoloration, and generating alerts for defects found.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual visual inspection is used to assess weld quality, then inspection accuracy can be maintained for individual welds, but inspection speed and productivity are severely limited due to the large number of welds
Solution Approach 1:
The patent replaces manual visual inspection with an automated optical inspection system using a camera to capture weld images and a neural network to analyze them. This substitution of mechanical/manual inspection with an automated vision system enables simultaneous inspection of multiple welds, dramatically increasing productivity while maintaining assessment accuracy through algorithmic analysis of weld characteristics
Solution Approach 2:
The patent creates digital copies of weld surfaces through camera imaging, allowing the neural network to analyze multiple weld images simultaneously. This copying approach enables parallel processing of numerous welds without requiring physical contact or sequential manual examination, thereby boosting inspection speed while preserving detailed quality assessment capability
2Productivity
If traditional image processing methods are used to assess weld quality automatically, then inspection speed improves, but measurement precision deteriorates due to surface appearance variations causing high false detection rates
Solution Approach 1:
The patent transforms the inspection approach by changing from traditional image processing parameters to neural network-based feature extraction. The neural network learns optimal parameters and patterns from training data, enabling it to distinguish between normal surface variations and actual defects. This parameter transformation allows the system to maintain high inspection speed while achieving superior detection accuracy by adapting to specific weld appearance variations
Solution Approach 2:
The patent implements a feedback mechanism where the neural network is trained on labeled weld images with known defects and quality outcomes. This training feedback enables the system to learn from examples, continuously improving its ability to distinguish between acceptable surface variations and actual defects. The feedback loop ensures high detection accuracy is achieved while maintaining rapid inspection throughput
3Loss of energy
If audit basis inspection is used for assembly line manufacturing, then resource consumption is reduced, but reliability of weld quality assurance deteriorates due to insufficient coverage of all welds
Solution Approach 1:
The patent enables continuous inspection of welds on the assembly line by implementing an automated vision system that can process multiple welds in sequence without interruption. The system continuously captures images and analyzes weld quality, providing uninterrupted quality assurance coverage. This continuous operation ensures comprehensive weld inspection while consuming resources efficiently through automated processing rather than intermittent manual auditing
Solution Approach 2:
The patent implements a self-service inspection system where the automated vision and neural network analysis perform quality assessment without requiring human intervention for each weld. The system independently captures images, analyzes defect patterns, and generates quality determinations, enabling comprehensive weld coverage while minimizing human resource consumption. This self-service capability ensures reliable quality assurance across all welds with efficient resource utilization
Data Source
AI summary
A method of inspecting an electric motor includes scanning an electric motor stator winding with a 2D or 3D camera, acquiring one or more images of a plurality welds between adjacent electrical wires forming the stator winding using the 2D camera, analyzing the one or more acquired images with at least one neural network such that the neural network determines if at least one of the plurality of welds has a weld defect. The at least one neural network is trained and distinguishes between surface discoloration on a surface of the welds and defect discoloration resulting from contamination during welding. Also, the method inspects over 150 welds per electric motor stator winding moving along an assembly line.