Stator Varnish Fill Verification Using Deep Learning Replicas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for evaluating the condition of varnish coating on stator windings are inefficient, costly, and prone to variability due to the location of the coating within the stator slots, necessitating destructive techniques for accurate assessment.
Innovation Solution
A deep learning tool using convolutional neural networks (CNN) processes images of stator sections to estimate varnish fill percentages, which is verified using master verification tools comprising replicas with known varnish amounts to ensure accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional destructive techniques are used to evaluate varnish coating condition, then measurement precision is improved, but productivity deteriorates and loss of time increases
Solution Approach 1:
The patent uses digital image copies and deep learning models to create virtual representations of the varnish coating evaluation process. Instead of physically examining or destroying the actual stator, the system captures images and processes them through trained neural networks, allowing rapid assessment without touching the physical object.
Solution Approach 2:
The patent replaces mechanical destructive examination methods with an optical-digital system. Cameras capture images of the stator windings, and deep learning algorithms automatically analyze the varnish coating condition, substituting physical manipulation with automated image processing.
2Measurement precision
If conventional destructive techniques are used to evaluate varnish coating condition, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent performs preliminary training of deep learning models using labeled datasets of stator images with known varnish conditions. This pre-trained knowledge enables the system to rapidly evaluate new stators without time-consuming manual analysis, as the AI has already learned the patterns during the preliminary training phase.
Solution Approach 2:
The system creates digital copies of stator images and processes these copies through the deep learning model, eliminating the need for time-consuming physical examination of the actual stator. The digital copying and processing occur in seconds compared to hours of manual inspection.
3Device complexity
If manual visual inspection is used to evaluate varnish coating, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent replaces human visual inspection with an automated deep learning-based image analysis system. The convolutional neural networks automatically extract features and classify varnish coating conditions, providing consistent and objective measurements without human subjectivity or fatigue.
Solution Approach 2:
The deep learning model performs self-assessment of the stator images by automatically detecting varnish coating characteristics without requiring human intervention. The system serves itself by autonomously completing the entire evaluation process from image input to condition classification.
Data Source
AI summary
Methods and systems are provided for verifying a deep learning tool for evaluating a varnish condition of a stator. In one example, a method for verifying the deep learning tool includes receiving images of replicas of a stator section at a processor of a computing system, the replicas of the stator section having different predetermined varnish fill percentages. The images are process and analyzed by the deep learning tool to output estimated varnish fill percentages. The estimated varnish fill percentages may be compared to the predetermined varnish fill percentages and a notification recommending at least one of further training of the deep learning tool and adjustments to an imaging setup for acquiring the images.