Abnormality determination device, abnormality determination method, and abnormality determination program
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing yarn defect detection systems in synthetic fiber manufacturing face challenges in accurately determining abnormalities due to the narrow width and positional changes of yarns, leading to frequent misidentification of defects.
Innovation Solution
An abnormality determination device that utilizes thermal imaging and machine learning to analyze yarn path regions, generating determination models based on training data to accurately identify abnormalities such as yarn path deviations, tension issues, and oil application anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If visible image analysis is used to detect yarn abnormalities, then the detection system is simple to implement, but the detection accuracy is low due to yarn narrow width and positional changes
Solution Approach 1:
The patent replaces the visible light-based detection system with a thermal imaging system. Instead of using cameras that detect reflected visible light, the invention uses thermal cameras to detect infrared radiation emitted by the yarn. This substitution of the detection mechanism allows for accurate yarn detection despite narrow width and positional changes, as thermal imaging captures heat distribution patterns rather than relying on visual appearance.
Solution Approach 2:
The patent changes the detection parameter from visible light intensity to thermal radiation intensity. By measuring temperature distribution and heat patterns of the yarn rather than its visual characteristics, the system can accurately detect abnormalities even when the yarn is narrow or its position varies. This parameter change fundamentally improves detection capability.
2Measurement precision
If thermal imaging with machine learning is used to accurately detect yarn abnormalities, then the detection accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent introduces a machine learning model as an intermediary between thermal image acquisition and abnormality determination. The model processes complex thermal image data and extracts meaningful patterns, acting as a mediator that translates raw thermal data into accurate abnormality detection results. This intermediary handles the computational complexity, allowing the physical detection device to remain relatively simple while achieving high accuracy.
Solution Approach 2:
The patent uses thermal images as a copy or representation of the yarn's physical state. Instead of directly measuring yarn properties, the system creates thermal image copies that represent the yarn's heat distribution. These image copies can be analyzed without physically contacting or disturbing the yarn, enabling non-invasive high-accuracy detection.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the accuracy of detecting yarn defects by leveraging thermal images and machine learning, allowing for precise identification of abnormalities in yarn paths, tension, and oil application, thereby improving the quality of synthetic fiber production.
Implementation Method 1
an image capture device 6 serving as a thermal camera that captures an image of the yarn path region 5a including a part of the yarn guide 5
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided is an abnormality determination device capable of accurately determining an abnormality in a yarn. The control device includes an input unit (50e), a determination unit (50c), and an output unit (50f). A thermal image including a yarn path region that is a part of a first yarn guide (5a) included in the spinning take-up machine (1) is input to the input unit (50e). The determination unit (50c) determines whether or not there is an abnormality regarding the yarn (93) based on the thermal image input to the input unit (50e). The output unit (50f) outputs the determination result of the determination unit (50c).