Thread Winding Tension Waveform Detection Using Autoencoders

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for detecting thread tension abnormalities in winding devices are inadequate in accurately identifying various patterns of tension deviations beyond average values and standard deviations.

Innovation Solution

A tension abnormality detection device utilizing an autoencoder trained on normal tension waveforms to detect abnormalities by comparing input and output waveforms, allowing for real-time detection and identification of abnormality types through machine learning and similarity analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If only average value and standard deviation of tension are used for detection, then the detection method is simple, but various patterns of tension abnormalities cannot be accurately detected

Engineering Contradiction:
Improvedetection method complexityVSAvoidabnormality detection accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent transitions from scalar tension magnitude analysis to waveform pattern analysis by comparing entire tension waveforms. The autoencoder model processes temporal patterns and shape characteristics of tension waveforms, adding dimensional information about tension variations over time rather than relying solely on average values and standard deviations.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent replaces traditional statistical mechanical analysis (average value and standard deviation calculations) with a machine learning-based autoencoder model. This neural network model automatically learns and extracts features from tension waveforms, substituting manual feature engineering with automated pattern recognition.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If autoencoder with machine learning is used for detection, then various tension abnormalities can be accurately detected, but the device complexity increases

Engineering Contradiction:
Improveabnormality detection accuracyVSAvoiddetection system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The autoencoder model performs self-service by automatically learning optimal feature representations from training data without requiring manual feature engineering. The model trains itself to compress normal tension waveforms and reconstruct them, automatically identifying patterns that distinguish normal from abnormal tension states.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system performs preliminary training action by pre-training the autoencoder model with normal tension waveform data before actual abnormality detection. This preliminary learning phase enables the model to establish a baseline of normal tension patterns, which is essential for subsequent accurate abnormality detection.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4385928B1Tension abnormality detection device, tension abnormality detection method, and tension abnormality detection program
Publication Date: 2026.01.14 TMT MACHINERY INC
  • EP4385928B1 patent drawingFigure 1
  • EP4385928B1 patent drawingFigure 2
  • EP4385928B1 patent drawingFigure 3

AI summary

Provided is a tension abnormality detection device (10) capable of accurately detecting various abnormalities related to tension of a thread wound by a thread winding machine (1), with a simple method. The tension abnormality detection device (10) includes a control device (50) and a tension sensor (13) for detecting the tension of the thread wound by the thread winding machine (1). The control device (50) executes processing for acquiring an autoencoder (124) trained to compress a normal tension waveform and then restore the normal tension waveform, and processing for detecting abnormality in the tension of the thread wound by the thread winding machine (1), based on a degree of similarity between an input tension waveform obtained from the tension sensor (13) and an output tension waveform obtained by inputting the input tension waveform to the autoencoder (124).