Autoencoder Feature Extraction for Composite Material Inference

Solution Overview

Problem

The materials engineering industry faces significant challenges in predicting composite characteristics from production recipes due to the non-linear relationship between material combinations and resulting properties, leading to costly and time-consuming trial-and-error processes.

Innovation Solution

A method and device utilizing autoencoders to extract features from latent spaces, enabling mutual inference between composite characteristics and production conditions through regressors, allowing for simulation and prediction of composite properties from given conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional trial-and-error methods are used to create new composites, then composite characteristics can be obtained through physical testing, but the process requires very large amounts of cost and time

Engineering Contradiction:

Improvecomposite characteristic measurementVSAvoiddevelopment time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent creates a virtual copy of the composite material system through digital twins and simulation models. Instead of physically testing every composite formulation, the system uses computational models to replicate material behavior and predict characteristics, dramatically reducing the need for physical prototypes and accelerated testing while maintaining measurement accuracy

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent performs preliminary computational screening and prediction of composite characteristics before physical manufacturing. By using machine learning models and simulation to pre-evaluate potential composite formulations, the system identifies promising candidates in advance, reducing the number of physical tests needed and accelerating the development timeline

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If traditional trial-and-error methods are used to create new composites, then composite characteristics can be obtained through physical testing, but the process requires very large amounts of cost

Engineering Contradiction:

Improvecomposite characteristic measurementVSAvoiddevelopment cost

Core Design Contradiction:

Measurement precisionVSLoss of energy

Solution Approach 1:

The patent creates a virtual copy of the composite material system through digital twins and simulation models. Instead of physically testing every composite formulation, the system uses computational models to replicate material behavior and predict characteristics, dramatically reducing the need for physical prototypes and accelerated testing while maintaining measurement accuracy

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent performs preliminary computational screening and prediction of composite characteristics before physical manufacturing. By using machine learning models and simulation to pre-evaluate potential composite formulations, the system identifies promising candidates in advance, reducing the number of physical tests needed and accelerating the development timeline

Inventive Principle:

Principle #10Preliminary action

3Productivity

If linear relationship assumptions are used to predict composite characteristics from production recipes, then predictions can be made quickly, but predictions cannot be made accurately because there is no linear relationship

Engineering Contradiction:

Improveprediction speedVSAvoidcharacteristic prediction accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent transforms the prediction approach by changing from linear parameter relationships to non-linear machine learning models. The system uses neural networks and other advanced algorithms that can capture complex non-linear interactions between production parameters and composite characteristics, maintaining computational efficiency while dramatically improving prediction accuracy

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces traditional mechanical/physical testing and linear mathematical models with computational intelligence and data-driven approaches. By substituting physical trial-and-error with virtual simulation and machine learning inference, the system achieves both speed and accuracy in predicting composite characteristics

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS20240256928A1Method and device for mutually inferring composite characteristics and composite production conditions through autoencoder feature extraction

Publication Date: 2024.08.01 SK PLANET CO LTD

AI summary

A method for mutually inferring a complex characteristic and a complex production condition through characteristic extraction of an autoencoder according to the present invention comprises the steps in which: during complex production in which a complex is produced using multiple materials, when there is an encoder trained to calculate a latent vector from a characteristic vector indicating a target complex characteristic, and a production condition vector indicating a production condition for expressing the complex characteristic is input to the encoder, an inference unit calculates, via a first regressor, a copied latent vector copying the latent vector from the input production condition vector; and the inference unit calculates, via a decoder, a copied characteristic vector copying the characteristic vector from the copied latent vector and an experimental condition vector indicating an experimental condition for measuring the complex characteristic.