Predicting Joining Strength of Dissimilar Materials Using Neural Networks

Solution Overview

Problem

The existing methods for joining dissimilar materials like plastic composite plates and steel plates in vehicle bodies are limited, requiring costly and time-consuming production and testing of various specimens to measure joining strengths, which becomes impractical with increasing diversity in material combinations and thicknesses.

Innovation Solution

A method using an artificial neural network model to predict joining strength by analyzing force-displacement data from specimens of dissimilar materials, allowing prediction without producing actual specimens, reducing time and cost by leveraging a prediction system connected through a network for inputting joining information and obtaining predicted force-displacement and joining strength values.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If specimens of various material combinations and thicknesses are produced and tested to measure joining strengths, then accurate joining strength data is obtained, but cost and time increase exponentially

Engineering Contradiction:

Improvejoining strength data accuracyVSAvoidspecimen production and testing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent creates a virtual copy of the specimen testing process through finite element analysis models. Instead of physically producing and testing numerous specimens for different material combinations and thicknesses, the invention uses computational models that replicate the mechanical behavior and joining strength characteristics, thereby eliminating the need for exhaustive physical specimen production while maintaining measurement accuracy

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces the mechanical specimen production and physical testing system with a computational simulation system. Finite element analysis models substitute for physical tensile testing apparatus, allowing joining strength prediction through numerical computation rather than physical experimentation, thus dramatically reducing time and resource requirements

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If specimens of various material combinations and thicknesses are produced and tested to measure joining strengths, then comprehensive material data is obtained, but cost increases exponentially

Engineering Contradiction:

Improvejoining strength data accuracyVSAvoidspecimen production cost

Core Design Contradiction:

Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent creates a virtual copy of the specimen testing process through finite element analysis models. Instead of physically producing and testing numerous specimens for different material combinations and thicknesses, the invention uses computational models that replicate the mechanical behavior and joining strength characteristics, thereby eliminating the need for exhaustive physical specimen production while maintaining measurement accuracy

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces the mechanical specimen production and physical testing system with a computational simulation system. Finite element analysis models substitute for physical tensile testing apparatus, allowing joining strength prediction through numerical computation rather than physical experimentation, thus dramatically reducing time and resource requirements

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If all individual material combinations are tested to ensure sufficient joining strength, then vehicle body strength requirements are met, but the process becomes impractical with diverse material combinations

Engineering Contradiction:

Improvevehicle body strengthVSAvoidmaterial selection efficiency

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent performs preliminary computational analysis through finite element models to predict joining strength for various material combinations before physical production. This preliminary virtual testing allows engineers to identify suitable material combinations and optimize designs without having to physically produce and test every possible combination, thereby ensuring strength requirements are met while maintaining high productivity

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent systematically varies key parameters such as material type, thickness, and joining method in finite element analysis models to predict their effect on joining strength. By changing these parameters computationally rather than physically, the invention can efficiently evaluate numerous material combinations and identify optimal solutions that meet vehicle body strength requirements without exhaustive physical testing

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11519835B2Method of predicting joining strength of joined dissimilar materials

Publication Date: 2022.12.06 HYUNDAI MOTOR CO LTD

AI summary

A method of predicting joining strength of joined dissimilar materials, includes performing a joining strength test on a plurality of specimens of joined dissimilar materials each having different joining information, and acquiring force-displacement data on a basis of the joining information; constructing, in a prediction system, an artificial neural network model for predicting the force-displacement data and the joining strengths from the joining information; learning the artificial neural network model by inputting the force-displacement data to the prediction system, the force-displacement data obtained through the joining strength test; inputting joining information to be predicted to the prediction system by using a computer running a software for performing prediction for the joining strength and connected to a host computer of the prediction system through a network; and predicting, by the learned artificial neural network model, force-displacement value and joining strength.