Composite Laminate Configuration Using Machine Learning Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The design of laminate configurations for composite material structures, particularly FRP pipes for optical observation satellites, is arduous due to the lack of a general relational expression depicting the relationship between demanded specifications and laminate configurations.
Innovation Solution
A method involving machine learning to derive a relational expression between physical property values and laminate configurations, using a database of theoretical, numerical, and actual measurement data to facilitate efficient design of composite material laminated structures, and a designing device to automate this process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a design assistance method using relational expressions is used, then design automation is improved, but the method is limited because general relational expressions are not available
Solution Approach 1:
The invention changes the approach from using fixed relational expressions to using machine learning models that can adapt to different parameters and structures. The system learns optimal laminate configurations by training on diverse datasets, enabling it to handle various composite material structures without requiring pre-defined relational expressions for each case.
Solution Approach 2:
The invention replaces the traditional mechanical/design-based approach of manually deriving relational expressions with an intelligent system using machine learning. This substitution allows the system to automatically learn complex relationships between laminate configurations and physical properties without requiring explicit mathematical formulations.
2Manufacturing precision
If manual designing of laminate configuration is performed, then design precision can be achieved, but the designing process becomes arduous and time-consuming
Solution Approach 1:
The invention performs preliminary actions by pre-training machine learning models on extensive datasets of laminate configurations and their physical properties. This pre-training enables the system to quickly provide accurate design recommendations without requiring time-consuming manual calculations or iterations during the actual design phase.
Solution Approach 2:
The invention creates a virtual model or copy of the design process through machine learning algorithms that simulate and learn from optimal laminate configurations. This digital copy allows rapid evaluation and optimization of design options without physical prototyping or extensive manual computation.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Provided is a method of designing a composite material laminated structure, the method including: a machine learning step of performing machine learning on a plurality of pieces of data each of which includes a pair of a physical property value of the composite material laminated structure and a laminate configuration of the composite material laminated structure, to obtain a relational expression depicting a relationship between the physical property value and the laminate configuration, the composite material laminated structure including a plurality of layers that are laminated; and a laminate configuration information calculation step of calculating, based on the relational expression and an objective value of the physical property value, laminate configuration information which is information of the laminate configuration that enables the objective value to be obtained.