Synthetic Composite Inspection Data from 3D Defect Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for training classification models for composite manufacturing inspection face challenges due to the scarcity of inconsistency data, leading to inefficiencies in resource usage, increased manufacturing time, and material waste.
Innovation Solution
Generate synthetic data by creating a three-dimensional model of a composite layup, introducing inconsistencies, and simulating camera positions and orientations to produce training images with specific colorations, which are then used to train an inconsistency identifier program.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real inconsistency data is collected for training classification models, then model accuracy is improved, but data collection time and resource consumption increase
Solution Approach 1:
The patent creates synthetic 3D models of composite materials with intentionally introduced inconsistencies (defects, misalignments, voids) to generate training data. These virtual copies replicate real inspection scenarios without requiring physical defective parts, thereby reducing data collection time while maintaining model training quality.
Solution Approach 2:
The system pre-generates diverse inconsistency types and their corresponding image data before actual inspection operations. By preparing training data in advance through simulation, the patent eliminates the need for time-consuming post-production data collection and enables faster model deployment.
2Measurement precision
If more inconsistency data is collected to improve model quality, then classification accuracy improves, but manufacturing time increases
Solution Approach 1:
The patent replaces physical data collection processes with computational simulation. Instead of physically creating and inspecting defective composite parts to gather training data, the system uses 3D modeling and rendering to generate synthetic inspection images, thereby reducing manufacturing time while maintaining comprehensive training data coverage.
Solution Approach 2:
The system varies multiple parameters (defect type, position, size, material properties, inspection conditions) within the 3D model to generate diverse training data efficiently. This parameter-driven approach allows unlimited data variation without additional physical manufacturing, improving both accuracy and productivity.
3Reliability
If physical inconsistency data is gathered for training, then model reliability improves, but material waste increases
Solution Approach 1:
The patent creates virtual copies of composite material structures with embedded inconsistencies instead of physically producing defective parts for training. This eliminates material waste associated with creating and discarding defective composite components while maintaining reliable training data for model development.
Solution Approach 2:
The system converts the harmful effect of material waste into a beneficial training approach by using computational methods to simulate defective conditions. The 'harm' of needing physical defective data is transformed into the advantage of being able to generate unlimited synthetic defect variations without material loss.
4Measurement precision
If comprehensive inspection data is collected, then detection capability improves, but resource consumption increases
Solution Approach 1:
The patent substitutes physical inspection resources (cameras, sensors, processing equipment) with computational rendering systems. The energy-intensive physical data collection process is replaced by software-based 3D modeling and image synthesis, achieving comprehensive detection capability with significantly reduced resource consumption.
Solution Approach 2:
The 3D modeling system serves multiple functions: it creates training data, simulates various inspection types, generates augmented reality overlays, and prepares manufacturing guidance documents. This multi-functionality reduces overall resource consumption by consolidating multiple processes into a single computational platform.
Data Source
AI summary
A method of generating synthetic data for composite manufacturing inspection comprising creating a three-dimensional model of a composite layup; generating an inconsistency in the three-dimensional model of the composite layup; setting a camera position and a camera orientation in the three-dimensional model space; generating a captured image of the three-dimensional model and the inconsistency from the camera position and the camera orientation; and creating a secondary image from the captured image, the secondary image having colorations of a preselected inspection type.