Prepreg Roll Quality Prediction for Real-Time Waste Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for testing unconsolidated composite components, such as prepreg rolls, are inefficient as they require testing multiple portions to identify out-of-tolerance sections, leading to time-consuming and costly discarding of entire rolls, and fail to determine the cause of quality issues in real-time, resulting in reduced manufacturing capacity.
Innovation Solution
A composite material management system that uses sensor data to predict properties of unconsolidated composite materials in real-time, allowing for identification of in-tolerance and out-of-tolerance portions during manufacturing, enabling corrective actions to be taken and reducing waste by marking or recycling only out-of-tolerance sections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional quality testing methods are used to inspect unconsolidated composite materials, then quality control is achieved, but manufacturing time increases and productivity decreases
Solution Approach 1:
The system performs real-time monitoring and prediction of quality parameters during the manufacturing process itself, rather than conducting separate post-manufacturing tests. Sensors collect data during fabrication and machine learning models predict final quality outcomes, enabling quality assurance to be integrated into the manufacturing workflow and eliminate separate testing phases
Solution Approach 2:
Traditional mechanical and manual quality testing methods are replaced with an automated system combining sensors, data processing, and machine learning algorithms. The system uses non-contact or minimal-contact sensing followed by computational analysis to predict quality parameters, substituting time-consuming physical testing with faster digital evaluation
2Measurement precision
If multiple portions of prepreg rolls are tested to identify out-of-tolerance sections, then quality assessment is improved, but time consumption and cost increase
Solution Approach 1:
The system creates a digital representation or 'copy' of the quality characteristics by collecting sensor data during manufacturing and using machine learning models to predict final quality parameters. This digital twin approach allows quality assessment without physically testing multiple portions of the material, reducing time and resource requirements while maintaining assessment accuracy
Solution Approach 2:
Instead of testing multiple discrete portions of the material, the system uses continuous sensor monitoring during manufacturing to capture sufficient data for predicting overall quality. The machine learning model processes this partial information during production to make accurate quality predictions, avoiding the need for extensive post-manufacturing sampling
3Reliability
If entire rolls are discarded due to quality issues in any portion, then quality standards are maintained, but material waste increases
Solution Approach 1:
The system divides the continuous prepreg roll into distinct segments or portions and evaluates the quality of each segment independently using sensor data and machine learning predictions. This segmentation enables identification of specific out-of-tolerance sections without compromising the entire roll, allowing only defective portions to be marked or discarded while preserving acceptable sections
Solution Approach 2:
The system applies different quality assessments and actions to different portions of the same material roll based on their specific predicted quality characteristics. Acceptable portions are marked for use, out-of-tolerance portions are marked for recycling or discard, and borderline portions may receive additional inspection, allowing each section to be treated according to its actual quality rather than applying a uniform standard to the entire roll
4Productivity
If real-time monitoring and prediction systems are implemented, then manufacturing efficiency increases, but device complexity increases
Solution Approach 1:
The system uses a multi-functional integrated platform that combines sensor data acquisition, real-time processing, machine learning prediction, and quality decision-making capabilities within a single system. This universal system handles multiple quality parameters and manufacturing scenarios using the same core architecture, reducing the need for separate specialized systems for each function and managing complexity through consolidation
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system, apparatus, computer program product, and a method for manufacturing an unconsolidated composite material. Sensor data is received from a sensor system for a composite material manufacturing system. The sensor data is received during manufacturing an unconsolidated composite material by the composite material manufacturing system. A set of predicted properties is determined for a number of portions of the unconsolidated composite material using the sensor data. The set of predicted properties is for the number of portions of the unconsolidated composite material as a completed product. A corrective action is performed based on a quality level for the number of portions of the unconsolidated composite material.