Composite Tow Layup Evaluation for NC Program Tolerance Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge in the fabrication of composite parts, such as Carbon Fiber Reinforced Polymer (CFRP), lies in determining the most beneficial Numerical Control (NC) programs for fiber placement machines, as existing methods require extensive trial and error due to the limitless number of possible programs, leading to inefficiencies and potential out-of-tolerance conditions, especially for complex and large parts like aircraft wings.
Innovation Solution
The use of machine learning models, specifically trained neural networks, to predictively identify fabrication discrepancies in new NC programs by analyzing tow placement information and measurements from previous programs, allowing for the determination of likelihoods of out-of-tolerance conditions on a localized basis, thereby facilitating the selection of optimal NC programs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multiple NC programs are tested through extensive fabrication trials to determine the most beneficial program, then manufacturing precision and reliability improve, but productivity and time consumption deteriorate significantly
Solution Approach 1:
The system performs preliminary analysis of NC programs using a trained neural network model before actual fabrication occurs. The model predicts fabrication discrepancies and out-of-tolerance conditions by analyzing tow placement information from the NC program against patterns learned from historical fabrication data, allowing selection of optimal NC programs without extensive trial fabrication
Solution Approach 2:
The system uses measurements and data from previously fabricated laminates as training data to create a neural network model that copies the knowledge and patterns from historical successful fabrications. This digital copy enables prediction and evaluation of new NC programs without physical trial runs
2Manufacturing precision
If the number of NC programs evaluated through fabrication trials is increased to ensure optimal selection, then manufacturing precision improves, but device complexity and process complexity increase
Solution Approach 1:
The neural network model serves as an intermediary between the NC program evaluation and physical fabrication. It acts as a virtual test bed that predicts fabrication outcomes by processing tow placement information and comparing it against historical patterns, eliminating the need for complex trial fabrication processes
3Measurement precision
If traditional trial-and-error methods are used to evaluate NC programs, then adaptability to new designs is reduced, but measurement precision and fabrication control improve through direct empirical data
Solution Approach 1:
The system implements feedback by using measurements from previously fabricated laminates as training data for the neural network model. This feedback loop allows the model to learn from historical empirical data while adapting to new NC programs and design variations, combining the benefits of empirical measurement with computational adaptability
Data Source
AI summary
Systems and methods are provided for facilitating fabrication of a composite part. An illustrative method includes loading a Numerical Control (NC) program that directs layup of tows by a fiber placement machine to create a laminate for curing into a composite part, identifying tow information recited in the NC program, applying inputs based on the tow information to a neural network that has been trained with measurements describing tow placement within other laminates that have been laid-up, and reporting likelihood of a fabrication discrepancy that is out of tolerance, based on an output of the neural network.