Session: 01-05-01: Applications of Artificial Intelligence/Machine Learning for Aerospace Structures 1

Paper Number: 152228

152228 - Machine Learning Modeling to Predict Mechanical Properties of Nylon Parts Fabricated via 3D Printing 

The aerospace industry consistently seeks lightweight and high-strength materials to enhance performance and fuel efficiency. Thermoplastic composites have emerged as promising candidates due to their ease of processing, cost-effectiveness, ability to fabricate complex geometries, and recyclability. The adoption of 3D printing has further facilitated rapid prototyping and the production of these composites with customizable properties. However, the mechanical properties, such as strain at break (%), ultimate tensile strength (MPa), and Young's modulus (MPa), are significantly influenced by 3D printing parameters. Therefore, proper control of the processing parameters is crucial for obtaining composites parts with improved mechanical properties. Today, machine learning approach has been employed to accurately forecast mechanical properties of composites materials without performing experimental investigations, which considerably reduced time and cost for manufacturing the composites materials.

This study investigates the impact of 3D printing parameters, including layer thickness, infill density, nanofiller composition, and print speed, on the mechanical properties of Nylon-Carbon Fiber (PA/CF) composites. Besides, four machine learning (ML) algorithms, including linear regression, support vector machines (SVM), Gaussian process regression (GPR), and boosted trees, were used to predict the mechanical properties of the composites materials. The models were trained using layer thickness, infill density, CF composition, and print speed as input features, while the mechanical properties (strain at break, ultimate tensile strength, and Young's modulus) served as the output predictions. Furthermore, to showcase the accuracy and efficiency of the proposed models in forecasting mechanical properties of PA/CF composites materials, the results of these approaches are compared using metrics functions, such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE).  

Optimizing the 3D printing parameters using the proposed ML models can offer a pathway to manufacturing 3D printing thermoplastic composites materials with improved mechanical properties, making them valuable tools for optimizing thermoplastic composites design in the aerospace sector.  

Presenting Author: Mahbub Ahmed Southern Arkansas University

Presenting Author Biography: Dr. Mabub Ahmed received his B.S. in Mechanical Engineering from Bangladesh University of Engineering & Technology (Dhaka, Bangladesh) in July 1997, and received his M.S. in Industrial Engineering from Lamar University (Beaumont, TX) in May 2001. He then completed his Ph.D. in Materials Science and Engineering with an emphasis in Mechanical Engineering at the University of Texas (El Paso, TX) in 2008. He spent two years working at a company in El Paso as a design engineer, where he created solid models, animations, and photo renderings of light fixtures. He taught at Georgia Southern University from 2008 to 2011 as a visiting assistant professor and from 2011 to 2012 as a lecturer in the department of Mechanical Engineering. He has worked part-time as a visiting faculty member at the University of Texas at El Paso in numerous summer semesters. Dr. Ahmed joined the SAU Engineering Program in August 2012, and his education and experience qualify him to teach a wide variety of courses in engineering, engineering physics, engineering technology, and industrial technology. Dr. Ahmed also does some consulting work for local industries, and has completed several years of consulting work with local aerospace company Amfuel. Dr. Ahmed currently holds a PE license in mechanical engineering in the state of Arkansas.