Machine Learning-Enabled Prognostication of Tensile Strength in 316L Stainless Steel Through Additive Manufacturing...

Area of Science:

• Materials Science
• Mechanical Engineering
• Additive Manufacturing

Background:

• Predicting the tensile strength of additively manufactured components is crucial for engineering applications.
• Process parameters in additive manufacturing (AM) make tensile strength prediction challenging.

Purpose of the Study:

• To develop a synergistic model integrating Convolutional Neural Networks (CNN) and Random Forest (RF) for predicting the tensile strength of 316L stainless steel components fabricated via Selective Laser Melting (SLM).
• To evaluate the model's performance against experimental data and compare it with existing methods.

Main Methods:

• Developed a predictive model using a combination of CNN and RF algorithms.
• Trained the model on 42 datasets and validated it using 12 experimental datasets.
• Quantified predictive performance using Mean Squared Error (MSE) and Mean Absolute Error (MAE).

Main Results:

• The integrated CNN-RF model achieved MSE of 0.00295 and MAE of 0.0344.
• This represents a significant improvement over CNN alone, with reduced MSE (3.28%) and MAE (31.88%).
• Achieved a high correlation coefficient of 0.9576, indicating accurate predictions, even with small datasets.

Conclusions:

• The synergistic CNN-RF model offers high-precision tensile strength prediction for SLM-fabricated 316L stainless steel, outperforming CNN alone.
• The model demonstrates effectiveness even with limited data, providing a framework for predicting mechanical properties across various materials.
• Adding relative density and Vickers hardness did not improve prediction accuracy with the same sample size.