Related Experiment Video
Updated: Jun 27, 2025
10:10
Three-Dimensional Particle Shape Analysis Using X-ray Computed Tomography: Experimental Procedure and Analysis Algorithms for Metal Powders
Published on: December 4, 2020
1.8K
Decision Support Tool in the Selection of Powder for 3D Printing
Ewelina Szczupak1, Marcin Małysza1,2, Dorota Wilk-Kołodziejczyk1,2
1Faculty of Metals Engineering and Industrial Computer Science, AGH University of Krakow, al. Mickiewicza 30, 30-059 Kraków, Poland.
Materials (Basel, Switzerland)
|April 27, 2024
Summary
This study introduces an AI-powered tool for selecting 3D printing powders, including steel, nickel, cobalt, and aluminum. The Random Forest algorithm proved most effective, achieving high accuracy in predicting mechanical parameters.
Area of Science:
- Materials Science and Engineering
- Additive Manufacturing
- Artificial Intelligence in Materials Selection
Background:
- Selecting appropriate powders is critical for successful 3D printing.
- Mechanical property prediction is essential for material selection in additive manufacturing.
- Various artificial intelligence algorithms can be applied to material selection problems.
Purpose of the Study:
- To develop a decision support tool for selecting 3D printing powders.
- To evaluate the performance of different AI algorithms for predicting mechanical parameters.
- To identify the most effective AI model for this application.
Main Methods:
- Focus on steel, nickel-based, cobalt-based, and aluminum-based powders.
- Implementation and comparison of AI algorithms: Random Forest, Decision Tree, K-Nearest Neighbors, Fuzzy K-Nearest Neighbors, Gradient Boosting, XGBoost, and AdaBoost.
- Utilized cross-validation and hyperparameter tuning for model optimization.
Main Results:
- Random Forest demonstrated superior performance.
- Achieved an F1 score of 98.66% with cross-validation.
- Attained an F1 score of 99.10% after hyperparameter tuning on the test set.
Conclusions:
- The Random Forest model is highly promising for a 3D printing powder selection decision support system.
- The optimized model provides accurate predictions for new data.
- Hyperparameter tuning can further enhance model performance, though overtraining is a consideration.
