A Data-Efficient Sequential Learning Framework for Melt Pool Defect Classification in Laser Powder Bed Fusion

TL;DR

This paper introduces SL-RF+, a sequential learning framework that enhances melt pool defect classification in Laser Powder Bed Fusion by efficiently using limited data and improving accuracy through iterative sample selection and synthetic data augmentation.

Contribution

The study presents SL-RF+, a novel sequential learning approach combining random forests with advanced sampling techniques for defect classification in additive manufacturing.

Findings

SL-RF+ outperforms traditional models in accuracy and robustness.

Efficiently captures complex defect patterns with minimal labeled data.

Demonstrates potential for real-world, data-scarce applications.

Abstract

Ensuring the quality and reliability of Metal Additive Manufacturing (MAM) components is crucial, especially in the Laser Powder Bed Fusion (L-PBF) process, where melt pool defects such as keyhole, balling, and lack of fusion can significantly compromise structural integrity. This study presents SL-RF+ (Sequentially Learned Random Forest with Enhanced Sampling), a novel Sequential Learning (SL) framework for melt pool defect classification designed to maximize data efficiency and model accuracy in data-scarce environments. SL-RF+ utilizes RF classifier combined with Least Confidence Sampling (LCS) and Sobol sequence-based synthetic sampling to iteratively select the most informative samples to learn from, thereby refining the model's decision boundaries with minimal labeled data. Results show that SL-RF+ outperformed traditional machine learning models across key performance metrics, including accuracy, precision, recall, and F1 score, demonstrating significant robustness in identifying melt pool defects with limited data. This framework efficiently captures complex defect patterns by focusing on high-uncertainty regions in the process parameter space, ultimately achieving superior classification performance without the need for extensive labeled datasets. While this study utilizes pre-existing experimental data, SL-RF+ shows strong potential for real-world applications in pure sequential learning settings, where data is acquired and labeled incrementally, mitigating the high costs and time constraints of sample acquisition.