Accurate predictions of keyhole depths using machine learning-aided simulations

2024-02-25

arXiv.org

AI Summary

To address the low accuracy and poor generalizability of dynamic keyhole depth prediction in laser processing, this study proposes a physics-informed machine learning framework. We introduce, for the first time, a closed-loop modeling architecture that compensates for real-time laser absorptivity deficits—overcoming long-standing accuracy limitations of conventional thermo-fluid simulations. The framework integrates multi-scale thermo-hydrodynamic and phase-change simulations, synchrotron X-ray experimental calibration, and a lightweight temporal regression neural network. Validated on titanium and aluminum alloys, it achieves ±10% prediction error in keyhole depth—significantly improving upon existing models’ 50–200% errors. The model features minimal parameters, rapid training, and strong transferability across materials and process conditions. This work establishes a new paradigm for high-fidelity, computationally efficient dynamic keyhole prediction.

Technology Category

• Machine Learning: Hardware-aware ML
• Computer Vision: Low Level & Physics-based Vision
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

The keyhole phenomenon is widely observed in laser materials processing, including laser welding, remelting, cladding, drilling, and additive manufacturing. Keyhole-induced defects, primarily pores, dramatically affect the performance of final products, impeding the broad use of these laser-based technologies. The formation of these pores is typically associated with the dynamic behavior of the keyhole. So far, the accurate characterization and prediction of keyhole features, particularly keyhole depth, as a function of time has been a challenging task. In situ characterization of keyhole dynamic behavior using a synchrotron X-ray is complicated and expensive. Current simulations are hindered by their poor accuracies in predicting keyhole depths due to the lack of real-time laser absorptance data. Here, we develop a machine learning-aided simulation method that allows us to accurately predict keyhole depth over a wide range of processing parameters. Based on titanium and aluminum alloys, two commonly used engineering materials as examples, we achieve an accuracy with an error margin of 10 %, surpassing those simulated using other existing models (with an error margin in a range of 50-200 %). Our machine learning-aided simulation method is affordable and readily deployable for a large variety of materials, opening new doors to eliminate or reduce defects for a wide range of laser materials processing techniques.

Problem

• Predict keyhole dynamics in laser materials processing accurately
• Reduce keyhole-induced defects like pores in final products
• Overcome poor accuracy and high cost of current simulations

Innovation

• Machine learning-aided simulation for keyhole dynamics
• Accurate keyhole depth prediction with 10% error
• Cost-effective alternative to synchrotron X-ray experiments