Realizing the Potential of Physics-Informed Neural Network in Modelling Laser Drilling Process.

Saved in:

Bibliographic Details
Title:	Realizing the Potential of Physics-Informed Neural Network in Modelling Laser Drilling Process.
Authors:	Kheirandish, Zahra¹ zahra.kheirandish@nld.rwth-aachen.de, Schulz, Wolfgang^1,2
Source:	Journal of Laser Micro / Nanoengineering. Dec2024, Vol. 19 Issue 3, p209-213. 5p.
Subjects:	Artificial neural networks, Computational physics, Deep learning, Microfabrication, Laser drilling, Lasers
Abstract:	This research introduces a novel paradigm by integrating physics-based models with neural networks, forming a cohesive and adaptive framework for laser drilling. The Physics-Informed Neural Network (PINN) leverages the underlying physical principles governing laser-material interactions to enhance the predictive accuracy of drilling outcomes. Through meticulous training on diverse datasets encompassing material properties, laser parameters, and resultant hole geometries, the PINN model exhibits a remarkable ability to generalize across varied conditions. Here, the Artificial Neural Network's (ANN) inputs are spatial and temporal coordinates inside the domain, on the boundaries and at initial moment. The output is the shape of laser drilled hole. During the training step, the output values, and their derivatives w.r.t. inputs are calculated and afterwards they are used to evaluate the loss value. Through the gradient-based optimization method, the weights of network are changed to reach minimum loss value. The trained model predicts the unknown coefficient in the physical governing equation, which here is ablation threshold intensity. Our study investigates the effectiveness of the PINN approach in optimizing laser drilling parameters and achieving superior accuracy. We demonstrate its robustness in predicting drilling outcomes for different laser configurations. This research contributes not only to the advancement of laser precision microfabrication but also to the broader field of computational physics. By seamlessly integrating data-driven approaches with fundamental physical insights, the PINN methodology offers a transformative pathway to optimize laser drilling processes, paving the way for enhanced efficiency and quality in microfabrication applications. The insights presented herein promise to shape the future of laser microfabrication, fostering innovation and pushing the boundaries of precision engineering. [ABSTRACT FROM AUTHOR]
	Copyright of Journal of Laser Micro / Nanoengineering is the property of Japan Laser Processing Society and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database:	Engineering Source

Description
Abstract:	This research introduces a novel paradigm by integrating physics-based models with neural networks, forming a cohesive and adaptive framework for laser drilling. The Physics-Informed Neural Network (PINN) leverages the underlying physical principles governing laser-material interactions to enhance the predictive accuracy of drilling outcomes. Through meticulous training on diverse datasets encompassing material properties, laser parameters, and resultant hole geometries, the PINN model exhibits a remarkable ability to generalize across varied conditions. Here, the Artificial Neural Network's (ANN) inputs are spatial and temporal coordinates inside the domain, on the boundaries and at initial moment. The output is the shape of laser drilled hole. During the training step, the output values, and their derivatives w.r.t. inputs are calculated and afterwards they are used to evaluate the loss value. Through the gradient-based optimization method, the weights of network are changed to reach minimum loss value. The trained model predicts the unknown coefficient in the physical governing equation, which here is ablation threshold intensity. Our study investigates the effectiveness of the PINN approach in optimizing laser drilling parameters and achieving superior accuracy. We demonstrate its robustness in predicting drilling outcomes for different laser configurations. This research contributes not only to the advancement of laser precision microfabrication but also to the broader field of computational physics. By seamlessly integrating data-driven approaches with fundamental physical insights, the PINN methodology offers a transformative pathway to optimize laser drilling processes, paving the way for enhanced efficiency and quality in microfabrication applications. The insights presented herein promise to shape the future of laser microfabrication, fostering innovation and pushing the boundaries of precision engineering. [ABSTRACT FROM AUTHOR]
ISSN:	18800688
DOI:	10.2961/jlmn.2024.03.2006