Physics-Informed Machine Learning in Machining Process Modeling

Abstract

Machining processes such as turning, milling, drilling, and grinding are complex manufacturing operations governed by nonlinear interactions among tool, workpiece, machine structure, and cutting environment. Accurate modeling of these processes is essential for improving productivity, surface integrity, tool life, and energy efficiency. Traditional physics-based models rely on analytical or numerical formulations derived from mechanics, thermodynamics, and material science, but they often suffer from simplifying assumptions and limited adaptability to real-world variations. On the other hand, purely data-driven machine learning (ML) models demonstrate strong predictive capability but lack physical interpretability and generalization outside the training domain. Physics-informed machine learning (PIML) has emerged as a promising paradigm that integrates domain knowledge in the form of physical laws, constraints, and governing equations with data-driven learning techniques. This paper presents a comprehensive review of physicsinformed machine learning approaches applied to machining process modeling. The fundamental principles of machining physics are discussed, followed by an overview of conventional modeling techniques and data-driven methods. Various PIML frameworks, including physics-informed neural networks, hybrid modeling, and constrained learning, are analyzed with respect to cutting force prediction, tool wear estimation, chatter detection, temperature modeling, and surface quality prediction. Current challenges, practical implementation issues, and future research directions are also highlighted. The review indicates thatphysics-informed machine learning provides a balanced approach, combining accuracy, robustness, and physical consistency, making it suitable for nextgeneration smart machining systems.

KEYWORDS: Physics-informed machine learning, machining process modeling, cutting forces, tool wear, hybrid models, smart manufacturing