How it cools? Studying the heat flow out of a semi-infinite slab in welding: An analytical approach

Fawzi Aly, Alex Kitt, Luke Mohr · 2026

Read How it cools? Studying the heat flow out of a semi-infinite slab in welding: An analytical approach by Fawzi Aly, Alex Kitt, Luke Mohr (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-21250.txt.

Introduction

Fawzi Aly, Alex Kitt, and Luke Mohr derive closed-form cooling formulas for welding and additive manufacturing, where poor heat control leaves residual stress, warping, and cracks. Older Rosenthal solutions miss finite 3D geometry, cooling, and transient effects. Their framework adds Newton-cooling boundary rules and, via Laplace transforms and Fourier series proven equivalent, solves transient and steady heat patterns for Gaussian, ellipsoidal, and switching heat sources. The formulas match numerical results closely, cut compute and test costs, and can seed synthetic data for machine learning.

Catalog record

CREATORFawzi Aly, Alex Kitt, Luke Mohr
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE84399
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.21250

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

How it cools? Studying the heat flow out of a semi-infinite slab in welding: An analytical approach Fawzi Aly, Alex Kitt, Luke Mohr Abstract Additive manufacturing and welding processes are highly sensitive to heat dissipation, where improper thermal management leads to residual stresses, distortions, and cracking. Existing heat transfer models, such as Rosenthal’s solutions, fail to handle finite 3D geometries, cooling effects, or transient behavior, limiting their accuracy. We overcome these limitations by developing an analytical framework that incorporates cooling boundary conditions mimicking Newton’s Law of Cooling. Using two different and proven-equivalent approaches—Laplace transform and Fourier series—we derive closed-form solutions for transient and steady-state temperature profiles under various heat sources, including Gaussian, ellipsoidal, double-ellipsoidal, and time-dependent on/off switch sources. We compare our analytical solutions to numerical implementations, demonstrating strong agreement while providing deeper physical insight. This approach significantly reduces computational cost and experimental requirements, making it a scalable tool for optimizing thermal predictions and mitigating residual stresses in metal-based manufacturing. Additionally, our framework enables the generation of synthetic datasets for machine learning models to predict heat distribution efficiently. I Introduction Additive Manufacturing (AM) excels in applications requiring complex geometries, rapid prototyping, and low-volume production without the need for costly tooling. Its ability to create lightweight, optimized structures makes it ideal for aerospace, medical implants, and specialized industrial components Gibson et al. 2021 ; Ngo et al. 2018 . However, for AM to truly compete with traditional methods, advancements in material properties and defect predictability are essential DebRoy et al. 2018 ; Grasso and Colosimo 2017 ; Lewandowski and Seifi 2016 . In AM, the laser moves methodically, melting the powdered material and fusing it into a solid structure. The part appears flawless as it takes shape—but beneath the surface, something unseen is happening . As the material cools, uneven heat dissipation creates thermal gradients, leading to internal stresses; without proper thermal management, microscopic cracks begin to form, weakening the structural integrity of the part DebRoy et al. 2018 ; Mercelis and Kruth 2006 ; Lewandowski and Seifi 2016 ; Khairallah et al. 2016 . Effective cooling management is essential to minimize residual stresses, microstructural changes, and crack formation, ultimately improving joint quality and mechanical performance Mercelis and Kruth 2006 ; Vrancken 2016 ; Patterson, Messimer, and Farrington 2019 ; Parry, Ashcroft, and Wildman 2019 ; Khairallah et al. 2016 ; Ganeriwala and Zohdi 2014 ; Hussein et al. 2013 . Nevertheless, the main question remains: Can we develop more accurate physics models to predict how a material cools? Accurately capturing heat dissipation in AM and welding remains a challenge, requiring a combination of sophisticated analytical and numerical approaches. This study contributes to this ongoing effort by proposing a more refined framework for modeling heat dissipation. The evolution of the system temperature is governed by the diffusion-convection equation, which serves as the fundamental model of heat transfer in this study. To accurately capture cooling effects, one must account for both the energy input from a moving heat source and the heat loss to the environment. However, this process is highly complex due to factors such as the material’s geometry, the non-uniform spatial distribution of the heat source, and its power variation as function of time Rosenthal 1946 ; Eagar and Tsai 1983 ; Nguyen et al. 1999 . Traditional analytical approaches, as demonstrated in the literature by Rosenthal and others Rosenthal 1946 ; Eagar and Tsai 1983 ; Nguyen et al. 1999 , often assume an open-domain system to take advantage of Fourier integral transformations. This method simplifies the problem by reducing it to a convolution of the heat source with the heat kernel over time. While effective for simple cases, these models fail to capture realistic cooling effects due to their inability to incorporate finite-width constraints and evolving source distributions. As a result, many previous models have struggled to accurately represent cooling, often oversimplifying heat dissipation and its impact on temperature evolution Kou 2003 ; Goldak and Akhlaghi 2005 ; Dowden 2001 ; Pavelic et al. 1969 ; Michaleris 2014 . While numerical approaches, such as the finite-element method (FEM), can effectively handle some of these complexities, they are computationally expensive and often fail to provide clear physical intuition about the underlying processes Michaleris 2014 ; Frazier 2014 ; Khairallah et al. 2016 . Analytical approaches, on the other hand, offer a more efficient alternative, not only reducing computational cost and experimental effort but also providing deeper physical insight into heat transfer mechanisms Carslaw and Jaeger 1959 ; Michaleris 2014 ; Rosenthal 1946 . Additionally, they can be seamlessly integrated into physics-informed machine learning models, further enhancing predictive capabilities in AM and thermal analysis Raissi, Perdikaris, and Karniadakis 2019 ; Karniadakis et al. 2021 . In this article, we address the challenges of solving the heat equation under more realistic boundary conditions (B.C.). We also account for the transient behavior of system’s temperature profile by employing the Laplace transformation (LT), often rendering the equation more amenable to separation of variables, and naturally incorporates B.C. and I.C. Schiff 1999 ; Krantz 2022 . This approach is not merely an alternative analytical method but introduces more physically realistic conditions, such as cooling effects and finite-width constraints, which are often absent in standard open-domain formulations. The resulting system of equations remains analytically tractable while capturing essential thermal behaviors. Moreover, these refined equations can still be implemented numerically if needed, providing a more physically accurate alternative to conventional numerical solutions in literature. While this method is particularly well-suited for initial-boundary value problems (IBVPs), the inversion from spectral space s back to the real-time domain t remains a challenge, often involving complex contour integrals with branch cuts. Nevertheless, we manage to overcome this obstacle and obtain a closed-form series solutions for the heat kernel, as well as for specific heat sources such as Gaussian and ellipsoidal spatial profiles and other time-dependent variations. These series solutions are shown to be equivalent to Fourier series (FS) expansions, constrained by the initial data. Since FS are particularly effective for handling finite-domain B.C., whereas LT are more suited for IVPs, we propose a more rigorous approach that begins with LT and later establishes its equivalence to FS expansions. This dual formulation provides greater flexibility in implementing the final solution, as we will demonstrate in this work. Figure 1 : Comparison between heat profile evolution for open-domain and finite-width solutions with cooling. The dashed lines represent the open-domain solution obtained via Fourier integral methods, while the solid lines correspond to the finite-width solution with NLOC with parameters: H_{1}=1.2 and H_{2}=1.5 . Initially, both solutions coincide, indicating that heat didn’t diffuse to the material boundaries yet. This overlap validates that, at early timescales, boundary cooling effects are not on yet. However, as time progresses to a critical point, characterized by the material width w , the source location z_{0} , and the thermal diffusivity \alpha , heat begins to reach the boundaries, activating cooling effects. Consequently, a noticeable difference between the two solutions develops, with substantial deviations at the boundaries over time, as shown in the red and blue profiles. This illustrates the significant role of boundary cooling in finite domains at later times of evolution. I.1 Executive Summary This work

Read the complete text