🔬 physics

Flow and Heat Transfer Characteristics of Forced Convection Past an Isoflux Circular Cylinder in Galinstan for Reynolds Numbers up to 600

This paper presents a numerical investigation of steady forced convection heat transfer from an isoflux circular cylinder in liquid metal Galinstan ($Pr=0.025$) for Reynolds numbers up to 600, utilizing a high-order finite difference scheme to analyze flow and thermal characteristics and proposing a new empirical correlation for the average Nusselt number with high accuracy.

Original authors: Dipjyoti Nath

Published 2026-07-14

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Dipjyoti Nath

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a tiny, super-hot metal cylinder sitting in a river of liquid metal called Galinstan. This isn't your average river; Galinstan is a special alloy that conducts heat like a champion sprinter but flows with the stickiness of honey. Scientists wanted to know: if we push this liquid past the cylinder at different speeds, how does the heat escape, and how does the liquid swirl around it?

To find out, the researchers built a super-precise digital model. Think of it like upgrading a video game from "blocky pixels" to "ultra-realistic graphics." While most previous studies used a "standard definition" math method (second-order accuracy), this team used a "4K Ultra HD" method called a fourth-order compact finite difference scheme. They paired this with a clever trick called "pseudo-time iteration," which is like running a simulation in fast-forward until the swirling patterns settle down into a steady, predictable state.

The Main Discovery: Speed Changes the Dance
The team simulated the liquid flowing past the cylinder at speeds ranging from a slow crawl (Reynolds number 1) to a fast zoom (Reynolds number 600).

At Slow Speeds (Re = 1): The liquid metal hugs the cylinder tightly, like a blanket. There are no swirls or gaps behind it. The heat spreads out evenly in perfect circles, mostly because the liquid is so good at conducting heat on its own.
At Fast Speeds (Re = 200 to 600): As the liquid speeds up, it can't keep up with the curve of the cylinder. It peels away, creating a "wake"—a chaotic, swirling tail behind the cylinder, much like the wake behind a boat. The faster it goes, the bigger this wake gets.

Here is the cool part: even though the liquid is moving fast, the heat doesn't stay trapped in a thin layer like it would in water or air. Because Galinstan is so good at conducting heat, the warmth spreads out widely into the fluid, creating a broad, diffuse thermal cloud. However, the speed does help push that heat downstream, making the heat transfer much more efficient than when the liquid was moving slowly.

The Numbers and the Rules
The researchers didn't just guess; they ran thousands of calculations to prove their method works. They checked their math against older, trusted studies for drag (the force pushing back on the cylinder) and heat transfer, and their results matched almost perfectly.

They found that as the Reynolds number goes up, the average heat transfer (called the Nusselt number) goes up too. But there's a catch: the more you speed it up, the less extra heat you get for your effort. It's like running faster and faster; eventually, you get tired, and running twice as fast doesn't make you twice as efficient.

To make this useful for engineers, the team created a new "recipe" (a mathematical formula) to predict exactly how much heat will be transferred for any speed between 1 and 600. This recipe is incredibly accurate, matching their simulation data with a score of 0.99939 out of 1.0.

What They Didn't Find
It's important to note what this study didn't do. The paper explicitly states that while real-world liquid metal flow eventually becomes chaotic and wobbly (unsteady) at high speeds, this specific study only looked at the "steady" version of the flow. They stopped their simulation at a Reynolds number of 600, so they didn't explore what happens if the liquid goes even faster and starts to vibrate wildly. They also didn't test other shapes or different types of liquid metals; this was strictly about a round cylinder in Galinstan.

The Bottom Line
In these simulations, the team showed that using a high-precision math method allows us to see exactly how heat moves in liquid metals. They proved that speeding up the flow helps pull heat away from the cylinder, but the liquid's super-high ability to conduct heat means the heat spreads out broadly rather than staying in a thin layer. The new formula they created is a reliable tool for anyone designing systems that use this special liquid metal, provided the flow stays steady and doesn't exceed a Reynolds number of 600.

Technical Summary: Flow and Heat Transfer Characteristics of Forced Convection Past an Isoflux Circular Cylinder in Galinstan for Reynolds Numbers up to 600

Problem Statement
This study presents a numerical investigation into steady forced convection heat transfer from a circular cylinder subjected to a constant heat flux (isoflux) boundary condition. The fluid medium is Galinstan, a liquid metal alloy characterized by a low Prandtl number ($Pr = 0.025$) and exceptionally high thermal conductivity. The research focuses on the hydrodynamic and thermal behavior of the flow over a Reynolds number ($Re$) range of $1 \le Re \le 600$ . While previous literature has extensively covered isothermal cylinders and moderate-to-high Prandtl number fluids, this work addresses the specific challenges of simulating low-Prandtl-number liquid metals under isoflux conditions, a regime relevant to advanced thermal management systems such as electronic cooling and liquid-metal-based energy systems.

Methodology
The governing equations—comprising the stream-function, vorticity, and energy equations in cylindrical coordinates—are solved using a fourth-order compact finite difference scheme (FOCS–CC). To ensure stability and convergence, this spatial discretization is coupled with a pseudo-time iteration (PTI) technique.

Key methodological features include:

Coordinate Transformation: The physical domain is mapped to a computational domain using the transformation $r = e^{\pi\lambda}$ and $\theta = \pi\eta$ , allowing for efficient grid resolution near the cylinder surface.
High-Order Accuracy: The scheme utilizes fourth-order compact formulations for spatial derivatives. A notable feature highlighted by the authors is the ease of implementation; the method extends existing second-order frameworks by incorporating specific correction coefficients ( $\alpha, b_1, \Gamma_1$ , etc.) without altering the underlying numerical structure.
Solution Algorithm: The resulting system of equations is integrated using a factorized approach to reduce large-banded matrices into tridiagonal systems, which are solved efficiently using the Tri-Diagonal Matrix Algorithm (TDMA).
Validation: The numerical code is validated through grid independence studies and by comparing the computed average Nusselt number ($Nu$) and total drag coefficient ( $C_D$ ) against established numerical results from Dennis and Chang (1970), Fornberg (1980), Ahmad (1996), and Morgan (1975).

Key Results and Observations
The study systematically analyzes the flow and thermal fields across the specified Reynolds number range:

Flow Structure: At low Reynolds numbers ($Re=1$), the flow remains fully attached to the cylinder surface with no separation, dominated by viscous forces. As $Re$ increases to 200 and 600, flow separation occurs, forming a wake region behind the cylinder. The size of this wake expands with increasing $Re$, indicating a growing dominance of inertial forces.
Thermal Field: For $Re=1$, isotherms form nearly concentric rings, signifying that heat transfer is governed primarily by thermal diffusion. As $Re$ increases, convective transport intensifies, causing isotherms to pack more closely at the upstream stagnation point and stretch downstream, forming a thermal plume. Despite this convective stretching, the thermal field remains relatively diffuse compared to conventional fluids due to Galinstan's high thermal diffusivity.
Heat Transfer Metrics: The local Nusselt number exhibits minimal angular variation at low $Re$ but shows significant variation at higher $Re$ due to the influence of the separated wake. The average Nusselt number increases monotonically with $Re$ throughout the investigated range, though the rate of enhancement diminishes at higher Reynolds numbers.
Empirical Correlation: Based on the numerical data, a new empirical correlation for the average Nusselt number is proposed for Galinstan over $1 \le Re \le 600$ :
$Nu = \alpha + \beta\sqrt{Re} + \gamma Re$
where $\alpha = 0.26133$ , $\beta = 0.13302$ , and $\gamma = -5.42758 \times 10^{-4}$ . This correlation demonstrates excellent agreement with the numerical data, yielding a coefficient of determination ( $R^2$ ) of 0.99939.

Significance and Claims
The authors claim that the proposed FOCS–CC scheme, integrated with the PTI technique, provides a robust and accurate numerical framework for simulating forced convection around circular cylinders in low-Prandtl-number liquid metals. The study establishes that:

Steady solutions for the incompressible Navier–Stokes equations can be successfully computed up to $Re = 600$ for this configuration, extending beyond the onset of unsteady flow instability.
The interplay between enhanced convective transport and dominant thermal diffusion is a defining characteristic of Galinstan, distinguishing its thermal behavior from fluids with moderate or high Prandtl numbers.
The high accuracy of the fourth-order compact scheme allows for reliable prediction of both hydrodynamic (drag) and thermal (Nusselt number) characteristics, as confirmed by validation against existing literature.

The paper concludes that the developed correlation and numerical method offer valuable tools for analyzing heat transfer in liquid metal applications, specifically highlighting the diminishing returns in convective heat transfer enhancement at higher Reynolds numbers due to the persistent influence of high thermal diffusivity.

More like this