2D differentially heated cavity

Simulation of natural convection in a differentially heated square cavity.

Industry
Energy
Physics
Heat transfer – Natural convection
Platform
Local workstation

2D differentially heated cavity

This video is hosted by YouTube. Loading it allows YouTube to receive technical connection data.

Objective

Demonstrate the model’s ability to reproduce a benchmark natural convection case in a differentially heated cavity, with coupled fluid flow and heat transfer.

Simulation setup

The computational domain is a two-dimensional square cavity. The left and right vertical walls are maintained at different temperatures, creating a horizontal thermal gradient that drives the convective motion. The horizontal walls are adiabatic. The simulation solves the coupled flow and heat transfer equations to capture the development and structure of the temperature field inside the cavity. This case is a classical benchmark for validating natural convection models.

Results

The simulation highlights a strongly unsteady natural convection regime, characteristic of a high Rayleigh number (Ra = 10¹⁰). After an initial conduction-dominated phase, the flow quickly becomes unsteady, with upward motion along the hot wall and downward motion along the cold wall.

The cavity core is strongly mixed. Thermal structures continuously deform, break up, and interact with each other. Thin thermal boundary layers develop along the hot and cold walls.

A clear transition toward a highly convective regime is observed. Thermal plumes become unsteady and actively contribute to mixing within the cavity.

Simulations performed for Rayleigh numbers ranging from 10³ to 10⁸ accurately reproduce the standard reference quantities of the differentially heated cavity. These results validate the model across a wide range of regimes, from conduction-dominated cases to well-established natural convection.

The table below compares Fugu predictions with the reference results of Dixit and Babu (2006) for the maximum horizontal velocity Umax and its location Y, the maximum vertical velocity Vmax and its location X, as well as the average Nusselt number Nu. The good agreement observed across all these indicators confirms the relevance of the model for this natural convection benchmark.

Maximum horizontal velocity Umax

RaRef.Fugu
10³3.65293.6462
10⁴16.16316.152
10⁵35.52134.738
10⁶64.18664.8294
10⁷148.58148.27
10⁸321.876319.310

Y position of Umax

RaRéf.Fugu
10³0.81250.8170
10⁴0.8280.818
10⁵0.85540.8540
10⁶0.84960.8500
10⁷0.8790.879
10⁸0.9280.929

Maximum vertical velocity Vmax

RaRéf.Fugu
10³3.6823.694
10⁴19.56919.604
10⁵68.65568.632
10⁶219.886220.558
10⁷699.236698.794
10⁸2222.392222.00

X position of Vmax

RaRéf.Fugu
10³0.171830.18300
10⁴0.1250.114
10⁵0.06640.0650
10⁶0.03710.0380
10⁷0.0210.021
10⁸0.0120.012

Average Nusselt number Nu

RaRéf.Fugu
10³1.12721.1177
10⁴2.2472.245
10⁵4.52264.5217
10⁶8.8058.825
10⁷16.52316.523
10⁸30.22530.226

Comparison between the reference values from Dixit and Babu (2006) and Fugu results for the 2D differentially heated cavity benchmark.

Additional resources