Docs

Convective Heat Transfer: UFAD Ventilated & Occupied Room

Overview

The goal of this case study is to validate the Convective Heat Transfer analysis type on the SimScale platform. Validation is performed against experimental data obtained by Chen et al. [1] who analyzed the particle transport and distribution in ventilated rooms. It is possible to use their results because it was concluded that those particles were a passive scalar and thus had no impact on the fluid behavior or the room temperature.

he setup consists of a room with an UFAD (Underfloor Air Distribution) ventilation system. The heat sources used were four human simulators and six ceiling lights. Cool air is fed to the room through the inlets on the floor while the humans and the lights heat up the surrounding air creating convection currents. Because the hottest air rises to the top of the room before cooling down and generating these convection currents, the exhaust is placed at the ceiling, thus ensuring that the hottest air is always leaving the room. The theory behind UFAD systems is that because of the setup previously explained, the air that is fed to the room can have a higher temperature while ensuring lower room temperatures, thus saving energy. Chen et al. [1] ran experiments to validate their own simulations for particle transport and distribution, while this study focuses solely on validating the CFD results.

Geometry

Fin1 Fin2

Fig.1. On the left is the geometry used for this case, on the right is the geometry provided by Chen et al. [1]]

Although the figure provided by Chen et al. [1] has a scale, the only details provided by the paper were the dimension of the room. Based on that and other schematic figures provided by Chen et al. [1] it was possible to estimate very closely the dimensions of all geometries: they are presented in table 1.

Table 1: Dimensions for all elements of the simulation.
Item \(x\) [m] \(y\) [m] \(z\) [m]
Room 4.8 4.2 2.4
Human Simulator 0.38 0.9 0.38
Inlets 0.25 N/A 0.25
Exhaust 0.25 N/A 0.25
Lamps 1.5 N/A 0.1

Analysis Type and Domain

Tool Type: OPENFOAM®

Analysis Type: Convective Heat Transfer (\(\kappa - \epsilon\) Turbulence Model)

Domain: The computational domain was that of the room with dimensions specified in table 1.

Mesh and Element types:

A hex-dominant mesh was generated with appropriate refinements on the lamps, human simulators, inlets, and exhaust surfaces, and an addition of boundary layer inflation on all surfaces to ensure the correct results of the thermal boundary layer. The mesh is shown on the following figures (2-4)

3-MeshZoomOut

Fig.2. Hex-dominant computational grid.

../../../_images/4-MeshExhaustCloseUp.png

Fig.3. Close-up of surface refinements for the exhaust.

../../../_images/5-HumanBoundaryLayer.png

Fig.4. Close-up of refinements for human simulators. Surface refinement + boundary layer inflation.

Simulation Setup

The simulation was a Convective Heat Transfer simulation with a \(\kappa - \epsilon\) Turbulence Model. The boundary conditions are specified below. Chen et al [1] provided the power generation for the human generators as a whole, so to run a simulation without a volumetric heat source this power was distributed accordingly to each face of the simulators.

Boussinesq approximation was used for this simulation as no significant temperature differences would be present.

Boundary Conditions

  • Inlet: constant volumetric flow at constant temperature
    • 0.0472 m3/s per outlet, at 293K
  • Outlet: constant pressure outlet
    • Gauge pressure = 0 Pa
  • Human Simulators (Top): No-slip walls with turbulent heat flux temperature
    • Power = 9.6W
  • Human Simulators (sides): No-slip walls with turbulent heat flux temperature
    • Power = 22.6W
  • Lamps: No-slip walls with turbulent heat flux temperature
    • Power = 64W
  • Walls: All walls were No-slip but each had a different surface temperature
    • Wall (+X): 297.7 K
    • Wall (-X): 298 K
    • Wall (+Z): 298.5 K
    • Wall (-Z): 298.3 K
    • Ceiling: 298.7 K
    • Floor: 297 K

Numerics

  • Relaxation factors:
    • Relaxation factor for field p_rgh = 0.7
    • Relaxation factor for equation U = 0.3
    • Relaxation factor for equation T = 0.5
    • Relaxation factor for equation k = 0.7
    • Relaxation factor for equation epsilon = 0.7
  • Divergence schemes:
    • Divergence scheme for div(phi,U) = Gauss linearUpwind limitedGrad
  • All other settings within numerics were left with the default selections/values.

Results

Chen et al. [1] probed seven lines across the middle of the room (at z=0) and for each line registred 7 points at y = {0.1, 0.3, 0.6 1.1, 1.4, 1.7, 2.2} m. The location of these seven lines was obtained from figure 5 (provided by Chen et al. [1]).

ProbeLineLocation

Fig.5. Probe line locations provided by Chen et al. [1]. The locations of interest for this study were V1 to V3.

Velocity

9-V1Velocity

Fig.6.Velocity results for probe line 1 (Blue = SimScale, Red = Chen et al. [1]).

9-V2Velocity

Fig.7. Velocity results for probe line 2.

9-V3Velocity

Fig.8. Velocity results for probe line 3.

A comparison of the velocity profiles reveals that the SimScale results follow the same trend as the experimental results provided by Chen et al. [1]. Due to the fact that many dimensions had to be estimated and tested, discrepancies were expected. Additionally, the results provided by Chen et al. [1] lack symmetry across the ZY plane, which should not be the case given that the geometry is symmetric. These errors in the experimental results add to the discrepancies between SimScale and Chen et al. [1]. So the most important thing is that both results follow the same trend.

Temperature

Results for velocity and temperature are shown next. Temperature was normalized using the following formula:

\[T_{norm} = \frac{ T - T_{supply} } { T_{exhaust} - T_{supply} }\]

With \(T_{supply}\) = 293K and \(T_{exhaust}\) = 299.18K

10-V1Temp

Fig.9. Normalized temperature results for probe line 1 (Blue = SimScale, Red = Chen et al. [1]).

11-V2Temp

Fig.10. Normalized temperature results for probe line 2.

12-V3Temp

Fig.11. Normalized temperature results for probe line 3.

A comparison of the temperature profiles reveals that the SimScale results follow the same trend as the experimental results provided by Chen et al. [1] even more closely than the velocity results. As explained above, an error was to be expected. Having such close results for the temperature indicates that the difference on the velocity profile comes from the estimated size of the inlet sizes which impact velocity, not temperature.

References

[1](1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14) Z. Zhang and Q. Chen, “Experimental measurements and numerical simulations of particle transport and distribution in ventilated rooms”, Atmospheric Environment, vol. 40, no. 18, pp. 3396-3408, 2006.

Disclaimer

This offering is not approved or endorsed by OpenCFD Limited, producer and distributor of the OpenFOAM software and owner of the OPENFOAM® and OpenCFD® trade marks. OPENFOAM® is a registered trade mark of OpenCFD Limited, producer and distributor of the OpenFOAM software.