Numerical analysis on the air conditioning system of the experimental hall at TPS

Taiwan Photon Source (TPS) had delivered the first synchrotron light on the last day of 2014. Installation of 16 beamlines of the first and second phases of TPS beamline project was completed. The third phase project also was launched in 2021. To confront the situation that the experimental hall is more compact, we performed Computational Fluid Dynamic (CFD) simulation to analyse the effects of the air conditioning system and various heat sources to the temperature and flow fields in the experimental hall.


Introduction
The air conditioning system plays a critical role in a stable operation of precision experimental facilities.Stable temperature and humidity is not only conducive to the experiment conducting, but provides a comfortable environment for experimenters.On the other hand, nowadays more people care about environment, energy conservation is an important issue.The experimental hall of TPS is large annular space, with about 87 m in inner radius, 104 m in outer radius and 10.7 m in height.For such a large space, the air conditioning system consumes huge power about 1 MW.Computational Fluid Dynamic (CFD) began from the early 30s of the 20th century to solve the linearized potential equations with 2D methods [1].As rapid development of numerical analysis and computer science, CFD has the advantage of reduction of time and cost.We had performed the CFD to simulate the temperature and flow field of the storage ring of Taiwan Light Source [2].

Numerical simulation
In this study, we performed CFD of 3D steady state numerical simulation using a commercial code ANSYS Fluent R.16.2 [3].To save computing memory and time, we simulate one twelfth of the TPS experimental hall, with 48 symmetric columns, as shown in Figure 1.There are three areas: A, B and C in the space.Area A, the inner ring, is where control instrumentation and air handling units are located.Area B is where the storage ring tunnel with wedge wall is located.Area C, the outer ring, is where beam-lines are located.In this study, we generate three virtual planes I, II and III to exam the simulation results.Plane I vertically and radially spans areas A, B and C passing through eight air exits.Plane II vertically spans areas B and C, as shown in Figure 1.Plane III is the horizontal plane at height of z = 1.0 M, which is not shown in the figure.Both Plane II and III respectively indicate the vertical and horizontal cross section areas where people sit or work around.

Governing equation
The basic governing equations include the continuity equation, the momentum equation and the energy equation.We apply the k-ε turbulence model [4] and SIMPLEC to solve the velocity and pressure problem.Mass conservation equation (continuity equation) where ρ is density of fluid (air in the study), t is time and u refers to air velocity vector.Momentum conservation equation where p is pressure, g is vector of gravitational acceleration, μ is dynamic viscosity of air, and τt is divergence of the turbulent stresses which accounts for auxiliary stress due to velocity fluctuations.Energy conservation equation where e is the specific internal energy, T is air temperature, k is heat conductivity, h is the specific enthalpy of fluid, j is the mass flux.

Geometry and Grid Generation
A detailed 3D model of one twelfth of experimental hall of TPS is set for the numerical simulation.The height of this space is 10.7 m.The volume of the simulation zone is about 12,313 m 3 .In this model, the air exits and exhausts are respectively located on the ceiling and area C, the outer ring.The lights, and air exits are distributed on the ceiling.According to the geometry of the model, we applied hybrid grid to discretize the model.The total number of the grid elements was 3,897,434.Figure 2 shows the generated grids near an air exit and a light.

Boundary Conditions
In the simulated area, there are 64 air exits and 8 air exhausts distributed on the ceiling and the outer ring, respectively.The boundary conditions are set according to actual conditions.The air exit velocity is 2 m/s normal to the ceiling.The air exit temperature is 20 o C. Air exhausts are set as outflow.Heat flux of some heat sources such as daylight windows, lights, cable trays, ceiling, hutch walls, panels, and PC screens are set between 60 W/m 2 to 30 W/m 2 .

Results and discussion
Figure 3 shows the air velocity (m/s) vectors on the Plane I.It clearly shows air velocity vectors at the air exits and one air exhaust.According to the continuity equation, the air velocities at air exhausts are larger than air velocities at air exits because the total area of air exit is larger than that of air exhaust.For better resolution, the upper temperature limit is set at 25 o C. In this figure, area in red color indicates temperature is equal or higher than 25 o C. For example, cable trays and panels attached on beam lines are also heat source, the temperature near the beam lines is higher than 25 o C. Figure 5 shows contours of temperature on the Plane II.As Plane II passes fewer air exits than Plane I, heat flux from ceiling is clearer than that in Figure 4. Air temperature near the beam exit is higher.Because cold air is supplied through air exits on the ceiling and most heat sources are near ground, cold air is in the upper zone, as shown in both Figures 4 and 5.   Figure 7 shows contours of temperature on the Plane III.It is clear that air temperature near beamlines is higher than other area.Especially in the area between two beamlines.On the other hand, there are few heat sources and the cold air from air exits flows to air exhausts in the left area, the temperature is lower.In this study case, the air exit velocity is set 2 m/s, as actual case.We also simulate the second case with air exit velocity set 10 m/s to examine the cooling effect.Figure 8 shows contours of temperature on the Planes I with air exit velocity 10 m/s.Please note that the air temperature distribution is set between 20 and 21 o C. In other words, air temperature in the whole area is cooled down in this case.Figure 9 showing contours of temperature on the Planes II with air exit velocity is set 10 m/s.Likewise, this figure is similar to Figure 5.Although high air exit velocity can enhance air conditioning effect, the fan of air handling unit (AHU) also consumes about double power.This is upper limit of the fan.However, noise and vibration through the AHU will be another issue.We also collect the actual temperature in the experimental hall.The collected temperature is higher about 2 o C than the simulated results on the Plane III.Some heat sources, such as some equipment and personnel.Besides, will try to modify the air exits layout closer the work areas to obtain better efficient cooling effect.Therefore, we will model and simulate the new case in near future for our design information.

Conclusion
We performed CFD simulation to analyse two cases of temperature and flow fields in the TPS experimental hall.The simulated temperature is lower than actual collected results.The simulated temperature variation at different location still helps us to improve our air conditioning system.

Acknowledgments
Authors would like to thank colleagues in the civil and utility and group of NSRRC for their assistance.

Figure 1 .
Figure 1.3D schematic view of one twelfth of the TPS experimental hall.

Figure 2 .
Figure 2. Generated grids near an air exit and a light.

Figure 3 .
Figure 3. Air velocity vectors on the Plane I.

Figure 4
Figure 4 shows contours of temperature on the Plane I. Based on the boundary condition, the air temperature at exits on the ceiling is 20 o C. Air temperature variation with distance from air exits is clear.The steady state simulation results show the temperature near the daylight windows is about 22 or 23 o C. Heat flux from ceiling may be observed.For better resolution, the upper temperature limit is set at 25 o C. In this figure, area in red color indicates temperature is equal or higher than 25 o C. For example, cable trays and panels attached on beam lines are also heat source, the temperature near the beam lines is higher than 25 o C.

Figure 4 .
Figure 4. Contours of temperature on the Plane I.

Figure 5 .
Figure 5. Contours of temperature on the Plane II.

Figure 6
Figure 6 shows simulated path lines on the Plane II of Case 1. Path lines are virtual massless particles released from air exits to visualize the flow.The color in the figure identifies the particle number.The storage ring tunnel blocks path lines in the right area while most path lines in the left area flow to air exhausts, as shown in the figure.

Figure 6 .
Figure 6.Simulated path lines on the Plane II.

Figure 7 .
Figure 7. Contours of temperature on the Plane III.

Figure 8 .
Figure 8. Contours of temperature on the Planes I with air exit velocity is set 10 m/s.

Figure 9 .
Figure 9. Contours of temperature on the Planes II with air exit velocity 10 m/s.

Figure 10
Figure 10 shows contours of temperature on the Planes III with air exit velocity is set 10 m/s.In this figure, we keep the temperature distribution between 23 and 28 o C. Compared with Figure 7, air temperature in whole area decreases.Temperature gradient near two beam lines is clearer.

Figure 10 .
Figure 10.Contours of temperature on the Planes III with air exit velocity 10 m/s.