- Email: [email protected]
A numerical airflow pattern study of a floor swirl diffuser for UFAD system
Accepted Manuscript Title: A numerical airflow pattern study of a floor swirl diffuser for UFAD system Authors: Y.H. Yau, K.S. Poh, A. Badarudin PII: DOI: Reference:
S0378-7788(17)30590-X https://doi.org/10.1016/j.enbuild.2017.10.037 ENB 8057
To appear in:
ENB
Received date: Revised date: Accepted date:
18-2-2017 24-9-2017 10-10-2017
Please cite this article as: Y.H.Yau, K.S.Poh, A.Badarudin, A numerical airflow pattern study of a floor swirl diffuser for UFAD system, Energy and Buildings https://doi.org/10.1016/j.enbuild.2017.10.037 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A numerical airflow pattern study of a floor swirl diffuser for UFAD system Y.H. Yau, K.S. Poh and A. Badarudin Department of Mechanical Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia email: [email protected]
Research Highlights
This present study is to examine the effect of geometrical design of a floor swirl diffuser. Laboratory experiment was carried out to validate both CFD and turbulence models. The results could be used as an important guide to design an optimised UFAD system. The optimum airflow rate per diffuser is required in the design of UFAD system to achieve thermal stratification.
Abstract Floor swirl diffuser is one of the commonly used diffusers in underfloor air distribution (UFAD) system. Geometrical design of the floor swirl diffuser will affect the airflow pattern of the indoor environment. Thus, the purpose of this study is to examine the effect of geometrical design of a floor swirl diffuser on the aspects of airflow rate, number of diffuser blades, angle of attack of diffuser blades, and the grille thickness. Laboratory experiment was carried out to validate CFD simulation results. The results of this study have shown that diffuser blades have an important role in the airflow pattern. The grille thickness, which is required to withstand the impact and the load of human and furniture, needs to be minimized as the thickness is detrimental to the function of diffuser blades. Besides, airflow rate per diffuser is studied in UFAD system in order to achieve thermal stratification. The air throw is lower compared to a high angle of attack at the low angles of attack of 30° and 45°. The optimum number of diffuser blades to create the swirl effect is 18 blades with a free area of 18.86%. Keywords Underfloor air distribution (UFAD); Floor swirl diffuser; Computational fluid dynamics (CFD); Airflow pattern; Thermal stratification Nomenclatures CFD HVAC OHAD UFAD VSD VAV
Computational fluid dynamics Heating, ventilating and air conditioning Overhead air distribution Underfloor air distribution Variable speed drive Variable air volume
CMH mm m/s m Vc
cubic meter per hour millimeter meter per second meter centerline velocity
1.
Introduction
UFAD diffusers are used to deliver conditioned air from the floor plenum to the occupants' space for human comfort in UFAD system. Generally, UFAD diffusers are categorized into six types: round omnidirectional diffusers, selective directional diffusers, displacement flow diffusers, linear diffusers, ducted diffusers and terminal units [1]. Although there are various UFAD diffusers available in the market since the emerging of underfloor air distribution technology in 1990s, there is limited research work carried out on the geometry of UFAD diffusers. The present study examines the geometrical effect of the floor swirl diffuser in terms of the airflow pattern. A well-designed UFAD system can have a potential to reduce energy consumption, reduce floor to floor height, improve thermal comfort, improve ventilation efficiency and provide flexibility in reconfiguring the interior layout of the building [2]. The performance of UFAD system has been examined from the aspects of fluid dynamics, energy performance, temperature distribution and thermal stratification [3-6]. However, one of the most crucial factors that will influence the thermal comfort and ventilation effectiveness of the indoor environment is the airflow pattern of UFAD diffusers. ASHRAE Standard 62.1-2013 awards the highest ventilation effectiveness of 1.2 to UFAD system that achieves thermal stratification or the vertical throw is less than or equal to 0.25 ms-1 at the height of 1.4m above the floor [7]. Both displacement ventilation and underfloor air distribution (UFAD) systems can have the potential to achieve values greater than 1.0 [8]. Thermal stratification is defined by the stratification level that divides the room into upper and lower zones. Lower zone has a relatively well mixed air due to the turbulence created by high velocity jets of the floor supply air outlets. Upper zone has a warm and contaminated air rising by the heat plumes. The average air velocity at upper zone is relatively low [9]. The ventilation effectiveness of UFAD system will be downgraded to 1.0 if the system does not achieve the thermal stratification or the required air throw, and thus the ventilation effectiveness will be similar to the conventional overhead air distribution (OHAD) system. The potential to achieve higher ventilation effectiveness means a reduction on the required minimum outdoor air rates. There will be a significant reduction on the cooling energy consumption for cases with high ventilation effectiveness. Based on study conducted by Lin and Tsai [10], the supply airflow rate is found to have a strong connection with the supply air momentum flux and thus affects the gradient of the vertical temperature profiles of the indoor environment. The study describes the throw height as the level where the temperature at the supply air outlet area is equal to the temperature of that height in the other area. The results show that a higher airflow rate will produce a lower thermal stratification. Schiavon et al. [6] studied the thermal stratification in perimeter zones using UFAD linear bar grilles and VAV directional diffusers. The study describes the relationship between the non-dimensional parameter, Gamma “Γ”, and the dimensionless temperature ratio, Phi “Φ”, of the diffusers. Gamma, Γ, is the ratio of buoyancy forces to the vertical momentum forces, while phi, Φ, is the dimensionless temperature at a specific height in the room. The results show that thermal stratification by using VAV directional diffusers is generally higher than linear bar grilles in UFAD system. Besides, diffusers with a lower airflow rate will have a lower throw height and produce less mixing. The thermal stratification of UFAD system is generally lower compared to the displacement ventilation (DV) system. Thus, the ventilation performance of UFAD system can be comparable with the DV system when a large number of diffusers are used. Raftery et al. [11] studied the stratification performance of a floor DV diffuser, which delivers air with mostly horizontal air momentum. The thermal stratification based on the dimensionless temperature ratio, Φ, shows that the floor DV diffuser produces a comparable result for the DV system. Sajadi [12] studied the effect of blade angle on the performance of ceiling type swirl diffuser. It is found that an angle of about 32° is optimum and is almost independent of the airflow rate. While for a floor type diffuser, a grille with minimum thickness is required to support load and impact. The air distribution is predicted to be different from the ceiling type diffuser where the diffuser blades will diffuse the supply air directly to the conditioned space. Literature review reveals that research studies conducted on the
geometrical effect of the floor swirl diffuser in terms of the airflow pattern are rarely found, and therefore, the objective of the present study is to investigate the airflow pattern of a floor swirl diffuser through various aspects, including the airflow rate, number of diffuser blades, the angle of attack of diffuser blades and grille thickness. It is reasonable to suggest that a thicker grille, and a higher airflow rate will cause the conditioned air supplied to the space like an air jet, which is detrimental to the thermal stratification [13]. 2.
Experimental Work
The experimental work was carried out in the air terminal testing laboratory in Prudentaire Marketing Pte Ltd. The laboratory was located inside a building with all the walls, ceiling and floor exposed to the ambient room temperature. Experiment has been conducted to validate the CFD simulation in studying various geometrical designs of the floor swirl diffuser. The schematic diagram of experimental setup for the testing of floor swirl diffuser is shown in Figure 1. A single inlet centrifugal fan with airfoil wheels running by a 3-phase induction motor was used to supply air to the diffuser. The fan was installed with a variable-speed drive (VSD) to control the airflow rate by motor frequency. The fan capacity was substantially larger compared with the testing required airflow rate of a unit of floor swirl diffuser. Hence, the variable air volume (VAV) system with an actuator and a damper was used to further regulate the airflow rate to the testing requirement. A smoke inlet was available at the flexible duct to conduct smoke tests. Smoke was supplied into the ducting system after the VAV control. Boundary conditions of the laboratory were measured prior to the testing. The dimension of the laboratory is W5400mm x L6600mm x H3100mm. The surface temperatures of the walls, ceiling and floor were measured by taping the sensors of the Center 309 K-type thermocouple on the surfaces using reflective aluminum tape. The average room temperature was measured by swinging Alnor Model 440-A hot-wired anemometer at the reference point in the laboratory. The measurement tools has an operating range of temperature of -10 to 60 °C, RH of 5 to 95% and velocity of 0 to 30 ms-1 with an accuracy of ± 0.3 °C for temperature, ± 3% for RH, and ± 3% of reading or ± 0.015ms-1 for velocity. After the fan has been operated to deliver a constant airflow of 290 CMH with a fluctuation of less than 10%, measurements were taken every one hour. The steady-state condition was achieved at the 5th hour of the fan operation as the supply air temperature, average room air temperature and return air temperature did not vary more than 1°C [14]. The hourly temperature difference of the room surface temperatures also dropped below 0.1°C after 5 hours of operation. Hence, the case can be assumed to be in an isothermal condition. The assumption is made that the diffuser is much smaller than the size of the room. A large side air return in experiment and four return air outlets at the four corners of the room will have a negligible effect in simulation on the discharge airflow pattern of the floor swirl diffuser [15].
3.
Numerical Simulation
Background Theory The governing equations of conservation of mass, momentum, and energy are resolved in the computational fluid dynamics (CFD) analysis [16]. The indoor airflow motions are predicted through solving these three mathematical transport equations of time-averaged Navier–Stokes equations using ANSYS Fluent. The equations are as follows [16]: Conservation of mass: 𝜕𝜌 𝜕𝑡
+
𝜕(𝜌𝑢) 𝜕𝑥
+
𝜕(𝜌𝑣) 𝜕𝑦
+
𝜕(𝜌𝑤) 𝜕𝑧
=0
Conservation of momentum: x-momentum equations,
(1)
𝜌
𝐷𝑢 𝐷𝑡
=
𝜕𝜎𝑥𝑥
+
𝜕𝑥
𝜕𝜏𝑦𝑥
𝜕𝜏𝑧𝑥
+
𝜕𝑦
𝜕𝑧
𝑏𝑜𝑑𝑦 𝑓𝑜𝑟𝑐𝑒𝑠
+ ∑ 𝐹𝑥
(2)
y-momentum equations, 𝜌
𝐷𝑣 𝐷𝑡
=
𝜕𝜏𝑥𝑦 𝜕𝑥
+
𝜕𝜎𝑦𝑦
𝜕𝜏𝑧𝑦
+
𝜕𝑦
𝜕𝑧
𝑏𝑜𝑑𝑦 𝑓𝑜𝑟𝑐𝑒𝑠
+ ∑ 𝐹𝑦
(3)
z-momentum equations, 𝜌
𝐷𝑤 𝐷𝑡
=
𝜕𝜏𝑥𝑧 𝜕𝑥
+
𝜕𝜏𝑦𝑧
𝜕𝜎𝑧𝑧
+
𝜕𝑦
𝜕𝑧
𝑏𝑜𝑑𝑦 𝑓𝑜𝑟𝑐𝑒𝑠
+ ∑ 𝐹𝑧
(4)
where 𝜎𝑥𝑥 , 𝜎𝑦𝑦 and 𝜎𝑧𝑧 are the normal stresses due to pressure and normal viscous stress components acting perpendicular to the volume. Conservation of energy: The energy equation is derived from the first law of thermodynamics. In order to relate with the continuity equation and momentum equations, the time rate of change of energy is represented in x, y and z directions as shown below: 𝜌
𝐷𝐸 𝐷𝑡
=
𝜕 𝜕𝑥
𝜕𝑇
[𝜆
𝜕𝑥
]+
𝜕 𝜕𝑦
𝜕𝑇
[𝜆
𝜕𝑦
]+
𝜕
[𝜆
𝜕𝑧
𝜕𝑇 𝜕𝑧
]−
𝜕(𝑢𝑝) 𝜕𝑥
−
𝜕(𝑣𝑝) 𝜕𝑦
−
𝜕(𝑤𝑝) 𝜕𝑧
+𝛷
(5)
The dissipation function, Φ, represents the energy work done on the fluid, which is then converted into heat. 𝛷=
𝜕(𝑢𝜏𝑥𝑥 ) 𝜕𝑥
+
𝜕(𝑢𝜏𝑦𝑥 ) 𝜕𝑦
+
𝜕(𝑢𝜏𝑧𝑥 ) 𝜕𝑧
+
𝜕(𝑣𝜏𝑥𝑦 ) 𝜕𝑥
+
𝜕(𝑣𝜏𝑦𝑦 ) 𝜕𝑦
+
𝜕(𝑣𝜏𝑧𝑦 ) 𝜕𝑧
+
𝜕(𝑤𝜏𝑥𝑦 ) 𝜕𝑥
+
𝜕(𝑤𝜏𝑦𝑧 ) 𝜕𝑦
+
𝜕(𝑤𝜏𝑧𝑧 ) 𝜕𝑧
(6)
In the present study, turbulence k-ɛ model was employed to solve the air distribution pattern as the just mentioned model is most widely used in industrial flow and heat transfer simulations [17]. It solves two model transport equations separately to obtain turbulence kinetic energy per unit mass, k and dissipation rate, ɛ. The equations for standard k-ɛ turbulence kinetic energy are shown below: Turbulence kinetic energy, k: 𝜕 𝜕𝑡
(𝜌𝑘) +
𝜕 𝜕𝑥𝑖
(𝜌𝑘𝑢𝑖 ) =
𝜕 𝜕𝑥𝑗
[(𝜇 +
𝜇𝑡
)
𝜕𝑘
𝜎𝑘 𝜕𝑥𝑗
] + 𝐺𝑘 + 𝐺𝑏 − 𝜌𝜀 − 𝑌𝑀 + 𝑆𝑘
(7)
Dissipation rate, ɛ: 𝜕 𝜕𝑡
(𝜌𝜀) +
𝜕 𝜕𝑥𝑖
(𝜌𝜀𝑢𝑖 ) =
𝜕 𝜕𝑥𝑗
[(𝜇 +
𝜇𝑡
)
𝜕𝜀
𝜎𝜀 𝜕𝑥𝑗
𝑘2
𝜀
𝜀2
𝑘
𝑘
] + 𝐶1𝜀 (𝐺𝑘 + 𝐶3𝜀 𝐺𝑏) − 𝐶2𝜀 𝜌
+ 𝑆𝜀
(8)
Turbulent viscosity, 𝜇𝑡 = 𝜌𝐶𝜇 (9) 𝜀 where 𝐺𝑘 = generation of turbulence kinetic energy due to the mean velocity gradients, 𝐺𝑘 = generation of turbulence kinetic energy due to buoyancy, 𝑌𝑀 = contribution of the fluctuating dilatation in compressible turbulence to the overall dissipation rate, 𝐶1𝜀 , 𝐶2𝜀 = constants of 1.44, 1.92 and 𝐶𝜇 = constants of 0.09, 𝜎𝑘 , = Prandtl number for k, 1.0, 𝜎𝜀 = Prandtl number for ɛ, 1.3, and 𝑆𝑘 , 𝑆𝜀 = source terms. The RNG k-ɛ model is an enhancement from the standard k-ɛ model to improve the accuracy for rapid strained flow and swirl flow by additional terms in k and ɛ transport equations. The realizable k-ɛ model is another enhancement from the standard k-ɛ model. This model has a different formulation for turbulent viscosity, which causes this model to outperform other variations in several cases. Thus, the CFD results from these turbulence models were compared in the current study to obtain the most suitable turbulence models.
Mesh Independence Study Seven unstructured tetrahedral meshing (I – VII) with different numbers of mesh elements were performed on the model in relation to the relevant center in ANSYS meshing to study the mesh independency of the results. Typically, in mesh independence studies, the meshing increment step is 1.5 times from the coarser mesh. However, the increment step is not viable when the mesh element exceeds 3 x 106 cells. The computational time is too costly. Hence, a low incremental step of up to 1.24 times was used in the fine mesh setting to produce more than 5 x 106 cells. A higher increment step of up to 2.58 times was used in the coarse mesh to reduce the number of analyzing models in the current study. The computer used to run the simulation was incorporated with Intel core i7-4510U CPU @ 2.00GHz processor, 8.00GB RAM and 64-bit Window 10 Pro operating system. Due to the limited computational power available, the mesh was studied from the coarse mesh of about 200,000 cells until the room temperature difference of two consecutive mesh elements dropped below 0.01°C. Figure 2 shows that the room mean temperature is consistent for the 3 meshes with the mesh elements of 3 million and above. A further analysis has been conducted to compare the center plane velocity profiles of the three meshes [18]. Figures 3a and 3b show the comparison between meshes V & VI, and meshes VI & VII respectively. The plane velocity profile is almost identical for the case in meshes VI &VII. Thus, the model can be considered to be mesh independence at mesh VI.
Data Verification The CFD investigation of diffuser airflow pattern is problematic due to the complex geometry of the diffuser, and the size of the diffuser is relatively small compared with the room dimensions. A correct turbulent model is crucial to predict accurately turbulent airflow pattern and the heat transfer between the inflow and the room. Hence, data comparison and verification are important to ensure that the numerical modeling represents the best physical model. The standard k-ɛ turbulence model has been widely used for diffusers’ airflow prediction in OHAD systems such as round diffusers, square diffusers and swirl diffusers [12, 19, 20]. In this study, the standard k-ɛ, RNG k-ɛ, and realization k-ɛ are compared with the experimental measured data. The plane airflow pattern is then compared with the smoke test data to find the most suitable turbulence model for the present investigation. For comparison and verification purposes, the centerline air velocity was extracted from the simulation to compare with the experimental measurement. Figure 4 shows the vertical centerline air velocity profiles of the measured and the predicted data for few turbulence models. No perfect agreement between the measured and predicted data is expected. However, both the standard and RNG k-ɛ models indicate a similar centerline air velocity pattern to the measured data. In fact, the RNG k-ɛ model has a closer prediction to the measured data at the location near to the floor swirl diffuser outlet. There is a drop at around y/H = 0.20, and a spike at around y/H = 0.40 in all turbulence models. This fluctuation is observed mainly due to the swirl mechanism that causes difficulty in measuring turbulent flow using the existing tool (i.e. Hot-wire anemometer). It is noted that the computed and measured results are consistent with the results found by Fathollahzadeh [21]. It is also noted that the spike of air velocity in the computed results were not captured in the measured data (Figure 4). This could be due to the limitation of the measurement tool just mentioned, where in the current work, a hot-wired anemometer was used for the measurement. Hence, the predicted models are acceptable in the present research. In addition, both measured and simulated results are consistently showing a velocity drop at about y/H = 0.20 and an acceleration at about y/H = 0.40 (Figure 4). This phenomenon was captured because the swirl diffusers create a circular air flow near the air inlet, and as the airflow moving upwards, the airflow changed from the horizontal air movement to the vertical air movement. Figure 4 has shown that the X-Y
plane air velocity, where the air velocity at the Z-direction (horizontal air velocity), was not captured. Thus, the airflow development can be viewed as the changes of the horizontal air movement to the vertical air movement. Figure 5 shows the plane air velocity profiles for standard, RNG and realizable k-ɛ turbulence models as compared to the smoke test photograph. The standard k-ɛ turbulence model, which has a higher centerline velocity compared with the air velocity at the edges of the flow plume shows the closest match to the actual airflow pattern. The RNG k-ɛ turbulence model has a higher air velocity at the edges of the flow plume compared with the centerline air velocity. Hence, the standard k-ɛ turbulence model is used for the present floor swirl diffuser model investigation.
4.
Results and Discussion
The base model used to study the geometrical design changes is a 250 mm nominal diameter floor swirl diffuser with 18 diffuser blades at the 45-degree angle of attack, and 3 mm of grille thickness tested at 195 cubic meter per hour (CMH). The following results to be elaborated in the subsequent paragraphs show the effects due to the geometrical design changes of the floor swirl diffuser. The Effect of Airflow Rates This study was conducted to compare airflow rates of 100, 195 and 290 CMH, which are the lowest to the highest recommended airflow rates used for this typical floor swirl diffuser [22]. As indicated by Lin [9], the airflow rate has a strong influence on vertical temperature profiles where a higher airflow rate lowers the thermal stratification. For the purpose of understanding the thermal stratification of the UFAD system in terms of vertical velocity profiles, a comparison on the centerline air velocity is reported and shown in Figure 6. The characteristics of a pure jet are defined by thermal length scale, which is significant for thermal stratification. In UFAD systems where the heat load is consistent, the thermal stratification depends highly on the velocity of the supply air [23]. The thermal length scale is defined as below [24, 25]: 3/4
𝑙𝑚 =
𝑀0
1/2
𝐵0 𝑀0 = 𝑄0 . 𝑢0 𝑔. 𝑄0 . ∆𝑇0 𝐵0 = 𝑇𝑟 where l_m is the thermal length scale, M is the momentum flux, B is the buoyancy flux, Q is the volume flux (m3s-1), u is the supply air velocity (ms-1), g is the gravitational acceleration, ∆T is the temperature difference between supply air of the diffuser and the room temperature and Tr is the reference air temperature. For the case of 100 CMH, a terminal velocity of 0.25 ms-1 has reached at around 1.3 m above the floor swirl diffuser. An UFAD system with a vertical throw of less than or equal to 0.25 ms-1 at the height of 1.4 m above the floor can have a zone air distribution effectiveness of 1.2 according to ASHRAE Standard 62.1 [7]. It is the only case that could achieve the requirement to receive the highest ventilation effectiveness of 1.2. Thus, the outdoor airflow rates can be reduced by a factor of 1.2. The terminal velocity for 195 CMH and 290 CMH is around 2.4 m and 2.8 m respectively. Hence, there is insufficient space to create lower and higher stratified zones. It will be a mixing air distribution system as the thermal stratification cannot be formed. Due to the importance of supply air velocity towards the thermal stratification as described earlier, it should be noted that the supply air velocities for the three cases are 1.73 ms-1, 3.38 ms-1 and 5.03 ms1 respectively. It is also observed from Figure 6 that the swirl diffusers again create a circular air flow near the air inlet as described earlier in Figure 4.
Although the airflow rate of 100 CMH achieves a higher ventilation effectiveness, Figure 7 shows that the airflow rate is insufficient in reducing the temperature of the environment as effective as the airflow rates of 195 CMH and 290 CMH, where the surrounding space has a lower temperature around 298 K (25°C). Besides, the two cases with higher airflow rates create a greater air mixing in the area surrounding the diffuser’s discharge outlet. The air movement is about 0.1 ms-1, which is slightly below the recommended air movement by Malaysian Standard (0.15 – 0.5 ms-1 for offices) [26]. A larger number of diffusers per room is needed when the airflow rate is small. It is also pertinent to mention that the supply air temperature of an UFAD system should be higher (usually 16-18°C) than the OHAD system to make sure that the occupants will not experience cold feet, and therefore the air jet characteristic length is a key to calculate the air distribution performance [8].
The Effect of the Number of Diffuser Blades This study examines the effect of the number of diffuser blades on the airflow pattern of the floor swirl diffuser. Diffusers with 12, 18 and 24 blades, as shown in Figure 8, were simulated separately at the same conditions at 195 CMH. Note that the 18-blade diffuser is commonly used by various air terminal manufacturers. The percentage of the free area for 12, 18 and 24 blades is 43.27%, 18.86% and 5.51% respectively. The blade used in the study has a curvature profile adopted from the upper surface of airfoil NACA 0012 to enhance the aerodynamic airflow. The maximum air velocities that discharged from the under-floor swirl diffuser for 12, 18 and 24 blades are 3.24 ms-1, 3.89 ms-1 and 4.19 ms-1 respectively. x In the case of 12-blades swirl mechanism as shown in Figure 9, almost half of the area is a free area (i.e. 43.27%). It makes the angled swirl blades ineffective in creating a swirl effect. The primary air moves vertically upwards like an air jet without expansion. Secondary air movement is induced near to the primary air up to the ceiling height. The 18-blades swirl mechanism, in which its free area is 18.86%, shows a well-developed plume air development due to the entrainment of room air. It reduces the velocity of the primary air significantly and shortens the throw. Hence, the secondary air movement is induced at a larger low level area of the room. The 18-blades swirl mechanism creates a greater swirl air effect compared to the 12-blades swirl mechanism. The 24-blades swirl mechanism has the most blades and the least free area at 5.51%. However, the primary air-throw is higher compared to the case of 18-blades. The swirl effect does not grow greater when compared to the case of 18-blades. This is due to the limited free area for the air to discharge and causes the air to be forced out from the diffuser at high velocity. In this case, the angled blades do not perform effectively in directing the air to flow in the swirl effect. The Effect of Diffuser Blade’s Angle of Attack This study elaborates the angle of attack that affects the room air movement and the room temperature. The angles studied are 30°, 45° and 60°. The percentage of the free area for 30°, 45° and 60° is 14.25%, 18.86% and 33.84% respectively. Figure 10 shows the plane air velocity profiles at the angles of 30°, 45° and 60° respectively. The airflow patterns in the cases of 30° and 45° have a similar establishment. In both cases, the room at the lower level has induced an air movement of about 0.1 ms-1. In fact, the primary airflow in the case of 45°-blade developed a greater air entrainment from the room air at the height of 1 – 2 meters above the diffuser. However, the 30°-blade creates a more evenly distributed temperature across the room compared with the other two cases. The primary airflow pattern in the 60°-blade is similar to the case of the 12-blades diffuser where the air is moving vertically without the expansion like an air jet. This causes the mixing of air at the higher level. The results are consistent with the findings by Sajadi [12] where the ceiling based swirl diffuser performs optimally at the swirl angle of 32°.
The Effect of Diffuser and Grille Thickness The floor swirl diffuser is installed on the floor. The design of grille top of the diffuser blade is important to support the impact and the loads of human or object. A 20mm thick grille is commonly used in the market. Figure 11 shows the floor swirl diffuser examined. This study discusses the effects of the presence of diffuser blades and grille thickness that affect the airflow pattern in the room. Although the design of the grille will affect the minimum grille thickness required, the impact and load study have not been conducted in the present work. The grille thickness test emphasizes on the airflow performance. The five cases studied are no diffuser blades, no grille, 3 mm grille, 10 mm grille and 20 mm grille. Note that a diffuser without a grille fitted to the UFAD system will not function properly due to the fact that the diffuser blades without a thick grille cannot withstand impact and load of a human or an object. Nevertheless, it is included in the current study for comparison purpose in order to have a better understanding of the airflow patterns.
Figure 12 shows the center plane air temperature profiles for all cases studied. The diffuser blade plays a key role in making the supply air rotating and mixing with the room air. When there is no diffuser blade, the air is delivered to the space vertically upward like an air jet with a minimum entrainment. The result is similar to the finding on ceiling type swirl diffuser by Sajadi [12]. Besides, the swirl angle of the diffuser plays a vital role in defining the characteristics of the swirl jet, especially when the grille is absent. The velocity profiles at the centerline show that a throw height of 0.25 ms-1 is reached at about 1.8 m above the floor. The primary air has entrained the room air more effectively compared to other cases. This can be seen by the expansion of the plume air jet, and the hollow cone pattern has formed. The case is consistent with the swirl angle study performed by Tavakoli et al. [27] on the ceiling based swirl diffuser, and the complete theory and flow solver details can be found in references [28, 29]. As the grille thickness increases, it is observed that the primary air is delivered into the space vertically upwards with a higher air throw and velocities. The primary air jet has induced a secondary air flow near to the jet. However, the room temperature becomes uneven as the cold air is delivered straight to the return air grille on the ceiling. The results in the case of no swirl diffuser blade have shown similarity to cases as the grille thickness increases such as 10 mm and 20 mm grille thickness. The thickness of the grille has proven to be detrimental to the function of diffuser blades. The Centerline Velocity, Vc The case on 195 CMH is the base model used to study the geometrical design changes. The simulation result suggests a good agreement with the experimental measurement. Figure 13 shows the centerline velocity of all models through experimental measurement and simulation. Among all the comparisons, the centerline velocity graph shows that the grille thickness has a major impact on the airflow from the floor swirl diffuser. The cases with 10mm and 20mm thick grilles have resulted in high air velocities near the air outlet at 1.91 ms-1 and 2.06 ms-1 respectively. The case without diffuser achieves a similar result of high centerline air velocity near the air outlet at 1.88 ms-1. This has proven that the air is delivered into the space without a proper swirling effect. For the cases with no grille and 3mm grille thicknesses, the centerline air velocities are greatly reduced to maximum values of 0.57 ms-1 and 0.77 ms-1 respectively. Therefore, having a thick grille is equal to not having diffuser blades, and the air will be delivered to the space without creating the swirling effect.
5.
Conclusions
In the present study, a comprehensive investigation has been successfully conducted on the airflow pattern of a floor swirl diffuser through various aspects, including the airflow rate, number of diffuser blades, the angle of attack of diffuser blades and grille thickness. The results can be used as an important guide to design an optimised UFAD system. Based on the results obtained, there are several major conclusions made as follows:
1)
2)
3)
4)
6.
In the airflow rate study, a 250 mm nominal diameter floor swirl diffuser with an airflow rate of 100 CMH and below can be awarded with a ventilation effectiveness of 1.2 according to ASHRAE Standard 62.1:2013. The minimum outdoor airflow rate can be reduced by dividing it with a factor of 1.2 if the design is in accordance to this flowrate or below. The momentum flux of the diffuser is directly proportional to the supply air velocity. Thus, the equivalent supply air velocity of 1.73 ms-1 is acceptable. An airflow rate of 195 CMH and above supplied using this diffuser will not obtain this better ventilation effectiveness at 1.2 because a higher air throw will cause the low and high stratification layers to be mixed by the turbulences of the flow. The diffuser performance will then be identical to an OHAD mixing system of 1.0. The future research on thermal stratification is recommended to include heat load profiles of the room in order to have a complete understanding on thermal stratification. It is suggested that the optimum number of diffuser blades to create a swirl effect is 18 blades, which has a free area of 18.86%. The airflow of this case shows a well-developed plume air development (Figure 9). The air plume developed is the largest among the three cases. A large free area in the case of 12 blades or a small free area in the case of 24 blades could lead to a direct air jet flow into the space, and thus ineffective in the air plume development, which is similar to the case without diffuser blades. The swirl blades are important in rotating the discharged air to improve the mixing effect. At the low angles of attack of 30° and 45°, the air throw is lower compared to a high angle of attack. The primary air is established and expanded by entraining room air. The room achieves an evenly distributed temperature profile at the low angle of attack due to the better air mixing effect. The blade angle can have a different effect on thermal stratification in cooling or heating mode. This can be an extension for the future research work. A thicker grille leads to air delivered as a straightened air jet into the space. The swirl effect from the diffuser blades is eliminated. Thus, a minimum grille thickness should be developed without detrimental effect to the strength to withstand the impact and load of people and furniture.
Acknowledgement and Declaration
The authors would like to thank University of Malaya for providing UMRG Grant RG030/15AET to the authors for research work to be conducted at University of Malaya. Thanks are also extended to University of Malaya PPP Grants PG111-2012B and PG042-2016A for providing the partial financial assistance to the co-author, Mr. K.S. Poh, for conducting the research work at HVAC&R Lab at the Department of Mechanical Engineering, University of Malaya. In addition, special thanks are extended to Prudentaire Marketing Sdn Bhd for their contribution in the laboratory testing conducted in this project.
References 1 2 3 4
5 6 7 8 9 10 11 12 13 14 15 16 17 18
19
20
ASHRAE. UFAD Guide. Atlanta (GA): ASHRAE; 2013; 323. Bauman F, Webster T. Outlook for Underfloor Air Distribution. ASHRAE Journal 2001; 43(6): 18-27. Liu QA, Linden PF. The fluid dynamics of an underfloor air distribution system. Journal of Fluid Mechanics 2006; 554(1): 323-341. Webster T, Linden P, Buhl F, et al. Energy performance of underfloor air distribution systems. Final Project Report submitted to California Energy Commission (CEC) Public Interest Energy Research (PIER) Program. California (USA): Center for the Built Environment, University of California, Berkeley, CA; 2007. Pasut W. Using Ductwork to Improve Supply Plenum Temperature Distribution in Underfloor Air Distribution (UFAD) System. Padua (Italy): University of Padua; 2011. Schiavon S, Webster T, Dickerhoff D, et al. Stratification prediction model for perimeter zone UFAD diffusers based on laboratory testing with solar simulator. Energy and Buildings 2014; 82: 786-794. ASHRAE Standing Standard Project Committee 62.1. ANSI/ASHRAE STANDARD 62.1-2013: Ventilation for Acceptable Indoor Air Quality. Atlanta (GA): ASHRAE, 2013; 52. ASHRAE. HVAC Applications Handbook, Chapter 57: Room air distribution. Atlanta (GA): ASHRAE; 2015. Webster T, Bauman F, Reese J. Underfloor air distribution: Thermal stratification. ASHRAE Journal 2002; 44(5): 28-36. Lin YJP, Tsai TY. An experimental study on a full-scale indoor thermal environment using an underfloor air distribution system. Energy and Buildings 2014; 80: 321-330. Raftery P, Bauman F, Schiavon S, et al. Laboratory testing of a displacement ventilation diffuser for underfloor air distribution systems. Energy and Buildings 2015; 108: 82-91. Sajadi B, Saidi MH, Mohebbian A. Numerical investigation of the swirling air diffuser: Parametric study and optimization. Energy and Buildings 2011; 43(6): 1329-1333. Zhang K, Zhang X, Li S. Simplified model for desired airflow rate in underfloor air distribution (UFAD) systems. Applied Thermal Engineering 2016; 93: 244-50. ASHRAE Standing Standard Project Committee 113. ANSI/ASHRAE Standard 113-2013: Method of Testing for Room Air Diffusion. Atlanta (GA): ASHRAE, 2013. Mohammed RH. A simplified method for modeling of round and square ceiling diffusers. Energy and Buildings 2013; 64: 473-482. Tu J, Yeoh GH, Liu C. Computational Fluid Dynamics - A Practical Approach. 2nd edition, Heinemann, Butterworth, United Kingdom, 2013: 440. ANSYS, ANSYS Fluent Theory Guide. 2013, Canonsburg, Pennsylvania: ANSYS Inc. Nada SA, El-Batsh HM, Elattar HF, et al. CFD investigation of airflow pattern, temperature distribution and thermal comfort of UFAD system for theater buildings applications. Journal of Building Engineering 2016; 6: 274-300. Aziz MA, Gad Ibrahim AM, Mohammed El Shahat FA, et al. Experimental and numerical study of influence of air ceiling diffusers on room air flow characteristics. Energy and Buildings 2012; 55: 738746. Hu SC. Airflow characteristics in the outlet region of a vortex room air diffuser. Building and Environment 2003; 38(4): 553-561.
21 Fathollahzadeh MH, Heidarinejad G, Pasdarshahri H. Prediction of thermal comfort, IAQ, and energy
consumption in a dense occupancy environment with the under floor air distribution system. Building and Environment 2015; 90: 96-104. 22 Trox Malaysia. Floor Diffusers. https://www.troxapo.com/floor-diffusers/type-fba-%C2%B7-fbk82577427340e09dc. Retrieved May 2015. 23 Fan Y, Li X, Yan Y, Tu J. Overall performance evaluation of underfloor air distribution system with
different heights of return vents. Energy and Buildings 2017; 147: 176-87. 24 Arghand T, Karimipanah T, Awbi HB, et al. An experimental investigation of the flow and comfort parameters for under-floor, confluent jets and mixing ventilation systems in an open-plan office. Building and Environment 2015; 92: 48-60. 25 Zhang, K., Zhang X, Li S, et al. Experimental study on the characteristics of supply air for UFAD system with perforated tiles. Energy and Buildings 2014; 80: 1-6.
26 Department of Standards Malaysia. MS 1525:2014: Code of Practice on Energy Efficiency and Use of Renewable Energy for Non-Residential Buildings. Selangor (Malaysia): Department of Standards Malaysia, 2014; 46 27 Tavakoli E, Hosseini R. Large eddy simulation of turbulent flow and mass transfer in far-field of swirl diffusers. Energy and Buildings 2013; 59: 194-202. 28 Yau YH, Badarudin A, Rubini P. A Numerical Study of Secondary Flow and Large Eddies in a Driven Cavity. Journal of Mechanical Science and Technology (JMST) 2011; 26 (1): 1-11. 29 Yau YH, Badarudin A, Rubini P. Numerical and Scalability Analysis of an Unstructured Cartesian Flow Solver. Proceedings of the Institution of Mechanical Engineers, Part C, Journal of Mechanical Engineering Science 2011; 225: 2138-2148.
Figure 1: Schematic diagram of the experimental setup in the laboratory
30.72
Temperature (°C)
30.71 II (603810, 30.7)
30.70
V (3483738, 30.69)
VII (5612559, 30.69)
30.69 VI (4311451, 30.69)
30.68
IV (2103269, 30.68)
30.67
III (1174073, 30.66)
30.66
I (233695, 30.66)
30.65 -
10,00,000
20,00,000 30,00,000 40,00,000 Number of mesh elements
50,00,000
60,00,000
Figure 2: Room mean temperature at different mesh elements.
Figure 3a: Plane air velocity for Mesh V (top left), plane air velocity for Mesh VI (top right) and air velocity differences between Meshes V and VI (bottom).
Figure 3b: Plane air velocity for Mesh VI (top left), plane air velocity for Mesh VII (top right) and air velocity differences between Meshes VI and VII (bottom)
y/H
Experiment
1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.00
0.20
Standard k-ɛ
0.40
Realizable k-ɛ
0.60 0.80 1.00 1.20 1.40 Centerline velocity (ms-1)
RNG k-ɛ
1.60
1.80
2.00
Figure 4: Vertical centerline air velocity profile of an actual model and various predicted turbulence models
(a)
(b)
(c) (d) Figure 5: Plane air velocity profiles for (a) smoke test; (b) standard k-ɛ; (c) RNG k-ɛ; (d) realizable k-ɛ
y/H
100CMH
195CMH
290CMH
1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0
0.2
0.4
0.6 0.8 Centerline velocity (ms-1)
1
1.2
1.4
Figure 6: The centerline vertical air velocity profiles for airflow rates of 100, 195 and 290 CMH
(a)
(b)
(c)
Figure 7: The center plane air temperature profiles for airflow rates of (a) 100 CMH; (b) 195 CMH; (c) 290 CMH
(a) (b) (c) Figure 8: Diffuser’s swirl mechanism with (a) 12 blades, (b) 18 blades and (c) 24 blades
(a) (b) (c) Figure 9: The center plane air velocity profiles for (a) 12-blades; (b) 18-blades; (c) 24-blades swirl mechanism
(a) (b) (c) Figure 10: Plane air velocity profiles for diffuser blades at the angles of (a) 30°; (b) 45°; (c) 60°
Figure 11: Floor swirl diffuser
(a)
(b)
(c)
(d) (e) Figure 12: The center plane air temperature profiles for the cases (a) without diffuser blades; (b) without grille; (c) with 3 mm grille thickness; (d) with 10 mm grille thickness; (e) with 20 mm grille thickness
Figure 13: The centerline velocity (m/s) for all the geometrical variation of floor swirl diffuser from the floor level of 0 meter to 3 meters above floor level
Table 1: Boundary conditions of the CFD simulation. Boundary Details Supply air inlet Airflow rate = 100, 195, 290 CMH Air temperature = 18.0°C Relative humidity = 60% Exhaust air outlet 4 numbers of 600 x 600 mm ceiling exhaust air outlet at the four corners of the room at 0 Pa gauge pressure Walls Surface temperature = 28.5°C Ceiling Surface temperature = 27.0°C Floor Surface temperature = 27.0°C Grille Adiabatic Air plenum Adiabatic
S0378-7788(17)30590-X https://doi.org/10.1016/j.enbuild.2017.10.037 ENB 8057
To appear in:
ENB
Received date: Revised date: Accepted date:
18-2-2017 24-9-2017 10-10-2017
Please cite this article as: Y.H.Yau, K.S.Poh, A.Badarudin, A numerical airflow pattern study of a floor swirl diffuser for UFAD system, Energy and Buildings https://doi.org/10.1016/j.enbuild.2017.10.037 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A numerical airflow pattern study of a floor swirl diffuser for UFAD system Y.H. Yau, K.S. Poh and A. Badarudin Department of Mechanical Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia email: [email protected]
Research Highlights
This present study is to examine the effect of geometrical design of a floor swirl diffuser. Laboratory experiment was carried out to validate both CFD and turbulence models. The results could be used as an important guide to design an optimised UFAD system. The optimum airflow rate per diffuser is required in the design of UFAD system to achieve thermal stratification.
Abstract Floor swirl diffuser is one of the commonly used diffusers in underfloor air distribution (UFAD) system. Geometrical design of the floor swirl diffuser will affect the airflow pattern of the indoor environment. Thus, the purpose of this study is to examine the effect of geometrical design of a floor swirl diffuser on the aspects of airflow rate, number of diffuser blades, angle of attack of diffuser blades, and the grille thickness. Laboratory experiment was carried out to validate CFD simulation results. The results of this study have shown that diffuser blades have an important role in the airflow pattern. The grille thickness, which is required to withstand the impact and the load of human and furniture, needs to be minimized as the thickness is detrimental to the function of diffuser blades. Besides, airflow rate per diffuser is studied in UFAD system in order to achieve thermal stratification. The air throw is lower compared to a high angle of attack at the low angles of attack of 30° and 45°. The optimum number of diffuser blades to create the swirl effect is 18 blades with a free area of 18.86%. Keywords Underfloor air distribution (UFAD); Floor swirl diffuser; Computational fluid dynamics (CFD); Airflow pattern; Thermal stratification Nomenclatures CFD HVAC OHAD UFAD VSD VAV
Computational fluid dynamics Heating, ventilating and air conditioning Overhead air distribution Underfloor air distribution Variable speed drive Variable air volume
CMH mm m/s m Vc
cubic meter per hour millimeter meter per second meter centerline velocity
1.
Introduction
UFAD diffusers are used to deliver conditioned air from the floor plenum to the occupants' space for human comfort in UFAD system. Generally, UFAD diffusers are categorized into six types: round omnidirectional diffusers, selective directional diffusers, displacement flow diffusers, linear diffusers, ducted diffusers and terminal units [1]. Although there are various UFAD diffusers available in the market since the emerging of underfloor air distribution technology in 1990s, there is limited research work carried out on the geometry of UFAD diffusers. The present study examines the geometrical effect of the floor swirl diffuser in terms of the airflow pattern. A well-designed UFAD system can have a potential to reduce energy consumption, reduce floor to floor height, improve thermal comfort, improve ventilation efficiency and provide flexibility in reconfiguring the interior layout of the building [2]. The performance of UFAD system has been examined from the aspects of fluid dynamics, energy performance, temperature distribution and thermal stratification [3-6]. However, one of the most crucial factors that will influence the thermal comfort and ventilation effectiveness of the indoor environment is the airflow pattern of UFAD diffusers. ASHRAE Standard 62.1-2013 awards the highest ventilation effectiveness of 1.2 to UFAD system that achieves thermal stratification or the vertical throw is less than or equal to 0.25 ms-1 at the height of 1.4m above the floor [7]. Both displacement ventilation and underfloor air distribution (UFAD) systems can have the potential to achieve values greater than 1.0 [8]. Thermal stratification is defined by the stratification level that divides the room into upper and lower zones. Lower zone has a relatively well mixed air due to the turbulence created by high velocity jets of the floor supply air outlets. Upper zone has a warm and contaminated air rising by the heat plumes. The average air velocity at upper zone is relatively low [9]. The ventilation effectiveness of UFAD system will be downgraded to 1.0 if the system does not achieve the thermal stratification or the required air throw, and thus the ventilation effectiveness will be similar to the conventional overhead air distribution (OHAD) system. The potential to achieve higher ventilation effectiveness means a reduction on the required minimum outdoor air rates. There will be a significant reduction on the cooling energy consumption for cases with high ventilation effectiveness. Based on study conducted by Lin and Tsai [10], the supply airflow rate is found to have a strong connection with the supply air momentum flux and thus affects the gradient of the vertical temperature profiles of the indoor environment. The study describes the throw height as the level where the temperature at the supply air outlet area is equal to the temperature of that height in the other area. The results show that a higher airflow rate will produce a lower thermal stratification. Schiavon et al. [6] studied the thermal stratification in perimeter zones using UFAD linear bar grilles and VAV directional diffusers. The study describes the relationship between the non-dimensional parameter, Gamma “Γ”, and the dimensionless temperature ratio, Phi “Φ”, of the diffusers. Gamma, Γ, is the ratio of buoyancy forces to the vertical momentum forces, while phi, Φ, is the dimensionless temperature at a specific height in the room. The results show that thermal stratification by using VAV directional diffusers is generally higher than linear bar grilles in UFAD system. Besides, diffusers with a lower airflow rate will have a lower throw height and produce less mixing. The thermal stratification of UFAD system is generally lower compared to the displacement ventilation (DV) system. Thus, the ventilation performance of UFAD system can be comparable with the DV system when a large number of diffusers are used. Raftery et al. [11] studied the stratification performance of a floor DV diffuser, which delivers air with mostly horizontal air momentum. The thermal stratification based on the dimensionless temperature ratio, Φ, shows that the floor DV diffuser produces a comparable result for the DV system. Sajadi [12] studied the effect of blade angle on the performance of ceiling type swirl diffuser. It is found that an angle of about 32° is optimum and is almost independent of the airflow rate. While for a floor type diffuser, a grille with minimum thickness is required to support load and impact. The air distribution is predicted to be different from the ceiling type diffuser where the diffuser blades will diffuse the supply air directly to the conditioned space. Literature review reveals that research studies conducted on the
geometrical effect of the floor swirl diffuser in terms of the airflow pattern are rarely found, and therefore, the objective of the present study is to investigate the airflow pattern of a floor swirl diffuser through various aspects, including the airflow rate, number of diffuser blades, the angle of attack of diffuser blades and grille thickness. It is reasonable to suggest that a thicker grille, and a higher airflow rate will cause the conditioned air supplied to the space like an air jet, which is detrimental to the thermal stratification [13]. 2.
Experimental Work
The experimental work was carried out in the air terminal testing laboratory in Prudentaire Marketing Pte Ltd. The laboratory was located inside a building with all the walls, ceiling and floor exposed to the ambient room temperature. Experiment has been conducted to validate the CFD simulation in studying various geometrical designs of the floor swirl diffuser. The schematic diagram of experimental setup for the testing of floor swirl diffuser is shown in Figure 1. A single inlet centrifugal fan with airfoil wheels running by a 3-phase induction motor was used to supply air to the diffuser. The fan was installed with a variable-speed drive (VSD) to control the airflow rate by motor frequency. The fan capacity was substantially larger compared with the testing required airflow rate of a unit of floor swirl diffuser. Hence, the variable air volume (VAV) system with an actuator and a damper was used to further regulate the airflow rate to the testing requirement. A smoke inlet was available at the flexible duct to conduct smoke tests. Smoke was supplied into the ducting system after the VAV control. Boundary conditions of the laboratory were measured prior to the testing. The dimension of the laboratory is W5400mm x L6600mm x H3100mm. The surface temperatures of the walls, ceiling and floor were measured by taping the sensors of the Center 309 K-type thermocouple on the surfaces using reflective aluminum tape. The average room temperature was measured by swinging Alnor Model 440-A hot-wired anemometer at the reference point in the laboratory. The measurement tools has an operating range of temperature of -10 to 60 °C, RH of 5 to 95% and velocity of 0 to 30 ms-1 with an accuracy of ± 0.3 °C for temperature, ± 3% for RH, and ± 3% of reading or ± 0.015ms-1 for velocity. After the fan has been operated to deliver a constant airflow of 290 CMH with a fluctuation of less than 10%, measurements were taken every one hour. The steady-state condition was achieved at the 5th hour of the fan operation as the supply air temperature, average room air temperature and return air temperature did not vary more than 1°C [14]. The hourly temperature difference of the room surface temperatures also dropped below 0.1°C after 5 hours of operation. Hence, the case can be assumed to be in an isothermal condition. The assumption is made that the diffuser is much smaller than the size of the room. A large side air return in experiment and four return air outlets at the four corners of the room will have a negligible effect in simulation on the discharge airflow pattern of the floor swirl diffuser [15].
3.
Numerical Simulation
Background Theory The governing equations of conservation of mass, momentum, and energy are resolved in the computational fluid dynamics (CFD) analysis [16]. The indoor airflow motions are predicted through solving these three mathematical transport equations of time-averaged Navier–Stokes equations using ANSYS Fluent. The equations are as follows [16]: Conservation of mass: 𝜕𝜌 𝜕𝑡
+
𝜕(𝜌𝑢) 𝜕𝑥
+
𝜕(𝜌𝑣) 𝜕𝑦
+
𝜕(𝜌𝑤) 𝜕𝑧
=0
Conservation of momentum: x-momentum equations,
(1)
𝜌
𝐷𝑢 𝐷𝑡
=
𝜕𝜎𝑥𝑥
+
𝜕𝑥
𝜕𝜏𝑦𝑥
𝜕𝜏𝑧𝑥
+
𝜕𝑦
𝜕𝑧
𝑏𝑜𝑑𝑦 𝑓𝑜𝑟𝑐𝑒𝑠
+ ∑ 𝐹𝑥
(2)
y-momentum equations, 𝜌
𝐷𝑣 𝐷𝑡
=
𝜕𝜏𝑥𝑦 𝜕𝑥
+
𝜕𝜎𝑦𝑦
𝜕𝜏𝑧𝑦
+
𝜕𝑦
𝜕𝑧
𝑏𝑜𝑑𝑦 𝑓𝑜𝑟𝑐𝑒𝑠
+ ∑ 𝐹𝑦
(3)
z-momentum equations, 𝜌
𝐷𝑤 𝐷𝑡
=
𝜕𝜏𝑥𝑧 𝜕𝑥
+
𝜕𝜏𝑦𝑧
𝜕𝜎𝑧𝑧
+
𝜕𝑦
𝜕𝑧
𝑏𝑜𝑑𝑦 𝑓𝑜𝑟𝑐𝑒𝑠
+ ∑ 𝐹𝑧
(4)
where 𝜎𝑥𝑥 , 𝜎𝑦𝑦 and 𝜎𝑧𝑧 are the normal stresses due to pressure and normal viscous stress components acting perpendicular to the volume. Conservation of energy: The energy equation is derived from the first law of thermodynamics. In order to relate with the continuity equation and momentum equations, the time rate of change of energy is represented in x, y and z directions as shown below: 𝜌
𝐷𝐸 𝐷𝑡
=
𝜕 𝜕𝑥
𝜕𝑇
[𝜆
𝜕𝑥
]+
𝜕 𝜕𝑦
𝜕𝑇
[𝜆
𝜕𝑦
]+
𝜕
[𝜆
𝜕𝑧
𝜕𝑇 𝜕𝑧
]−
𝜕(𝑢𝑝) 𝜕𝑥
−
𝜕(𝑣𝑝) 𝜕𝑦
−
𝜕(𝑤𝑝) 𝜕𝑧
+𝛷
(5)
The dissipation function, Φ, represents the energy work done on the fluid, which is then converted into heat. 𝛷=
𝜕(𝑢𝜏𝑥𝑥 ) 𝜕𝑥
+
𝜕(𝑢𝜏𝑦𝑥 ) 𝜕𝑦
+
𝜕(𝑢𝜏𝑧𝑥 ) 𝜕𝑧
+
𝜕(𝑣𝜏𝑥𝑦 ) 𝜕𝑥
+
𝜕(𝑣𝜏𝑦𝑦 ) 𝜕𝑦
+
𝜕(𝑣𝜏𝑧𝑦 ) 𝜕𝑧
+
𝜕(𝑤𝜏𝑥𝑦 ) 𝜕𝑥
+
𝜕(𝑤𝜏𝑦𝑧 ) 𝜕𝑦
+
𝜕(𝑤𝜏𝑧𝑧 ) 𝜕𝑧
(6)
In the present study, turbulence k-ɛ model was employed to solve the air distribution pattern as the just mentioned model is most widely used in industrial flow and heat transfer simulations [17]. It solves two model transport equations separately to obtain turbulence kinetic energy per unit mass, k and dissipation rate, ɛ. The equations for standard k-ɛ turbulence kinetic energy are shown below: Turbulence kinetic energy, k: 𝜕 𝜕𝑡
(𝜌𝑘) +
𝜕 𝜕𝑥𝑖
(𝜌𝑘𝑢𝑖 ) =
𝜕 𝜕𝑥𝑗
[(𝜇 +
𝜇𝑡
)
𝜕𝑘
𝜎𝑘 𝜕𝑥𝑗
] + 𝐺𝑘 + 𝐺𝑏 − 𝜌𝜀 − 𝑌𝑀 + 𝑆𝑘
(7)
Dissipation rate, ɛ: 𝜕 𝜕𝑡
(𝜌𝜀) +
𝜕 𝜕𝑥𝑖
(𝜌𝜀𝑢𝑖 ) =
𝜕 𝜕𝑥𝑗
[(𝜇 +
𝜇𝑡
)
𝜕𝜀
𝜎𝜀 𝜕𝑥𝑗
𝑘2
𝜀
𝜀2
𝑘
𝑘
] + 𝐶1𝜀 (𝐺𝑘 + 𝐶3𝜀 𝐺𝑏) − 𝐶2𝜀 𝜌
+ 𝑆𝜀
(8)
Turbulent viscosity, 𝜇𝑡 = 𝜌𝐶𝜇 (9) 𝜀 where 𝐺𝑘 = generation of turbulence kinetic energy due to the mean velocity gradients, 𝐺𝑘 = generation of turbulence kinetic energy due to buoyancy, 𝑌𝑀 = contribution of the fluctuating dilatation in compressible turbulence to the overall dissipation rate, 𝐶1𝜀 , 𝐶2𝜀 = constants of 1.44, 1.92 and 𝐶𝜇 = constants of 0.09, 𝜎𝑘 , = Prandtl number for k, 1.0, 𝜎𝜀 = Prandtl number for ɛ, 1.3, and 𝑆𝑘 , 𝑆𝜀 = source terms. The RNG k-ɛ model is an enhancement from the standard k-ɛ model to improve the accuracy for rapid strained flow and swirl flow by additional terms in k and ɛ transport equations. The realizable k-ɛ model is another enhancement from the standard k-ɛ model. This model has a different formulation for turbulent viscosity, which causes this model to outperform other variations in several cases. Thus, the CFD results from these turbulence models were compared in the current study to obtain the most suitable turbulence models.
Mesh Independence Study Seven unstructured tetrahedral meshing (I – VII) with different numbers of mesh elements were performed on the model in relation to the relevant center in ANSYS meshing to study the mesh independency of the results. Typically, in mesh independence studies, the meshing increment step is 1.5 times from the coarser mesh. However, the increment step is not viable when the mesh element exceeds 3 x 106 cells. The computational time is too costly. Hence, a low incremental step of up to 1.24 times was used in the fine mesh setting to produce more than 5 x 106 cells. A higher increment step of up to 2.58 times was used in the coarse mesh to reduce the number of analyzing models in the current study. The computer used to run the simulation was incorporated with Intel core i7-4510U CPU @ 2.00GHz processor, 8.00GB RAM and 64-bit Window 10 Pro operating system. Due to the limited computational power available, the mesh was studied from the coarse mesh of about 200,000 cells until the room temperature difference of two consecutive mesh elements dropped below 0.01°C. Figure 2 shows that the room mean temperature is consistent for the 3 meshes with the mesh elements of 3 million and above. A further analysis has been conducted to compare the center plane velocity profiles of the three meshes [18]. Figures 3a and 3b show the comparison between meshes V & VI, and meshes VI & VII respectively. The plane velocity profile is almost identical for the case in meshes VI &VII. Thus, the model can be considered to be mesh independence at mesh VI.
Data Verification The CFD investigation of diffuser airflow pattern is problematic due to the complex geometry of the diffuser, and the size of the diffuser is relatively small compared with the room dimensions. A correct turbulent model is crucial to predict accurately turbulent airflow pattern and the heat transfer between the inflow and the room. Hence, data comparison and verification are important to ensure that the numerical modeling represents the best physical model. The standard k-ɛ turbulence model has been widely used for diffusers’ airflow prediction in OHAD systems such as round diffusers, square diffusers and swirl diffusers [12, 19, 20]. In this study, the standard k-ɛ, RNG k-ɛ, and realization k-ɛ are compared with the experimental measured data. The plane airflow pattern is then compared with the smoke test data to find the most suitable turbulence model for the present investigation. For comparison and verification purposes, the centerline air velocity was extracted from the simulation to compare with the experimental measurement. Figure 4 shows the vertical centerline air velocity profiles of the measured and the predicted data for few turbulence models. No perfect agreement between the measured and predicted data is expected. However, both the standard and RNG k-ɛ models indicate a similar centerline air velocity pattern to the measured data. In fact, the RNG k-ɛ model has a closer prediction to the measured data at the location near to the floor swirl diffuser outlet. There is a drop at around y/H = 0.20, and a spike at around y/H = 0.40 in all turbulence models. This fluctuation is observed mainly due to the swirl mechanism that causes difficulty in measuring turbulent flow using the existing tool (i.e. Hot-wire anemometer). It is noted that the computed and measured results are consistent with the results found by Fathollahzadeh [21]. It is also noted that the spike of air velocity in the computed results were not captured in the measured data (Figure 4). This could be due to the limitation of the measurement tool just mentioned, where in the current work, a hot-wired anemometer was used for the measurement. Hence, the predicted models are acceptable in the present research. In addition, both measured and simulated results are consistently showing a velocity drop at about y/H = 0.20 and an acceleration at about y/H = 0.40 (Figure 4). This phenomenon was captured because the swirl diffusers create a circular air flow near the air inlet, and as the airflow moving upwards, the airflow changed from the horizontal air movement to the vertical air movement. Figure 4 has shown that the X-Y
plane air velocity, where the air velocity at the Z-direction (horizontal air velocity), was not captured. Thus, the airflow development can be viewed as the changes of the horizontal air movement to the vertical air movement. Figure 5 shows the plane air velocity profiles for standard, RNG and realizable k-ɛ turbulence models as compared to the smoke test photograph. The standard k-ɛ turbulence model, which has a higher centerline velocity compared with the air velocity at the edges of the flow plume shows the closest match to the actual airflow pattern. The RNG k-ɛ turbulence model has a higher air velocity at the edges of the flow plume compared with the centerline air velocity. Hence, the standard k-ɛ turbulence model is used for the present floor swirl diffuser model investigation.
4.
Results and Discussion
The base model used to study the geometrical design changes is a 250 mm nominal diameter floor swirl diffuser with 18 diffuser blades at the 45-degree angle of attack, and 3 mm of grille thickness tested at 195 cubic meter per hour (CMH). The following results to be elaborated in the subsequent paragraphs show the effects due to the geometrical design changes of the floor swirl diffuser. The Effect of Airflow Rates This study was conducted to compare airflow rates of 100, 195 and 290 CMH, which are the lowest to the highest recommended airflow rates used for this typical floor swirl diffuser [22]. As indicated by Lin [9], the airflow rate has a strong influence on vertical temperature profiles where a higher airflow rate lowers the thermal stratification. For the purpose of understanding the thermal stratification of the UFAD system in terms of vertical velocity profiles, a comparison on the centerline air velocity is reported and shown in Figure 6. The characteristics of a pure jet are defined by thermal length scale, which is significant for thermal stratification. In UFAD systems where the heat load is consistent, the thermal stratification depends highly on the velocity of the supply air [23]. The thermal length scale is defined as below [24, 25]: 3/4
𝑙𝑚 =
𝑀0
1/2
𝐵0 𝑀0 = 𝑄0 . 𝑢0 𝑔. 𝑄0 . ∆𝑇0 𝐵0 = 𝑇𝑟 where l_m is the thermal length scale, M is the momentum flux, B is the buoyancy flux, Q is the volume flux (m3s-1), u is the supply air velocity (ms-1), g is the gravitational acceleration, ∆T is the temperature difference between supply air of the diffuser and the room temperature and Tr is the reference air temperature. For the case of 100 CMH, a terminal velocity of 0.25 ms-1 has reached at around 1.3 m above the floor swirl diffuser. An UFAD system with a vertical throw of less than or equal to 0.25 ms-1 at the height of 1.4 m above the floor can have a zone air distribution effectiveness of 1.2 according to ASHRAE Standard 62.1 [7]. It is the only case that could achieve the requirement to receive the highest ventilation effectiveness of 1.2. Thus, the outdoor airflow rates can be reduced by a factor of 1.2. The terminal velocity for 195 CMH and 290 CMH is around 2.4 m and 2.8 m respectively. Hence, there is insufficient space to create lower and higher stratified zones. It will be a mixing air distribution system as the thermal stratification cannot be formed. Due to the importance of supply air velocity towards the thermal stratification as described earlier, it should be noted that the supply air velocities for the three cases are 1.73 ms-1, 3.38 ms-1 and 5.03 ms1 respectively. It is also observed from Figure 6 that the swirl diffusers again create a circular air flow near the air inlet as described earlier in Figure 4.
Although the airflow rate of 100 CMH achieves a higher ventilation effectiveness, Figure 7 shows that the airflow rate is insufficient in reducing the temperature of the environment as effective as the airflow rates of 195 CMH and 290 CMH, where the surrounding space has a lower temperature around 298 K (25°C). Besides, the two cases with higher airflow rates create a greater air mixing in the area surrounding the diffuser’s discharge outlet. The air movement is about 0.1 ms-1, which is slightly below the recommended air movement by Malaysian Standard (0.15 – 0.5 ms-1 for offices) [26]. A larger number of diffusers per room is needed when the airflow rate is small. It is also pertinent to mention that the supply air temperature of an UFAD system should be higher (usually 16-18°C) than the OHAD system to make sure that the occupants will not experience cold feet, and therefore the air jet characteristic length is a key to calculate the air distribution performance [8].
The Effect of the Number of Diffuser Blades This study examines the effect of the number of diffuser blades on the airflow pattern of the floor swirl diffuser. Diffusers with 12, 18 and 24 blades, as shown in Figure 8, were simulated separately at the same conditions at 195 CMH. Note that the 18-blade diffuser is commonly used by various air terminal manufacturers. The percentage of the free area for 12, 18 and 24 blades is 43.27%, 18.86% and 5.51% respectively. The blade used in the study has a curvature profile adopted from the upper surface of airfoil NACA 0012 to enhance the aerodynamic airflow. The maximum air velocities that discharged from the under-floor swirl diffuser for 12, 18 and 24 blades are 3.24 ms-1, 3.89 ms-1 and 4.19 ms-1 respectively. x In the case of 12-blades swirl mechanism as shown in Figure 9, almost half of the area is a free area (i.e. 43.27%). It makes the angled swirl blades ineffective in creating a swirl effect. The primary air moves vertically upwards like an air jet without expansion. Secondary air movement is induced near to the primary air up to the ceiling height. The 18-blades swirl mechanism, in which its free area is 18.86%, shows a well-developed plume air development due to the entrainment of room air. It reduces the velocity of the primary air significantly and shortens the throw. Hence, the secondary air movement is induced at a larger low level area of the room. The 18-blades swirl mechanism creates a greater swirl air effect compared to the 12-blades swirl mechanism. The 24-blades swirl mechanism has the most blades and the least free area at 5.51%. However, the primary air-throw is higher compared to the case of 18-blades. The swirl effect does not grow greater when compared to the case of 18-blades. This is due to the limited free area for the air to discharge and causes the air to be forced out from the diffuser at high velocity. In this case, the angled blades do not perform effectively in directing the air to flow in the swirl effect. The Effect of Diffuser Blade’s Angle of Attack This study elaborates the angle of attack that affects the room air movement and the room temperature. The angles studied are 30°, 45° and 60°. The percentage of the free area for 30°, 45° and 60° is 14.25%, 18.86% and 33.84% respectively. Figure 10 shows the plane air velocity profiles at the angles of 30°, 45° and 60° respectively. The airflow patterns in the cases of 30° and 45° have a similar establishment. In both cases, the room at the lower level has induced an air movement of about 0.1 ms-1. In fact, the primary airflow in the case of 45°-blade developed a greater air entrainment from the room air at the height of 1 – 2 meters above the diffuser. However, the 30°-blade creates a more evenly distributed temperature across the room compared with the other two cases. The primary airflow pattern in the 60°-blade is similar to the case of the 12-blades diffuser where the air is moving vertically without the expansion like an air jet. This causes the mixing of air at the higher level. The results are consistent with the findings by Sajadi [12] where the ceiling based swirl diffuser performs optimally at the swirl angle of 32°.
The Effect of Diffuser and Grille Thickness The floor swirl diffuser is installed on the floor. The design of grille top of the diffuser blade is important to support the impact and the loads of human or object. A 20mm thick grille is commonly used in the market. Figure 11 shows the floor swirl diffuser examined. This study discusses the effects of the presence of diffuser blades and grille thickness that affect the airflow pattern in the room. Although the design of the grille will affect the minimum grille thickness required, the impact and load study have not been conducted in the present work. The grille thickness test emphasizes on the airflow performance. The five cases studied are no diffuser blades, no grille, 3 mm grille, 10 mm grille and 20 mm grille. Note that a diffuser without a grille fitted to the UFAD system will not function properly due to the fact that the diffuser blades without a thick grille cannot withstand impact and load of a human or an object. Nevertheless, it is included in the current study for comparison purpose in order to have a better understanding of the airflow patterns.
Figure 12 shows the center plane air temperature profiles for all cases studied. The diffuser blade plays a key role in making the supply air rotating and mixing with the room air. When there is no diffuser blade, the air is delivered to the space vertically upward like an air jet with a minimum entrainment. The result is similar to the finding on ceiling type swirl diffuser by Sajadi [12]. Besides, the swirl angle of the diffuser plays a vital role in defining the characteristics of the swirl jet, especially when the grille is absent. The velocity profiles at the centerline show that a throw height of 0.25 ms-1 is reached at about 1.8 m above the floor. The primary air has entrained the room air more effectively compared to other cases. This can be seen by the expansion of the plume air jet, and the hollow cone pattern has formed. The case is consistent with the swirl angle study performed by Tavakoli et al. [27] on the ceiling based swirl diffuser, and the complete theory and flow solver details can be found in references [28, 29]. As the grille thickness increases, it is observed that the primary air is delivered into the space vertically upwards with a higher air throw and velocities. The primary air jet has induced a secondary air flow near to the jet. However, the room temperature becomes uneven as the cold air is delivered straight to the return air grille on the ceiling. The results in the case of no swirl diffuser blade have shown similarity to cases as the grille thickness increases such as 10 mm and 20 mm grille thickness. The thickness of the grille has proven to be detrimental to the function of diffuser blades. The Centerline Velocity, Vc The case on 195 CMH is the base model used to study the geometrical design changes. The simulation result suggests a good agreement with the experimental measurement. Figure 13 shows the centerline velocity of all models through experimental measurement and simulation. Among all the comparisons, the centerline velocity graph shows that the grille thickness has a major impact on the airflow from the floor swirl diffuser. The cases with 10mm and 20mm thick grilles have resulted in high air velocities near the air outlet at 1.91 ms-1 and 2.06 ms-1 respectively. The case without diffuser achieves a similar result of high centerline air velocity near the air outlet at 1.88 ms-1. This has proven that the air is delivered into the space without a proper swirling effect. For the cases with no grille and 3mm grille thicknesses, the centerline air velocities are greatly reduced to maximum values of 0.57 ms-1 and 0.77 ms-1 respectively. Therefore, having a thick grille is equal to not having diffuser blades, and the air will be delivered to the space without creating the swirling effect.
5.
Conclusions
In the present study, a comprehensive investigation has been successfully conducted on the airflow pattern of a floor swirl diffuser through various aspects, including the airflow rate, number of diffuser blades, the angle of attack of diffuser blades and grille thickness. The results can be used as an important guide to design an optimised UFAD system. Based on the results obtained, there are several major conclusions made as follows:
1)
2)
3)
4)
6.
In the airflow rate study, a 250 mm nominal diameter floor swirl diffuser with an airflow rate of 100 CMH and below can be awarded with a ventilation effectiveness of 1.2 according to ASHRAE Standard 62.1:2013. The minimum outdoor airflow rate can be reduced by dividing it with a factor of 1.2 if the design is in accordance to this flowrate or below. The momentum flux of the diffuser is directly proportional to the supply air velocity. Thus, the equivalent supply air velocity of 1.73 ms-1 is acceptable. An airflow rate of 195 CMH and above supplied using this diffuser will not obtain this better ventilation effectiveness at 1.2 because a higher air throw will cause the low and high stratification layers to be mixed by the turbulences of the flow. The diffuser performance will then be identical to an OHAD mixing system of 1.0. The future research on thermal stratification is recommended to include heat load profiles of the room in order to have a complete understanding on thermal stratification. It is suggested that the optimum number of diffuser blades to create a swirl effect is 18 blades, which has a free area of 18.86%. The airflow of this case shows a well-developed plume air development (Figure 9). The air plume developed is the largest among the three cases. A large free area in the case of 12 blades or a small free area in the case of 24 blades could lead to a direct air jet flow into the space, and thus ineffective in the air plume development, which is similar to the case without diffuser blades. The swirl blades are important in rotating the discharged air to improve the mixing effect. At the low angles of attack of 30° and 45°, the air throw is lower compared to a high angle of attack. The primary air is established and expanded by entraining room air. The room achieves an evenly distributed temperature profile at the low angle of attack due to the better air mixing effect. The blade angle can have a different effect on thermal stratification in cooling or heating mode. This can be an extension for the future research work. A thicker grille leads to air delivered as a straightened air jet into the space. The swirl effect from the diffuser blades is eliminated. Thus, a minimum grille thickness should be developed without detrimental effect to the strength to withstand the impact and load of people and furniture.
Acknowledgement and Declaration
The authors would like to thank University of Malaya for providing UMRG Grant RG030/15AET to the authors for research work to be conducted at University of Malaya. Thanks are also extended to University of Malaya PPP Grants PG111-2012B and PG042-2016A for providing the partial financial assistance to the co-author, Mr. K.S. Poh, for conducting the research work at HVAC&R Lab at the Department of Mechanical Engineering, University of Malaya. In addition, special thanks are extended to Prudentaire Marketing Sdn Bhd for their contribution in the laboratory testing conducted in this project.
References 1 2 3 4
5 6 7 8 9 10 11 12 13 14 15 16 17 18
19
20
ASHRAE. UFAD Guide. Atlanta (GA): ASHRAE; 2013; 323. Bauman F, Webster T. Outlook for Underfloor Air Distribution. ASHRAE Journal 2001; 43(6): 18-27. Liu QA, Linden PF. The fluid dynamics of an underfloor air distribution system. Journal of Fluid Mechanics 2006; 554(1): 323-341. Webster T, Linden P, Buhl F, et al. Energy performance of underfloor air distribution systems. Final Project Report submitted to California Energy Commission (CEC) Public Interest Energy Research (PIER) Program. California (USA): Center for the Built Environment, University of California, Berkeley, CA; 2007. Pasut W. Using Ductwork to Improve Supply Plenum Temperature Distribution in Underfloor Air Distribution (UFAD) System. Padua (Italy): University of Padua; 2011. Schiavon S, Webster T, Dickerhoff D, et al. Stratification prediction model for perimeter zone UFAD diffusers based on laboratory testing with solar simulator. Energy and Buildings 2014; 82: 786-794. ASHRAE Standing Standard Project Committee 62.1. ANSI/ASHRAE STANDARD 62.1-2013: Ventilation for Acceptable Indoor Air Quality. Atlanta (GA): ASHRAE, 2013; 52. ASHRAE. HVAC Applications Handbook, Chapter 57: Room air distribution. Atlanta (GA): ASHRAE; 2015. Webster T, Bauman F, Reese J. Underfloor air distribution: Thermal stratification. ASHRAE Journal 2002; 44(5): 28-36. Lin YJP, Tsai TY. An experimental study on a full-scale indoor thermal environment using an underfloor air distribution system. Energy and Buildings 2014; 80: 321-330. Raftery P, Bauman F, Schiavon S, et al. Laboratory testing of a displacement ventilation diffuser for underfloor air distribution systems. Energy and Buildings 2015; 108: 82-91. Sajadi B, Saidi MH, Mohebbian A. Numerical investigation of the swirling air diffuser: Parametric study and optimization. Energy and Buildings 2011; 43(6): 1329-1333. Zhang K, Zhang X, Li S. Simplified model for desired airflow rate in underfloor air distribution (UFAD) systems. Applied Thermal Engineering 2016; 93: 244-50. ASHRAE Standing Standard Project Committee 113. ANSI/ASHRAE Standard 113-2013: Method of Testing for Room Air Diffusion. Atlanta (GA): ASHRAE, 2013. Mohammed RH. A simplified method for modeling of round and square ceiling diffusers. Energy and Buildings 2013; 64: 473-482. Tu J, Yeoh GH, Liu C. Computational Fluid Dynamics - A Practical Approach. 2nd edition, Heinemann, Butterworth, United Kingdom, 2013: 440. ANSYS, ANSYS Fluent Theory Guide. 2013, Canonsburg, Pennsylvania: ANSYS Inc. Nada SA, El-Batsh HM, Elattar HF, et al. CFD investigation of airflow pattern, temperature distribution and thermal comfort of UFAD system for theater buildings applications. Journal of Building Engineering 2016; 6: 274-300. Aziz MA, Gad Ibrahim AM, Mohammed El Shahat FA, et al. Experimental and numerical study of influence of air ceiling diffusers on room air flow characteristics. Energy and Buildings 2012; 55: 738746. Hu SC. Airflow characteristics in the outlet region of a vortex room air diffuser. Building and Environment 2003; 38(4): 553-561.
21 Fathollahzadeh MH, Heidarinejad G, Pasdarshahri H. Prediction of thermal comfort, IAQ, and energy
consumption in a dense occupancy environment with the under floor air distribution system. Building and Environment 2015; 90: 96-104. 22 Trox Malaysia. Floor Diffusers. https://www.troxapo.com/floor-diffusers/type-fba-%C2%B7-fbk82577427340e09dc. Retrieved May 2015. 23 Fan Y, Li X, Yan Y, Tu J. Overall performance evaluation of underfloor air distribution system with
different heights of return vents. Energy and Buildings 2017; 147: 176-87. 24 Arghand T, Karimipanah T, Awbi HB, et al. An experimental investigation of the flow and comfort parameters for under-floor, confluent jets and mixing ventilation systems in an open-plan office. Building and Environment 2015; 92: 48-60. 25 Zhang, K., Zhang X, Li S, et al. Experimental study on the characteristics of supply air for UFAD system with perforated tiles. Energy and Buildings 2014; 80: 1-6.
26 Department of Standards Malaysia. MS 1525:2014: Code of Practice on Energy Efficiency and Use of Renewable Energy for Non-Residential Buildings. Selangor (Malaysia): Department of Standards Malaysia, 2014; 46 27 Tavakoli E, Hosseini R. Large eddy simulation of turbulent flow and mass transfer in far-field of swirl diffusers. Energy and Buildings 2013; 59: 194-202. 28 Yau YH, Badarudin A, Rubini P. A Numerical Study of Secondary Flow and Large Eddies in a Driven Cavity. Journal of Mechanical Science and Technology (JMST) 2011; 26 (1): 1-11. 29 Yau YH, Badarudin A, Rubini P. Numerical and Scalability Analysis of an Unstructured Cartesian Flow Solver. Proceedings of the Institution of Mechanical Engineers, Part C, Journal of Mechanical Engineering Science 2011; 225: 2138-2148.
Figure 1: Schematic diagram of the experimental setup in the laboratory
30.72
Temperature (°C)
30.71 II (603810, 30.7)
30.70
V (3483738, 30.69)
VII (5612559, 30.69)
30.69 VI (4311451, 30.69)
30.68
IV (2103269, 30.68)
30.67
III (1174073, 30.66)
30.66
I (233695, 30.66)
30.65 -
10,00,000
20,00,000 30,00,000 40,00,000 Number of mesh elements
50,00,000
60,00,000
Figure 2: Room mean temperature at different mesh elements.
Figure 3a: Plane air velocity for Mesh V (top left), plane air velocity for Mesh VI (top right) and air velocity differences between Meshes V and VI (bottom).
Figure 3b: Plane air velocity for Mesh VI (top left), plane air velocity for Mesh VII (top right) and air velocity differences between Meshes VI and VII (bottom)
y/H
Experiment
1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.00
0.20
Standard k-ɛ
0.40
Realizable k-ɛ
0.60 0.80 1.00 1.20 1.40 Centerline velocity (ms-1)
RNG k-ɛ
1.60
1.80
2.00
Figure 4: Vertical centerline air velocity profile of an actual model and various predicted turbulence models
(a)
(b)
(c) (d) Figure 5: Plane air velocity profiles for (a) smoke test; (b) standard k-ɛ; (c) RNG k-ɛ; (d) realizable k-ɛ
y/H
100CMH
195CMH
290CMH
1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0
0.2
0.4
0.6 0.8 Centerline velocity (ms-1)
1
1.2
1.4
Figure 6: The centerline vertical air velocity profiles for airflow rates of 100, 195 and 290 CMH
(a)
(b)
(c)
Figure 7: The center plane air temperature profiles for airflow rates of (a) 100 CMH; (b) 195 CMH; (c) 290 CMH
(a) (b) (c) Figure 8: Diffuser’s swirl mechanism with (a) 12 blades, (b) 18 blades and (c) 24 blades
(a) (b) (c) Figure 9: The center plane air velocity profiles for (a) 12-blades; (b) 18-blades; (c) 24-blades swirl mechanism
(a) (b) (c) Figure 10: Plane air velocity profiles for diffuser blades at the angles of (a) 30°; (b) 45°; (c) 60°
Figure 11: Floor swirl diffuser
(a)
(b)
(c)
(d) (e) Figure 12: The center plane air temperature profiles for the cases (a) without diffuser blades; (b) without grille; (c) with 3 mm grille thickness; (d) with 10 mm grille thickness; (e) with 20 mm grille thickness
Figure 13: The centerline velocity (m/s) for all the geometrical variation of floor swirl diffuser from the floor level of 0 meter to 3 meters above floor level
Table 1: Boundary conditions of the CFD simulation. Boundary Details Supply air inlet Airflow rate = 100, 195, 290 CMH Air temperature = 18.0°C Relative humidity = 60% Exhaust air outlet 4 numbers of 600 x 600 mm ceiling exhaust air outlet at the four corners of the room at 0 Pa gauge pressure Walls Surface temperature = 28.5°C Ceiling Surface temperature = 27.0°C Floor Surface temperature = 27.0°C Grille Adiabatic Air plenum Adiabatic