Air distribution in a room with ceiling-mounted diffusers--comparison with wall-mounted diffuser, vertical ventilation, and displacement ventilation.

ABSTRACT

Experiments with room air distribution that is generated by a radial ceiling-mounted diffuser and a diffuser generating flow with swirl are compared with the air distribution obtained by mixing ventilation from a wall-mounted diffuser, vertical ventilation, and displacement ventilation. The air distribution generated by a radial diffuser is partly controlled by the momentum flow from the diffusers and partly from gravity forces where the thermal load and the temperature difference between room air and supply air deflect the radial wall jet down into the occupied zone. The ceiling diffuser with swirling flow generates a flow pattern in the room that is rather uninfluenced by the thermal load. The flow is highly mixed above the occupied zone, and the air movement penetrates the occupied zone close to the walls.

All systems were tested in the same room with a load consisting of two manikins, each sitting at a desk with two PCs and two desk lamps, producing a total heat load of 480 W.

In all five cases, the design of the air distribution system was based on flow elements from the diffuser, a maximum velocity assumption, and a critical vertical temperature gradient in the room. The characteristics of the air distribution systems are addressed by analyzing the acceptable conditions for the supply flow rate and the temperature difference for the different systems.

This paper shows that an air distribution system with ceiling-mounted air terminal units is able to generate comfortable velocity and temperature conditions at the same and at slightly higher loads as can be obtained by a vertical ventilation system, a mixing ventilation system with wall-mounted diffuser, and a displacement ventilation system with a low-velocity wall-mounted diffuser.

The comparison is extended by considering both the local discomfort caused by draft and the percentage of dissatisfied due to the temperature gradient when this is relevant to the systems. The draft rating is low for the ceiling-mounted diffusers as well as for the low impulse system (textile terminals), and the temperature gradient is also low because of the high level of room air mixing.

INTRODUCTION

The aim of an air-conditioning system is to remove excess heat in a room and replace room air with fresh air to obtain a high air quality. It is not sufficient to remove heat and contaminated air; it is also necessary to distribute and control the air movement in the room to create thermal comfort in the occupied zone.

Most air distribution systems are based on mixing ventilation with ceiling or wall-mounted diffusers or on displacement ventilation with wall-mounted low-velocity diffusers. New principles for room air distribution are introduced with the textile terminals mounted at the ceiling and radial diffusers with swirling flow also mounted at the ceiling.

This paper addresses five air distribution systems in all, namely, one low impulse system based on a textile terminal, mixing ventilation from a wall-mounted terminal, displacement ventilation from a wall-mounted low-velocity diffuser, mixing ventilation from a ceiling-mounted diffuser, and mixing ventilation from a ceiling-mounted diffuser with a swirling flow. All of the experiments are summarized in Table 1.

The two systems--a ceiling-mounted radial diffuser and a ceiling-mounted swirl diffuser--are discussed in detail in this paper, while the four other systems have been discussed in earlier papers (see Jacobsen et al. [2002], Nielsen et al. [2003], and Nielsen et al. [2005]). All of the systems are compared with respect to the design strategies. The supply flow rate [q.sub.0] and the temperature difference [DELTA][T.sub.o] between return and supply are chosen as design parameters. The local discomfort caused by draft rating and the dissatisfied due to the temperature gradient are also addressed.

Test Room

Figure 1 shows the full-scale room and the location of the ceiling-mounted diffusers. The dimensions of the room are in accordance with the requirements of the International Energy Agency Annex 20 work with length, width, and height equal to 4.2 m, 3.6 m, and 2.5 m, respectively.

Figure 2 shows the furnishings and the heat load of the room. The heat load consists of two PCs, two desk lamps, and two manikins, producing a total heat load of 480 W. The room dimensions, the furnishings, and the heat load are identical to the earlier test cases with mixing and displacement ventilation (see Jacobsen et al. [2002], Nielsen et al. [2003], and Nielsen et al. [2005]). Therefore, it is possible to make a direct comparison between five different air distribution systems.

The temperature distribution is measured along three vertical lines and in supply and return with type-K thermocouples connected to a data logger. The velocities are measured along a line at the height of 1.8 m (height of the occupied zone). The measurement line is moved down along the whole length of the room. Velocities are also measured at the head, chest, and ankle height of the two thermal manikins and on the two tables. All velocity measurements are made by 18 hot sphere anemometers connected to a multichannel flow analyzer.

[FIGURE 1 OMITTED]

The thermal load is constant in all experiments and the flow rate has been raised from 2.75 [h.sup.-1] up to 10.45 [h.sup.-1] with a variation in the temperature difference [DELTA][T.sub.o] between 12 K and down to 3.5 K. The room temperature is close to 22.9[degrees]C in all the experiments.

[FIGURE 2 OMITTED]

The Design Graph for the Ceiling-Mounted Diffusers

Three important parameters are considered in the design of room air distribution. The parameters are the air velocity (draft), the vertical temperature gradient, and the asymmetry in the mean radiant temperature. The air velocity can either be the velocity [u.sub.ocz] when it penetrates the upper boundaries of the occupied zone (1.8 m above the floor), the velocity [u.sub.01] close to the ankles of the manikins, or [u.sub.11] in head height of the seated manikins.

The principles of similarity show that any dimensionless velocity in the room can be given as a unique function of the Archimedes number if the flow in the room is a fully developed turbulent flow (high Reynolds number flow) (see Tahti and Goodfellow [2001]). The similarity principle is used to obtain a formulation of the velocity level for [u.sub.ocz], [u.sub.01], and [u.sub.11].

[u.sub.ocz]/[u.sub.o] = fun[c.sub.1] (Ar) (1)

[u.sub.01]/[u.sub.o] = fun[c.sub.2] (Ar) (2)

[u.sub.11]/[u.sub.o] = fun[c.sub.3] (Ar) (3)

The Archimedes number (Ar) is given by

Ar = [[beta]gd[DELTA][T.sub.0]]/[u.sub.o.sup.2] (4)

where [beta], g, d, and [DELTA][T.sub.o] are the thermal expansion coefficient, gravitational acceleration, length scale in diffuser, and temperature difference between return and supply flow, respectively; [u.sub.o] is the supply velocity given as a face velocity [q.sub.o]/[a.sub.o]; [q.sub.o] is the flow rate to the room; [a.sub.o] is a reference area of the diffuser (0.0226 [m.sup.2] for the radial diffuser and 0.0314 [m.sup.2] for the diffuser with swirl).

[FIGURE 3 OMITTED]

Figures 3a and 3b show the measured dimensionless velocities [u.sub.ocz]/[u.sub.o], [u.sub.01]/[u.sub.o], and [u.sub.11]/[u.sub.o] as functions of the Archimedes number. It is seen that the velocity at the upper surface of the occupied zone outside the vertical plumes ([u.sub.ocz]) in both cases is higher than the velocity at head ([u.sub.11]) and ankle ([u.sub.01]) height of the manikin. Although the velocity [u.sub.ocz] at the height of 1.80 m is larger than [u.sub.01] and [u.sub.11], it will never be a limiting design velocity because it only exists outside the thermal plume of the persons. The persons are protected from this velocity by their own thermal plumes in all of the experiments given in Figures 3a and 3b.

The experiments show that the velocity at ankle height ([u.sub.01]) is higher than the velocity at head height ([u.sub.11]). The velocity at ankle height is therefore the limiting parameter for the air distribution system.

Figure 4 describes the idea behind the design graph. The graph is based on the minimum and maximum allowable flow rate [q.sub.o] to the room and on the maximum temperature difference between return and supply. It is expressed as a [q.sub.o]-[DELTA][T.sub.o] graph. Figure 4 shows that it is necessary to have a minimum flow rate of fresh air to the room to obtain the necessary air quality. It is also expected that there must be a limit for the maximum flow rate to avoid draft in the room as well as a maximum flow rate in connection with the design of the duct system. The temperature difference [DELTA][T.sub.o] between return and supply is also restricted. A too-high temperature difference may either cause draft in the occupied zone or create a temperature gradient in the room that is too large. There is also a restriction on the temperature difference [DELTA][T.sub.o] in connection with design of the cooling system.

[FIGURE 4 OMITTED]

Figure 4 indicates an area for [q.sub.o] and [DELTA][T.sub.o] that supplies sufficiently fresh air and ensures a draft-free air movement in the occupied zone plus a restricted vertical temperature gradient. This area is the design area for a given air distribution system.

Figure 5 shows the design graph for the two mixing ventilation systems with ceiling-mounted diffusers. The [q.sub.o] - [DELTA][T.sub.o] values are restricted to a level where [u.sub.01] and [u.sub.11] are smaller than 0.15 m/s to avoid draft. The draft at the ankles causes most of the restrictions indicated in Figure 5.

The experiments with the radial diffuser show that [u.sub.01] equal to 0.15 m/s corresponds to a [u.sub.ocz] value close to 0.3 m/s as well as a maximum [DELTA][T.sub.o] value of 10 K. It can thus be concluded that [u.sub.ocz] equal to 0.3 m/s combined with [DELTA][T.sub.o] = 10 K should be the maximum entering velocity in the occupied zone and the maximum temperature difference for a design with a radial diffuser.

The experiments with the swirl diffuser show that [u.sub.01] equal to 0.15 m/s corresponds to a [u.sub.ocz] value in the range of 0.3-0.4 m/s. It can thus be concluded that [u.sub.ocz] equal to 0.35 m/s should be the maximum entering velocity in the occupied zone for a design with a ceiling-mounted diffuser with swirl.

The design area for the diffuser with swirling flow is located at lower airflow rates than the design area for the radial diffuser. A swirl diffuser with large supply openings could, if necessary, move the design area toward large airflow rates because an increase in the supply area decreases the corresponding velocity level in the occupied zone. There is no upper limits for [DELTA][T.sub.o] for this diffuser. The high mixing rate in the initial flow removes temperature differences in the flow and eliminates a large part of the gravity forces.

[FIGURE 5 OMITTED]

It is also a requirement that the temperature gradient dT/dy be smaller than 3 K/m, but this is easily obtained and does not add further restrictions to the [q.sub.o]-[DELTA][T.sub.o] graph.

Local Discomfort in a Room Ventilated by the Ceiling-Mounted Diffusers

The design models discussed in the previous section can only give the limits for the operation of the air distribution system. It is necessary to consider thermal comfort for all flow rates if the system has to be optimized.

The thermal environment often shows temperature gradients, velocity gradients, different turbulence levels, and an asymmetric radiant temperature distribution. The local discomfort that is the result of this environment is found from measurements of the local values of air temperature, air velocity, and turbulence level and from measurements of surface temperatures or asymmetric radiant temperatures (see Fanger and Langkilde [1975], Olesen et al. [1979], Fanger et al. [1989], and Toftum et al. [1997]).

The number of dissatisfied because of draft, the draft rating (DR), is used as a measure of local discomfort. The draft rating (DR) is defined as

DR = [e.sup.d([t.sub.a] - 24)](34 - [t.sub.a])(u - 0.05)[.sup.0.62](0.37uTu + 3.14), (5)

where [t.sub.a], u, and Tu are ambient air temperature, air velocity, and turbulence intensity, respectively (see Toftum et al. [1997]); d is a factor that is dependent on direction.

[FIGURE 6 OMITTED]

Equation 5 is especially based on a situation where the velocity is exposed to the back of the neck of the subjects. In this paper the highest velocity is often found at ankle level, and the sensitivity at this position is slightly lower than that of the back of the neck. The draft rating given from Equation 5 (shown in Figures 6 and 8) therefore represents a conservative estimate.

Figures 6a and 6b show the percentage of dissatisfied because of draft, DR. The draft rating is based on the velocity measured at the height of 0.1 m. It is shown that the diffuser with swirling flow has the lowest draft rating for an airflow rate below 0.05 [m.sup.3]/s. The draft rating of 5% is obtained for a flow of 0.03 [m.sup.3]/s. The radial diffuser has the optimal conditions at a higher flow rate with DR equal to 7.5% for a flow rate [q.sub.o] of 0.055 [m.sup.3]/s. Both diffusers are generating a similar level of draft at both positions A and B in the room.

The measurements show small temperature gradients and a low level of asymmetric radiation. This could be explained by the high mixing of air in the room.

Comparison between Five Different Air Distribution Systems

It is possible to make a direct comparison between the different air distribution systems because the experiments were done with the same office equipment: two computers, two desk lamps, and two manikins. The thermal load is in all cases equal to 480 W.

The air distribution systems are:

* mixing ventilation with a wall-mounted diffuser

* displacement ventilation with a wall-mounted low-velocity diffuser

* vertical ventilation with a ceiling-mounted textile terminal

* mixing ventilation with a ceiling-mounted radial diffuser

* mixing ventilation with a ceiling-mounted swirl diffuser

The limiting design parameter for mixing ventilation generated by a wall-mounted diffuser is the velocity [u.sub.ocz] of the jet when it penetrates the upper boundaries of the occupied zone in the case of non-isothermal flow. The vertical temperature gradient is not considered to be important for this case. Nielsen et al. (2003) and Jacobsen et al. (2002) show details of this design method.

Displacement ventilation has a high air velocity in the stratified flow at the floor. Reduced velocity in the occupied zone is therefore obtained by restricting the velocity in the stratified flow where it enters the vertical boundary of the occupied zone in front of the diffuser. This is expressed by the length of the adjacent zone [l.sub.n], which is the distance from the diffuser to a given velocity level in the stratified flow (see Nielsen [2000] and Skistad et al. [2002]). The vertical temperature gradient is also important in displacement ventilation and should be kept below a certain level.

The limiting design parameter for vertical ventilation from a ceiling-mounted textile terminal is the velocity [u.sub.ocz] of the downward-directed jet established when the flow is non-isothermal. The vertical temperature gradient is not important for this type of flow because the thermal plumes from heat sources and the downward-directed displacement flow from the terminal generate a large mixing of the room air (see Nielsen et al. [2005]).

Figure 7 shows the design graph for all five air distribution systems. Air quality considerations require a flow rate that is larger than 0.02 [m.sup.3]/s for all systems.

The maximum penetration velocity through the occupied zone [u.sub.ocz] is equal to 0.4 m/s in the case of mixing ventilation from an end wall-mounted diffuser. The rather high velocity seems to give the same level of comfort as obtained by the textile terminal for [u.sub.ocz] equal to 0.2 m/s.

The length of the adjacent zone [l.sub.n] is given to 1 m with a reference velocity of 0.2 m/s in the case of displacement ventilation, and the temperature difference [DELTA][T.sub.o] is in principle restricted to 12.5[degrees]C in the room, giving a gradient of 3 K/m. The curves in Figure 7 show the corresponding area for fulfilment of thermal comfort in the [DELTA][T.sub.o]-[q.sub.o] diagram. It should be mentioned that there is a possibility of reducing the minimum airflow rate for the displacement ventilation system because the stratification of supply air improves the air in the inhalation zone when the contaminant source in the room is also a heat source. This effect has not been considered in Figure 7.

Figure 7 indicates that, to some extent, the room has the same level of comfort (in terms of maximum velocity and temperature gradient) in the case of mixing ventilation with a wall-mounted diffuser or displacement ventilation with a wall-mounted low-velocity diffuser. The figure shows also that the vertical ventilation (low impulse) systems are superior to both mixing ventilation based on a wall-mounted diffuser and to displacement ventilation. In connection with vertical ventilation it should be noticed that the layout with two workplaces in a small room may result in a [q.sub.o]-[DELTA][T.sub.o] curve, which is favorable compared with the situation in a large room with vertical ventilation. (The lower curve in Figure 7 for vertical ventilation is obtained for one manikin in the room, corresponding to two persons in a larger room).

[FIGURE 7 OMITTED]

Mixing ventilation generated by ceiling-mounted diffusers is able to handle a higher heat load than any of the other systems.

It should also be emphasized that the design graph, Figure 7, to some extent is dependent on room size, room layout, and the layout and design of the terminal units (number, location, etc.).

It is possible to use the method behind the design graph in larger rooms rather than the test room in this paper. It is necessary to modify the flow elements (wall jet, penetration length, stratified flow, etc.) to the actual room dimension, which results in a specific design graph for the larger room.

Local Discomfort

Figures 8a and 8b show the draft rating (DR) for the persons at positions A and B. The draft rating is dependent on the location of the workplace in mixing and displacement ventilation, which, however, is not the case for the vertical ventilation system. Position A is the workplace located farthest from the wall with the inlet openings, and position B is the workplace closest to this wall. Generally, the best results are obtained for displacement ventilation with a workplace far away from the diffuser, while mixing ventilation with a wall-mounted diffuser has a high draft rating. The vertical ventilation system gains from the high level of buoyant flow in the room, which, in this case, protects the manikins from draft. Mixing ventilation with a ceiling-mounted diffuser generating flow with swirl has a draft rating close to the best result obtained by displacement ventilation at position A. The draft rating for the mixing ventilation with a ceiling-mounted radial diffuser also shows a low value at both positions under the optimal conditions of around 0.055 [m.sup.3]/s. All of the systems with ceiling-mounted diffusers (radial diffuser, swirl diffuser, textile terminal) are superior to systems with end wall-mounted diffusers.

Measurements show that the temperature gradient and asymmetric radiation are only important for displacement ventilation.

CONCLUSIONS

Air distribution systems based on mixing ventilation and ceiling-mounted diffusers are able to generate comfortable conditions up to a thermal load that is slightly superior to systems with mixing ventilation from wall-mounted diffusers, displacement ventilation with a wall-mounted low-velocity diffuser, or vertical ventilation systems with ceiling-mounted textile terminals.

It is characteristic that a mixing ventilation system with a ceiling-mounted swirl diffuser does not give a restriction on the temperature difference between return and supply within the level of traditional design practice. This effect is seen in Figures 5 and 7 as a lack of limitation of [DELTA][T.sub.o]. A similar effect can be seen for the vertical ventilation system, which is without restrictions on both temperature difference [DELTA][T.sub.o] and flow rate [q.sub.o].

[FIGURE 8 OMITTED]

The draft rating (DR) is low for a ceiling-mounted mixing ventilation system with an optimal design and restricted heat load. A mixing ventilation system with an end wall diffuser and a displacement ventilation system with an end wall-mounted low-velocity diffuser both have a higher draft rating under the design conditions.

REFERENCES

Fanger, P.O., and G. Langkilde. 1975. Interinvidual differences in ambient temperature preferred by seated persons. ASHRAE Transactions. 81(2):140-47.

Fanger, P.O., A.K. Melikov, H. Hanzawa, and J. Ring. 1989. Turbulence and draft. ASHRAE Journal. April.

Jacobsen, T.S., P.V. Nielsen, R. Hansen, E. Mathiesen, and C. Topp. 2002. Thermal comfort in a mixing ventilated room with high velocities near the occupied zone. ASHRAE Transactions 108(2).

Nielsen, P.V. 2000. Velocity distribution in a room ventilated by displacement ventilation and wall-mounted air terminal devices. Energy and Buildings 31(3)179-87.

Nielsen, P.V., T.S. Larsen, and C. Topp. 2003. Design methods for air distribution systems and comparison between mixing ventilation and displacement ventilation. Proceedings of Healthy Buildings 2003, Singapore.

Nielsen, P.V., C. Topp, M. Soennichsen, and H. Andersen. 2005. Air distribution in rooms generated by a textile terminal--Comparison with mixing ventilation and displacement ventilation. ASHRAE Transactions.

Olesen, B.W., M. Scholer, and P.O. Fanger. 1979. Vertical air temperature differences and comfort. In Indoor Climate, P.O. Fanger and O. Valbjoern, eds. Danish Building Research Institute, Copenhagen, pp. 561-79.

Skistad, H., E. Mundt, P.V. Nielsen, K. Hagstrom, and J. Railio. 2002. Displacement ventilation in non-industrial premises. REHVA Guidebook No 1.

Tahti, E., and H. Goodfellow. 2001. Handbook of Industrial Ventilation. San Diego: Academic Press.

Toftum, J., G. Zhou, and A. Melikov. 1997. Effect of airflow direction on human perception of draught. Proceedings of Clima 2000, Brussels.

Peter V. Nielsen, PhD

Thomas Heby

Bertil Moeller-Jensen

Fellow ASHRAE

Peter V. Nielsen is a professor in the Department of Building Technology and Structural Engineering, Aalborg University, Aalborg, Denmark. Thomas Heby is with I-klima, Aalborg. Bertil Moeller-Jensen is with Ramboell Denmark, Aalborg.

Table 1. Five Different Air Distribution Systems

System                    Supply                       Return

Mixing ventilation        Radial ceiling diffuser      End wall-mounted
                                                         below ceiling
Mixing ventilation        Ceiling swirl diffuser       End wall-mounted
                                                         below ceiling
Mixing ventilation        End wall-mounted             Return opening
                                                         below supply
                                                         terminal
Displacement ventilation  End wall-mounted             End wall-mounted
                                                         below ceiling
Vertical ventilation      Ceiling-mounted low impulse  End wall-mounted
                            textile terminal             at floor level