Plenum airflow distribution and its influence on the performance of a diffuse ceiling ventilation

. Implementation of diffuse ceiling ventilation (DCV) is slowly gaining momentum and applications in building ventilation have taken off with countries like Denmark, Finland and Netherlands taking the lead in Europe. However, DCV is yet to gain a foothold in Sweden and so not many installations are known, and their performance in relation to Swedish building practice is not yet established. A school in southern Sweden was subsequently renovated and two classrooms were equipped with a sound-absorbent suspended ceiling compatible with DCV. DCV has possible benefits for educational environments including improved thermal comfort as well as lower costs and noise levels. However, it is currently still unknown how supply conditions in the plenum affect the diffusion of air and resulting conditions within the room. To assess airflow characteristics and whether these influence flow conditions in the classroom, we investigated and compared the performance of DCV with two different supply conditions in the plenum. Air speeds and temperature distribution measurements in the plenum and classroom were performed with constant temperature anemometers and thermocouples respectively. The general observation from this study and the system setup herein is that airflow and temperature characteristics in the classroom were independent of the airflow conditions in the plenum. Further investigations in a controlled climate chamber are recommended to investigate and optimise system performance in accordance with Swedish building practice.


Introduction
Different air distribution systems, like mixing systems that generate homogeneous indoor environmental conditions or stratification systems for heterogeneous indoor environmental (e.g., displacement or underfloor air distribution), are widely used. A review of different air distribution systems and their operational principles are discussed by Yang et al. [1] for more details. The chosen air distribution system will offer different advantages/disadvantages depending on the fluid dynamic principles employed when in operation. For example, mixing air distribution systems promote dilution of contaminants through increased turbulence transfer but is often problematic if not well designed as it promotes indoor contaminant flooding due to shortcut ventilation [2]. On the other, stratified air distribution systems focus on purging (removal) of contaminants from the occupied region but it is vulnerable to the distribution of heat sources in the room.
Recently, diffuse ceiling air distribution, which is a total mixing air distribution system, has gained attention for use in rooms with high heat loads like classrooms and open-plan offices. Literature suggests benefits for modern diffuse ceiling ventilation (DCV) systems, which are typically smaller, use less energy and create more uniform indoor environmental conditions [3]. The operational concept of DCV is to supply fresh air in the * Corresponding author: alan.kabanshi@hig.se plenum, above the ceiling tiles of the suspended ceiling. The air is then distributed in the plenum and eventually diffused into the room (occupied zone) through perforations (slots or holes) in the ceiling tiles or through gaps between ceiling tiles and the supporting grid profiles distributed throughout the entire ceiling area. The large opening area accorded by the diffusion surface results in low momentum diffusion of air from the plenum into the room. Since air is mostly supplied at low temperatures, the system relies on buoyancy forces generated by the heat sources to generate and drive airflow circulation and mixing in the room [3]. Zhang and Heiselberg [4] discussed three different ceiling diffusion types based on the air path, related products, and mixing ability with room air.
To our knowledge, the concept of DCV has not been adopted in building ventilation in Sweden, but other countries have implemented the systems with a lot of research and case studies in Denmark and Netherlands [5,6]. In addition, few studies have performed measurements on the airflow types in the plenum and how it influences the air distribution in the room. The airflow distribution in the plenum will differ depending on the choice and point of air release from the ventilation supply duct and the physical characteristics of plenum. For example, is the duct opening in the centre and supplying in only one direction? Are structural components like beams creating physical obstructions for airflow? Or how do service installations like fire sprinkler pipes, recessed ceiling lamps or exhaust ventilation ducts (which introduce heat transfer effects) interact with the plenum airflow? The current study explores two cases of airflow supply and direction in the plenum. The aim was to determine whether changes in airflow characteristics in the plenum influences the diffusion of air and the resulting conditions into the occupied room.

Methods
In this project, we performed field measurements at a school in Eslöv in the southern part of Sweden. The school was recently renovated, and two classrooms were fitted with DCV. Figure 1 shows measurement points and two different air supply conditions in the plenum. The supply airflow rate was about 280 l/s (0.28 m 3 /s) and no temperature or humidity control was applied. Thus, the outdoor air conditions were used as supply air conditions (measured supply temperature = 18.13 ºC ).

Plenum measurements
The plenum is dimensioned as shown in Figure 1 (top), and had overall depth of 0.8 m. A small area above the windows had a plenum extension with ceiling tiles laid in an upward trend to flush with the top plane of the plenum. No structural beams were present only exhast ventilation service pipes, as seen in Figure 1, and wiring cables for electrical installations (sensors and lamps). The room had ceiling grid made of 600x600x40 mm "Ecophon Master E T15" sound absorbing tiles. The air was supplied into the room from the plenum through gaps between ceiling tiles and the supporting grid profiles.
T-type thermocouples (class 1) with a tip diameter of 1.5 mm and a factory accuracy of ±0.3 º C were used in surface temperature measurements in both the plenum and the classroom. One central measurement point was chosen on each surface for the measurements. The thermocouples were calibrated in the range of 10 -30 º C with 2 Hz frequency operation mode with an uncertainty of ± 0.1 º C. CTA88 (Constant Temperature Anemometers) were used to measure the air speed and local temperature at 12 points about 0.2 m above the ceiling tiles. The CTA probes were calibrated for an air speed measurement range of 0.05 -3 m/s and had an air speed accuracy of 0.05 m/s and temperature range of 10 -40 º C with a sensor accuracy of 0.2 º C. The sampling interval for all measurements was set to 60 s, with the response time of 0.2 s to 90% of a step change.

Classroom measurements
The classroom is dimensioned as shown in Figure 2 and had a ceiling height of 3 m. 60 W incandescent lamps distributed in the room in 32 positions were used as heat sources to represent sensible thermal plumes generated by students and 100 W lamp was used to represent the heat source from the teacher. The classroom was unfurnished. Surface temperatures were measured with thermocouples and room conditions were measured at points shown in Figure 2, with room air temperature measured with thermocouples and air speed with CTA. The measurements were performed at the height of 0.1, 0.6, 1.1, 1.7 and 2.6 meters from the floor at all points in positions shown in Figure 2 for CTA and thermocouple measurements. CTA measured temperatures are not considered in the current analysis. 3 Results and discussion. Figure 3 shows a contour plot of plenum air speeds. This is an aerial capture of the plenum plan and the velocity distribution at different points in the measured plane. As can be seen, the air speed distribution was very different between the two conditions. The single-sided supply registered non-homogeneous air speeds and high values on one side of the plenum which was exposed to the supply jet throw. In this condition the air speed in the plenum varied from 0.162 m/s to 1.08 m/s. This implies that the pressure distribution in this setup varied across the plane, and it is probable that this influenced the diffusion of air through the ceiling. On the other hand, two-sided supply has a much more evenly distributed air speeds with a minimum being 0.14 m/s and a maximum of 0.35 m/s. The implication here is that if the pressure distribution is quite uniform and thus airflow supply diffusion into the room will also be uniform across the ceiling plane.  Figure 4 shows the air temperature distribution in the plenum. The two-sided supply showed a much cooler air distribution than the single-sided supply. Since the supply temperatures are critical in DCV, minimizing heat gain in the plenum is important to minimize the loss in the cooling capacity of the supply air. Also, a well distributed temperature profiles helps prevent variations in airflow currents in the occupied room. This can have implications on occupants' perception of draft (i.e., thermal comfort) where certain points would be perceived as drafty if the air currents are much cooler than other regions.

Plenum air speed and temperature distribution.
In both cases, the wall close to the school corridors was warmer and this might have implications for heat exchange between the supply air and the surfaces in the plenum. It is important to investigate the influence of having different surface temperatures and heat transfer characteristics in the plenum and how this will affect the performance of DCV. This also applies for service ducts that have an effect of introducing heat transfer between the plenum air temperature and the service installations. In this study we did not investigate this as the average air temperature values were similar between the two conditions.  Figure 5 shows the room air speeds at different heights. the measured air speeds were less than 0.2 m/s which is in compliance with recommended indoor air speeds according to ISO 7730. No differences were observed between the conditions although there were slightly higher velocities in the centre (Point E) at 1.1 m and air speed variations at 0.1 m which can be attributed to convective flows due the heat loads (lamps). For most of the points other than 0.1 m height and in both conditions, the recorded air speeds are within a standard deviation of less than 0.03 m/s implying that the distribution in the room was quite homogeneous. Thus, there is low risk of draft perception in the room.  Figure 6 shows the room temperature values measures at different locations and heights in the room. As observed in each condition, the temperatures at different points are close to each other showing that the room temperature field is homogeneous. In DCV, temperature is driving the airflows in the room through buoyancy, and as such a level of stratification is expected. However, the measurements do not show any indication of vertical temperature gradients, except at 0.1 m in the singlesided supply condition where the temperature is higher by about 0.5 ºC . Measurements with regards to tracer gas and thermal comfort were also conducted. No significant differences in the room were observed between the two plenum conditions. However, further measurements could provide insights on the plenum conditions and a follow-up study is planned to assess system performance over a service year.

Room air speed and temperature distribution.
Additionally, more detailed measurements in a controlled climate chamber are planned in order to evaluate different plenum setups and conditions and whether they have an influence on the resulting room conditions. There is additionally a need to investigate the performance of diffuse ceiling ventilation in Sweden since it is not common practice to cool or dehumidify the supply air. Whether DCV would perform equally well with fluctuating or varying supply conditions, and which factors need to be considered regarding Swedish building regulations are potential areas of future investigation.

Conclusions
The current study evaluated how airflow characteristics and temperature distribution in the plenum influence the air speed and temperature profiles in the room. Using the setup and conditions reported herein, we can deduce that differences in air speed supply and temperature in the plenum had minimal/no influence on the temperature and air speed distribution in the room with diffused ceiling ventilation system. However, further studies are recommended to assess system performance in controlled environments and how the system can be optimized in accordance with the Swedish building codes and practice.