SEARCH

SEARCH BY CITATION

Keywords:

  • Convective boundary layer;
  • Thermal manikin;
  • Particle image velocimetry;
  • Pseudocolor visualization;
  • Ventilation flow;
  • Airflow interaction

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

This study investigates the interaction between the human convective boundary layer (CBL) and uniform airflow with different velocity and from different directions. Human body is resembled by a thermal manikin with complex body shape and surface temperature distribution as the skin temperature of an average person. Particle image velocimetry (PIV) and pseudocolor visualization (PCV) are applied to identify the flow around the manikin's body. The findings show that the direction and magnitude of the surrounding airflows considerably influence the airflow distribution around the human body. Downward flow with velocity of 0.175 m/s does not influence the convective flow in the breathing zone, while flow at 0.30 m/s collides with the CBL at the nose level reducing the peak velocity from 0.185 to 0.10 m/s. Transverse horizontal flow disturbs the CBL at the breathing zone even at 0.175 m/s. A sitting manikin exposed to airflow from below with velocity of 0.30 and 0.425 m/s assisting the CBL reduces the peak velocity in the breathing zone and changes the flow pattern around the body, compared to the assisting flow of 0.175 m/s or quiescent conditions. In this case, the airflow interaction is strongly affected by the presence of the chair.

Practical Implications

Interaction of the human convective boundary layer flow with the surrounding flows modifies the airflow around the human body, affects spread of pollution, personal exposure and convection heat transfer from the body. Understanding the nature of this interaction can be used for better distribution of ventilation air and improving occupants' thermal comfort.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

Understanding the air movement in the indoor environment is important as people spend 80–90% of their time indoors (Spengler and Sexton, 1983). One of the most common ways to reduce human exposure to airborne pollutants is to use mechanical ventilation. In current room air distribution design practice, airflows induced by the building occupants are not taken into account resulting in inaccurate prediction of the personal exposure. Several recent studies have challenged existing ventilation standards by showing that increase in the air change rate can lead to the exposure increase (Bolashikov et al., 2012; Pantelic and Tham, 2013; Popiołek et al., 2012). These studies emphasize the importance of understanding the complex interactions between airflows generated by the ventilation system and buoyant airflows induced by human body heat.

A healthy human body dissipates substantial part of metabolic heat by means of convection (Murakami et al., 2000; Zukowska et al., 2012). The convective heat loss, caused by the temperature gradient between a human body surface and cooler surrounding air, induces upward natural flow of the surrounding air, creating a convective boundary layer (CBL) around the human body. The CBL further rises above the head developing into a human thermal plume. Understanding the physics of the human CBL is essential to both indoor air quality and thermal comfort, because it has the ability to transport potentially contaminated air to the breathing zone and influences the body heat release, respectively.

Several studies in the past investigated the CBL in a quiescent indoor environment. Lewis et al. (1969) found that a nude standing man induces a buoyant flow which remains laminar at the lower leg region and develops into a fully turbulent flow at about 1.5 m from the floor. Homma and Yakiyama (1988) revealed that the CBL has a thickness of 150 mm at the head level with a mean velocity up to 0.25 m/s. These studies as well as several other studies performed in a quiescent environment (Clark and Toy, 1975; Craven and Settles, 2006; Licina et al., 2014) provide a basic knowledge of the airflow behavior in the vicinity of a human body. Air motion in mechanically ventilated spaces is highly unpredictable and influenced by momentum flux at the supply terminal, type and location of supply and exhaust terminals, room geometry, movement of the occupants, obstacles and furniture, and thermal loads in the room (Etheridge and Sandberg, 1996). These suggest that mechanical ventilation will generate different airflow patterns in every room, hence any generalization is challenging. In practice, human CBL interacts with the airflow in the occupied spaces. To understand the phenomenon, a simplified approach is needed which is based on the physics of the interaction between the CBL and the surrounding airflows. In such an approach, the interacting flows can be assisting, opposing, or transverse to each other. So far, little is known about the nature of the interaction between these flows.

The study that used a combination of two personalized ventilation systems coupled with mixing or displacement ventilation highlighted the importance of airflow interaction around the human body on inhaled air quality (Melikov et al., 2003). Interaction between the CBL and surrounding airflows is largely influenced by the characteristics of the invading flow generated by the air delivery systems, such as velocity, direction, and turbulence intensity. These factors will determine the extent of disturbance that invading flow exerts over the human natural convection flow. Interaction between the CBL and locally supplied personalized ventilation airflow from the front (Bolashikov et al., 2011a), assisting from below (Bolashikov et al., 2011b), and opposing from above (Yang et al., 2009) has been studied and reported. An airflow of 6 l/s supplied from the front was able to penetrate the CBL, while personalized jet at the lower flow rate was deflected upwards by the CBL without reaching the breathing zone. In the study with two assisting confluent jets, increased airflow from below led to an increase in the velocity in the breathing zone. The velocity increase from 0.17 to 0.34 m/s was recorded across the mouth, when the airflow of 4 l/s increased to 10 l/s. These studies investigated only the local airflow interaction in the breathing zone and did not focus on how airflow patterns in the occupied spaces globally affect the development of the CBL.

Melikov and Zhou (1996) studied airflow interaction between the human free convection flow and uniform horizontal airflow from behind. Results indicated that the convection flow with a maximum velocity of 0.13 m/s measured at the neck of a seated manikin was penetrated by invading flow with the velocity of 0.1 m/s, resulting in 4°C lower air temperature near the skin. Similar study performed in a low-speed uniform environment revealed that heated and non-heated manikin creates very distinctive airflow patterns with respect to the free air stream (Johnson et al., 1996). Both studies focused only on the interaction between the human convection flow and the airflow approaching from behind the manikin. In another, low-speed wind tunnel study, it was found that the human metabolic heat considerably alters the airflow patterns on the downwind side of the child-sized manikin and that increased wind velocity reduces the importance of the human convection flow (Heist et al., 2003).

As seen, most of the measurements on the interaction between airflows induced by the human body heat and the surrounding airflows have been carried out in the wind tunnel that gives limited body orientation with respect to the free stream or localized invading flows with relatively small cross-section area. Neither of the previous studies considered interaction between the entire CBL and assisting flow from below nor opposing flow from above. Previous experiments have also shown that increase in the free stream velocity decreases the relative importance of the buoyancy effect produced by the human body. The widely adopted velocity at which invading flow starts to disturb the CBL is 0.1 m/s or above (Bjørn and Nielsen, 2002; Melikov and Zhou, 1996). However, these findings apply only to disturbance due to horizontal airflow or due to a walking person. None of the studies in the past considered the velocity magnitude at which opposing and assisting airflows create disturbance to the CBL. Finally, most of the aforementioned studies had been performed with point-wise measurement technique, which may not be sufficient to fully understand the nature of air movement near the human body.

This paper presents findings from an experimental investigation of the CBL and its interaction with opposing flow from above, transverse flow from front, and assisting flow from below. Measuring techniques employed for this study consist of particle image velocimetry (PIV) complemented with pseudocolor visualization (PCV) technique, which have been shown to provide a good synergy between quantitative and qualitative airflow characteristics and can be adequately employed for the CBL investigation (Licina et al., 2014). Findings of this study contribute to the fundamental knowledge of how the human CBL interacts with the airflow in the occupied spaces of the room.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

Experimental facility

Measurements were performed in an environmental chamber: 11.1 × 8 × 2.6 m (L × W × H) equipped with displacement ventilation. A thermal manikin was positioned in the center of the chamber. Air from a dedicated air-handling unit was supplied via six low-momentum floor standing diffusers and exhausted through six ceiling mounted grills. All supply and exhaust devices were located at far enough distance from the thermal manikin to eliminate interference between the supply airflow and the CBL, as shown in Figure 1 (left). Furthermore, supply diffusers were sealed on the side facing the manikin to prevent any such interference. To minimize radiant heat exchange with the surroundings, one external wall was insulated and heat sources from the lighting fixtures were eliminated. The other three walls and the floor were adiabatic, while the ceiling was suspended below an insulated roof of the building.

image

Figure 1. The environmental chamber (left) and experimental setup for the sitting manikin (right)

Download figure to PowerPoint

Experimental equipment

Airflow generator

A specially designed rectangular box with dimensions 1.8 × 1 × 0.2 m (L × W × H) was used to generate a uniform forced convection flow around the thermal manikin. The top plate of the box was perforated with 18 000 holes of 5 mm diameter. The box contained 66 DC fans that were controlled with a fine-tuning frequency regulator capable of providing a velocity range from 0 to 1 m/s. The amount of heat added to the airstream by the fans was negligible. The uniformity test was performed at 0.7 and 1 m distance above the perforated plate to assure that the velocity remains constant at both distances. Test was performed in accordance with the standard (ASHRAE, 2012) that uses the coefficient of variation as an indicator of the flow uniformity. Results obtained with omnidirectional thermal anemometers (Dantec Dynamics, Copenhagen, Denmark) with ±0.02 m/s and ±0.5 K accuracy showed that at both 0.7 and 1 m distances, the velocity remained constant with a low discrepancy (<10%) among the corresponding 32 horizontal grid points. These findings suggested that the box generated a uniform flow of room air before reaching the CBL of the manikin.

Thermal manikin

The non-breathing thermal manikin with a complex female body shape of 1.68 m height in the standing and 1.23 m height in the sitting posture was used to resemble a realistic human body. The body consisted of 26 independently controlled segments that released 65 W/m2 of heat which approximates the heat loss from the human body in a thermally comfortable state. The recorded temperature of the manikin's surface was in the range of 30.5–34.5°C, similar to the body surface temperature distribution in the environment studied. The tight-fitted clothing of the manikin, consisting of trousers, T-shirt, shoes, socks and underwear, corresponds to a typical attire level of occupants in tropical buildings. Tight clothing ensemble enabled an easier reproduction of the measurement results. The thermal manikin was calibrated prior to the experiments.

Particle image velocimetry

In this study, PIV system was employed to capture and analyze the target flow field in front of the thermal manikin. It consists of dual YAG laser (New Wave Research, Inc., Fremont, CA, USA), double-pulse 190 mJ, and a wavelength of 532 nm, synchronizer, computer and 2 MP CCD camera with 28 mm lens that offers 1600 × 1200 pixel resolution. The image acquisition frequency of the CCD camera was set to 10 Hz which well exceeds typical frequencies of turbulent eddies found in ventilated spaces (Hanzawa et al., 1987). Atomized olive oil particles (mean diameter of 1 μm) generated in a six-jet atomizer (Model 9306; TSI Inc., Shoreview, MN, USA) were used for seeding the flow around the thermal manikin. The degree of coupling between the particles and the fluid was high (Stokes number close to zero), which suggests that the seeding particles behaved like tracers. More details related to the flow seeding, calibration, and the measurement accuracy are discussed in Licina et al. (2014).

Pseudocolor visualization

The PCV technique was employed to visualize the flow field around the thermal manikin. Both the PIV and PCV are image acquisition techniques which are similar to two differences: (i) in the area of the flow field captured by the camera and (ii) the image processing algorithms. In the case of PCV technique, as the individual seeding particles need not be visible to the camera, it enabled a substantially larger field of view. The PCV technique has the ability to assign a spectrum of colors to the particles illuminated by the laser, depending on the pixel intensity obtained by the camera. For the purpose of this study, all PCV images have entirely qualitative character and the color spectrum does not have any quantitative content. More details on PCV technique can be found in the manual (Insight 3G™, 2004).

Experimental design

The CBL of the sitting thermal manikin and its interaction with surrounding airflows was examined under three different scenarios: (i) opposing airflow from above; (ii) transverse airflow from front; and (iii) assisting airflow from below. In the three scenarios, the airflow generator was positioned above, in front, and below the thermal manikin at the distances indicated in Figure 1 (right). For scenarios of opposing and assisting flows, the airflow generator was positioned 12 cm away from the surface (respectively ceiling and floor) to allow air suction from its bottom plate. The results were compared to the case obtained in a quiescent indoor environment, where the CBL so generated was taken as the reference (and henceforth referred as ‘pure CBL’). Interaction between the CBL and assisting flow was performed primarily for the sitting body posture. The manikin was seated in a basic type of the plastic chair shown in Figure 1 (right), which is better described in Licina et al. (2014). In the case of assisting flow, in response to the finding that the chair created an interference that significantly altered the air flow patterns that impacted the breathing zone, additional experiments were conducted for the standing posture. In all scenarios, uniform airflow was supplied through the airflow generator to achieve velocities of 0.175, 0.3, and 0.425 m/s at the surface of the manikin when the manikin was not heated (i.e., adiabatic with the room temperature). Velocities of 0.175 and 0.3 m/s were chosen to represent typical velocities occurring in most occupied indoor spaces (Baldwin and Maynard, 1998), while a velocity of 0.425 m/s is similar to one induced by the mechanical fan and/or personalized ventilation (Melikov et al., 2003).

Minimizing the total heat gain in the chamber facilitated a constant room air temperature of 23°C to be maintained at the minimum ventilation rate (1/h). The low ventilation rate minimized interference between the CBL and ventilation flow. Velocity and temperature measurements were obtained with Dantec thermal anemometers at four locations around the thermal manikin at 1.2 m distance, vertically placed at 0.5, 1, 1.5, and 1.9 m from the floor at each of the locations. Based on 300 recordings obtained during 5 min, the mean velocity magnitude measured was below 0.05 m/s, which indicated that quiescent indoor environment had been achieved (Murakami et al., 2000). Air temperature at the highest measurement point was 0.5°C higher than the air temperature near the floor, which suggested a low degree of vertical thermal stratification.

To achieve a good mixing between the olive oil particles and air, seeding was performed through the bottom side of the airflow generator. Uniformly distributed seeding particles were discharged through the perforated plate of the box by the mechanical fans. Particles were illuminated by the PIV laser that was aligned with the centerline of the manikin. In the case of the transverse flow, the PIV laser was mounted at the top of the airflow generator to avoid interference with the flow discharged from it. To capture instantaneous flow field in the breathing zone delimited with 20 × 15 cm area, the CCD camera was positioned at 0.48 m orthogonal distance from the laser sheet.

Data analysis

After performing an optimal number of images independence test, it was found that variation in the average velocity distribution beyond 540 image pairs becomes insignificant (<3%). Therefore, for each measurement, 540 image pairs were analyzed that is equivalent to 54 s of the flow. Captured images were analyzed using Insight 3G software (version 9.1.0.0; TSI Inc.). A processor type ‘Recursive Nyquist’, which starts with 64 × 64 pixels and ends with 32 × 32 pixels interrogation area, was used to analyze instantaneous images. As a result, 2–3% of spurious vectors per image were detected and rejected. Image post-processing functions or filters were not applied in the data analysis.

Analyzed data are presented as a mean velocity plot along with the velocity vectors where the length of an arrow is proportional to the velocity magnitude and the direction of the arrow corresponds to the flow direction at the point. In addition, the velocity and RMS of fluctuating velocity are presented as a function of the horizontal distance from the middle section of the mouth. The magnitudes of the velocity and RMS represent the resultant vectors from the X and Y components. The relative turbulence intensity was obtained as the ratio of the standard deviation of velocity fluctuations (RMS) and the mean velocity at each measurement point.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

Three sets of experiments were carried out, which describes the interaction between the CBL with mutually opposing, transverse, and assisting flows.

Interaction with opposing flow from above

Figure 2 shows the mean velocity contours in the breathing zone of the CBL in a quiescent environment, as well as its interaction with the opposing flow from above. Figure 3 shows the influence of the opposing flow on the CBL mean velocity profiles and the RMS of fluctuating velocities as a function of the horizontal distance from the mouth. In the case of no surrounding airflow, the thermal manikin created a velocity profile that is characteristic of airflows adjacent to a vertical heated plate. The peak velocity in the mouth region was 0.185 m/s, recorded at a distance of 14 mm, while at 150 mm distance from the mouth, the velocity dropped nearly by half (0.095 m/s). Adding an opposing (downward) flow at 0.175 m/s created an insignificant effect on the airflow characteristics in the breathing zone, as shown in Figures 2 and 3 (top). Further velocity increase created substantial disturbances to the CBL at the height of the head. A downward velocity of 0.30 m/s was not able to completely offset the plume; however, it reduced the peak velocity by 46% (from 0.185 to 0.10 m/s), while at 150 mm distance from the mouth, the velocity dropped down to 0.06 m/s. As the velocity of the CBL close to the surface is naturally higher, stronger disturbances of the CBL are expected to occur at some distance from the surface. As seen in Figure 2, at 120 mm distance in front of the nose, the downward flow collided with the CBL, thus creating the vortex in that region. Opposing flow at 0.425 m/s was able to completely break away the human thermal plume and created a predominantly downward velocity in front of the face. In this case, invading downward flow followed the natural curvature of the skull peaking at 0.26 m/s at 90 mm distance in front of the mouth (Figure 3, top). As the top of the head created a physical barrier for the downward flow, the region below the chin and very close to the mouth preserved the upward air movement.

image

Figure 2. Particle image velocimetry of the convective boundary layer (CBL) and its interaction with opposing flow from above

Download figure to PowerPoint

image

Figure 3. Mean velocity (top) and RMS of fluctuating velocities (bottom) in the mouth region: Impact of the opposing flow from above

Download figure to PowerPoint

Although the results of the RMS of fluctuating velocities are presented (Figure 3, bottom), the results were discussed based on the relative turbulence intensity values. In case of the pure CBL, the fluctuating velocity component remained relatively constant further from the body surface, which indicated that the relative turbulence intensity gradually increased with increasing the distance from the surface, reaching 85% at 150 mm distance from the mouth. Due to the small disturbance in case of the opposing flow at 0.175 m/s, similar relative turbulence intensity profile was observed. The opposing flow at 0.30 m/s created a collision point between two flows at the nose level, thus causing a high degree of turbulence intensity up to 110%. When downward flow of 0.425 m/s was applied, a change in the flow direction occurred within 20–30 mm from the mouth which resulted in a region of higher turbulent intensity with turbulence levels of 110%. Further ahead of the mouth, the relative turbulence intensity declined progressively to 27% at 150 mm from the mouth due to a single predominant airflow direction.

To fully comprehend how the uniform downward airstream impinges upon the thermal plume above the head and the CBL in front of an occupant, PCV technique was employed (Figure 4). The top part of the Figure 4 displays a pure thermal plume produced from the manikin's body heat. As there were no mechanically driven airflows in the surroundings, the plume rose vertically upwards. The thermal plume and mutually opposing flow at 0.175 m/s collided at the height of about 0.4 m above the head (Figure 4). At this height, the rising thermal plume is spread in the horizontal direction, similar to the effect of the jet impinging on an orthogonal plate. A study in which locally supplied airflow opposes the thermal plume showed that the airflow of 4 l/s, which corresponds to the target velocity of 0.24 m/s, creates a collision point at a similar height (Yang et al., 2009). Downward flow at 0.30 m/s further penetrated the rising plume creating a collision point at the height of approximately 0.15 m above the head. Due to a stronger CBL in a narrow space enveloping the human body, thermal plume above the head remained upward. Finally, when the opposing flow approached the manikin at the velocity of 0.425 m/s, the thermal plume did not exist as it was fully overpowered by the invading flow. Four additional videos that describe interaction between the human thermal plume and downward flow at these three velocities are enclosed as Supporting information.

image

Figure 4. Pseudocolor visualization of the convective boundary layer (CBL) at 0.5 s interval: Influence of opposing flow from above

Download figure to PowerPoint

Interaction with transverse flow from front

Figure 5 shows the mean velocity contours of the CBL under its interaction with the transverse flow from the front. Figure 6 shows the change of the mean velocity and standard deviation of velocity fluctuations (RMS) as a function of the distance at the level of the mouth. The results indicate that the pure CBL flow was strongly disturbed by invading transverse flow at a relatively low velocity of 0.175 m/s (see also Figure 2—for comparison with pure CBL). In this case, the maximum velocity of the CBL dropped sharply from 0.185 m/s (for the pure CBL) to 0.08 m/s at 15 mm horizontal distance from the mouth. Further increase in the invading flow velocity to 0.30 and 0.425 m/s completely replaced the upward CBL flow with horizontal airstreams, generating mean velocities of 0.10 and 0.125 m/s at 15 mm distance from the mouth, respectively. When each of the three invading velocities was applied, the mean velocity at 15 mm horizontal distance from the mouth was lower than in the case of the pure CBL (Figure 6, top). The velocity of invading flow steadily decreased as it approached the manikin, which is partially due to its collision with the manikin's surface (Yang et al., 2009) and partially due to the collision with the upward CBL.

image

Figure 5. Particle image velocimetry of the convective boundary layer and its interaction with transverse flow from front

Download figure to PowerPoint

image

Figure 6. Mean velocity (top) and RMS of fluctuating velocities (bottom) in the mouth region: Impact of the transverse flow from front

Download figure to PowerPoint

Exposing the manikin to horizontal airstream from the front at 0.425 m/s induced higher velocity close to the surface of the manikin, compared to those introduced at 0.30 and 0.175 m/s. As the CBL was completely replaced by the invading flow at 0.425 m/s, the flow maintained a predominantly horizontal flow direction. This predominant flow direction resulted in the lowest values of the standard deviation of velocity fluctuations (RMS) and consequently the lowest relative turbulence intensity levels. At 100 mm distance in front of the mouth, the relative turbulence intensity values for the free stream velocities of 0.425, 0.30 and 0.175 m/s were 35%, 45%, and 60%, respectively.

Interaction with assisting flow from below—sitting manikin

Airflow interaction between the CBL and assisting flow from below was investigated when the thermal manikin was standing or seated in the chair. The velocity contours in front of the sitting manikin exposed to a uniform upward airflow at 0.175, 0.30, and 0.425 m/s are presented in Figure 7. Expectedly, assisting flow at 0.175 m/s resulted in increased mean velocity in the breathing zone, relative to the case of the pure CBL (see also Figure 2). Although, assisting flow at 0.175 m/s reduced the peak velocity from 0.185 to 0.165 m/s, the velocity along the horizontal distance from the mouth was higher and remained constant at 0.16 m/s, as shown in Figure 8 (top). The RMS of fluctuating velocities in the same region was reduced and remained relatively steady at 0.065 m/s, which corresponds to a turbulence intensity of 40%. In contrast to expected results, further increase in the upward velocity to 0.30 and 0.425 m/s reduced the velocity in the breathing zone and changed the direction of the flow. The velocity reduction occurred up to 70 mm distance from the mouth compared to the pure CBL scenario, and across the whole breathing zone compared to the case of assisting flow at 0.175 m/s (Figure 8, top). The velocity profile for assisting flow at 0.30 and 0.425 m/s was rather similar, starting from 0.12 m/s at 10 mm in front of the mouth and steadily increasing to 0.14 m/s at 150 mm distance. The upward vector direction was replaced with vectors approaching the face at 45° angle, which will be discussed in the following paragraphs. Unlike the velocity, the level of the relative turbulence intensity in front of the mouth was considerably higher for the case of assisting flow at 0.425 m/s (85%), compared to the turbulence intensity at 0.30 m/s (55%), as shown in Figure 8 (bottom).

image

Figure 7. Particle image velocimetry of the convective boundary layer and its interaction with assisting flow from below—sitting manikin

Download figure to PowerPoint

image

Figure 8. Mean velocity (top) and RMS of fluctuating velocities (bottom) in the mouth region of a sitting manikin: Impact of assisting flow from below

Download figure to PowerPoint

To understand the effect of velocity reduction in the breathing zone due to velocity increase in the assisting flow, additional PIV experiments were performed. The target measurement area was positioned in the chest and abdominal zone of the manikin, as indicated in Figure 1 (right). Due to the space limitation, the results in Figure 9 show only the velocity contours in case of assisting flow at 0.175 and 0.425 m/s; however, scenarios of the pure CBL and assisting flow at 0.30 m/s are also discussed. Thermal manikin placed in a quiescent environment induced an upward air movement of approximately 0.10 m/s in the abdominal region that steadily accelerated toward the chest (not shown). As seen in Figure 9, adding the assisting flow at 0.175 m/s created an area of low velocities at 0.03 m/s in abdominal region. Further in front of the abdomen, at the distance that precisely matched the horizontal edge of the chair (X = 160 mm), there was a strong upward movement that corresponds to the flow induced by the airflow generator. After passing the edge of the chair, this flow did not separate much from the chair and maintained a confluent character with respect to the rising CBL flow, as shown in the chest region. The confluent character between the CBL and mutually assisting flow at 0.175 m/s clearly explains the velocity increase in the breathing zone.

image

Figure 9. Particle image velocimetry of the convective boundary layer and its interaction with assisting flow from below at 0.175 m/s (left) and 0.425 m/s (right)—chest (top) and abdominal (bottom) region in front of the sitting manikin

Download figure to PowerPoint

After its collision with the chair, the assisting flow at 0.425 m/s separated from the edge of the chair, forming a large vortex in front of the chest with a downward airstream in the abdominal region, as shown in Figure 9 (right). From the fluid dynamics view point, this problem is typical for bluff body flow where the passing flow experiences frictional and pressure drag force. In this case, pressure drag was predominant and related to the cross-section of the chair and flow separation. As the upward velocity from the airflow generator increased, Reynolds number increased, and consequently, the pressure gradient on the bottom side of the chair increased in magnitude. Once the adverse pressure gradient became sufficiently strong, it produced separation of the flow. This separation from the edge of the chair increased the size of a wake, which explains the existence of a large vortex in the chest region. This vortex was responsible for change of the airflow direction in the breathing zone resulting in vectors 45° inclined toward the face. A similar effect was observed in case of assisting flow at 0.30 m/s, where the only difference was the size and location of the vortex. The vortex was smaller and shifted closer to the body due to the lower velocity from the airflow generator and respective pressure drag. Four additional videos that describe interaction between the human CBL and upward flow at three velocities are enclosed with this article (Supporting information).

Interaction with assisting flow from below—standing manikin

Findings that velocity increase in the assisting flow decreases the velocity in the breathing zone led to further investigations that involve a standing body posture. Figure 10 shows the mean velocity contours in front of the standing manikin under its interaction with the assisting flow from the airflow generator. Unlike the sitting posture where the chair created a blocking effect to the rising flow from the airflow generator, a standing manikin created no obstruction for the assisting flow. Consequently, the higher velocity of the assisting flow generated higher velocities in the breathing zone of the manikin. In the case of a pure CBL, after peaking at 0.165 m/s, the velocity profile steadily decreased to 0.125 m/s at 130 mm horizontal distance from the mouth (Figure 11, top). Assisting flow at 0.175 m/s created the same velocity pattern in a higher magnitude range, peaking at 0.245 m/s and continually decreasing down to 0.21 m/s. Relative turbulence intensity along the horizontal distance from the mouth substantially dropped when assisting flow at 0.175 m/s was applied, from 55% down to 30% (Figure 11, bottom). Assisting flow at velocities of 0.30 and 0.425 m/s created a low turbulence intensity of about 30% and the velocity profile that gradually increased with the horizontal distance from the mouth, reaching 0.32 and 0.38 m/s at 100 mm distance, respectively.

image

Figure 10. Particle image velocimetry of the convective boundary layer (CBL) and its interaction with assisting flow from below—standing manikin

Download figure to PowerPoint

image

Figure 11. Mean velocity (top) and RMS of fluctuating velocities (bottom) in the mouth region of a standing manikin: Impact of assisting flow from below

Download figure to PowerPoint

A comparison of the velocities in the breathing zone when the thermal manikin was standing and sitting in a quiescent environment (Figures 8 and 11) confirms the previous findings that the sitting posture generates a higher peak velocity (Licina et al., 2014). Moving the manikin from seated to standing posture reduced the peak velocity from 0.185 to 0.165 m/s (by 11%). Nevertheless, at 10 cm horizontal distance from the mouth, standing manikin generated higher velocity of 0.125 m/s compared to 0.105 m/s for the sitting posture, which suggest a milder velocity decay profile in the case of a standing manikin.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

The present study is designed to enhance understanding of the convection flow enveloping the human body and its interaction with mutually opposing, transverse, and assisting airflow. The results are discussed with respect to personal exposure to ambient pollutants, thermal comfort, and optimal ventilation design.

Ventilation with a vertically downward air distribution is recommended as an efficient pollutant removal strategy that reduces the risk of airborne cross-infection (ASHRAE, 2013; CDC, 2003). In this approach, however, a uniform downward airflow generated by the ceiling mounted air terminal devices is disturbed in the regions above the heat sources that intensify the air mixing (Nielsen et al., 2007). Our results show that the downward uniform flow at the velocity of 0.30 m/s penetrates the human thermal plume to the breathing zone causing intensive air mixing resulting in a high degree of turbulence intensity up to 110%. This is well above the turbulence intensity of 27% recorded when the plume was offset by downward flow at a velocity of 0.425 m/s, which highlights the importance of the velocity magnitude of the ventilation flow and airflow interaction in general. Other examples of flow opposing the CBL are downward diffuse ventilation, downward flow induced by the ceiling mounted personalized ventilation or ceiling mounted fan that is commonly found in residential and other non-conditioned spaces. The level of penetration to the breathing zone depends on characteristics and interaction of the downward flow and the thermal plume. Our finding that downward flow of 0.175 m/s does not affect the CBL is different from widely adopted assumption that invading flow of 0.1 m/s starts to disturb the CBL. In terms of optimal ventilation design that is aimed to deliver fresh and clean air to the occupant's inhalation zone without draught discomfort, downward flow approach may have several issues. The downward flow confronts the upward thermal plume, and hence, a higher momentum is necessary to reach the breathing zone, compared to assisting and transverse flow scenarios. Concerning the ability of the CBL to entrain and transport potentially infectious pollutants from the lower room level to the breathing zone, the magnitude of the opposing flow can have different implications on the human exposure. Downward flow at 0.425 m/s peels of the CBL in the breathing zone and protects from pollution located in the middle or lower room level, or in the case that an occupant is a pollution source itself. In this case, however, special attention must be paid to avoid thermal discomfort, as the velocity of 0.425 m/s exceeds the comfortable velocity range recommended by the standard (ISO, 2005). Increased velocity also poses an additional energy penalty. At 0.30 m/s, downward flow reduces the draught risk but collides with the CBL close to the head which may easily increase particle dwell time in the same region and thus increase the exposure to such particles. Opposing flow at 0.175 m/s has a negligible impact on the CBL; thus, the pollution at the floor level near the human body will end up in the breathing zone unobstructed. Our finding that a complete destruction of the thermal plume occurs at a downward velocity of 0.425 m/s differs from the one found by Yang et al. (2009). They reported that the downward personalized airflow supplied from ceiling mounted nozzle destroys the plume at an airflow above 16 l/s, which corresponds to the velocity of 0.8 m/s in the target area. This discrepancy may be due to the difference in the cross-section area of the local invading flow (diameter 0.35–0.5 m), and in our case, a cross-sectional area generated by the airflow generator (1.8 m2). Comparison of these two studies emphasizes that the CBL may interact differently with global airflow patterns in the occupied spaces and locally supplied airflows.

Occupants exposed to traversing flows can be commonly found in practice, for example desktop fan, personalized ventilation, and stratum ventilation. In contrast to opposing flow, transverse flow penetrates the CBL at the breathing zone at a relatively low velocity of 0.175 m/s. This finding can be useful for an optimal design of personalized ventilation system, because transverse flow from front can deliver fresh air to the breathing zone within comfortable velocity limits and with lower energy penalty, compared to the opposing flow scenario. The ability of transverse flow to penetrate to the breathing zone at a lower flow rate can also be important from the protective point of view against ambient pollutants entrained by the CBL or those shed from the clothing and skin. The latter ones apply to products of chemical reactions of bioeffluents from the body. In this case, invading transverse flow from the front easily pushes the rising pollution toward the side of the body and prevents them from reaching the breathing zone. In this case, special attention must be paid to reduce cross-infection risk because the pollutants from the occupant, including the potentially infected exhaled air if he/she is sick, may be discharged toward the other occupants. One way is to combine supply with personalized exhaust that could serve as a source control and mitigate fraction of pollution from the infected occupant that is inhaled by other occupants (Melikov and Dzhartov, 2013; Yang et al., 2013). Occupants exposed to a uniform transverse flow from the front are likely have a higher draught discomfort level compared to opposing flow at equal flow rate, as there is no need to offset the thermal plume, but also improved perceived air quality (Melikov and Kaczmarczyk, 2012).

Occupants exposed to an upward air movement create another commonly encountered airflow interaction in an indoor environment. Typical examples are occupants exposed to an upward flow generated by displacement or under-floor air distribution, as well as upward piston flow from the perforated floor customary met in clean rooms. The results of this study reveal the importance of furniture arrangement that augment the complexity of interaction between the CBL and mutually assisting flow. The blocking effect of the chair creates a vortex in abdominal and chest region and completely alters the air patterns around the sitting manikin at the velocities above 0.175 m/s. In another study, it was found that the presence of the table in front of an occupant can reduce the peak velocity in front of the mouth by 45% (Licina et al., 2014). On the other hand, increasing the airflow of assisting flow supplied vertically upwards from the table to 10 l/s increased the velocity in the breathing zone up to 0.34 m/s, because there were no obstacles encountered (Bolashikov et al., 2011b). These findings clearly imply that the chair, the table, and other furniture play an important role in formation of air patterns around the human body and should be carefully considered in numerical predictions and optimal ventilation design. More research is needed on the topic of different furniture design and arrangement to achieve a desired air distribution around the human body. Airflow patterns in front of the manikin placed in a uniform upward airstream substantially differ between sitting and standing body postures. The upward airflow induced by the airflow generator and the CBL of a standing manikin are mutually confluent and do not cause substantial air mixing. On the other hand, downward air distribution within the comfortable velocity range is disturbed by the manikin's plume which makes it difficult to create a uniform downward air distribution due to high mixing of air and high turbulence intensity levels. Increased turbulence intensity of the flow adjacent to the human body may cause an increase in the heat transfer from the body (Melikov and Zhou, 1996). These findings show that the body posture and its orientation relative to the airflow direction are important for airflow distribution in the microenvironment around the human body. Considering the pollution from the floor that is brought up to the breathing zone via the CBL, adding the clean assisting flow would probably reduce the exposure in both the cases of standing and sitting manikin, due to a dilution effect. The same apply for the other pollutants generated close to the occupant such as bioeffluents from the body or formaldehyde from the table that are expected to be diluted by the upward uniform flow and reduce the exposure. Pollutants located at some distance from the body would be transported to the upper room level where they can be exhausted. The impact of the critical distance at which the pollution would not be transported to the breathing zone needs to be studied in the future.

Although discussed, this study does not show the results of how interaction between the CBL and mutually opposing, transverse, and assisting flow affects personal exposure to surrounding pollutants. These personal exposure results are yet to be reported. It should be noted that the practical implications of this study are valid at a surrounding air temperature of 23°C and may be different at higher or lower temperatures. The effects of other air movements, such as created by other occupants or heat sources, have also not been studied. In addition, room geometry such as increasing the ceiling height modifies the CBL. The current work does not consider a human respiratory airflow which needs to be studied in the future. Future research and design trends are expected to go through the paradigm shift from the coarse total volume description and design to the description and design that considers the airflow distribution in the microclimate around the occupants and its impact on personal exposure. This shows the necessity for a thoughtful understanding of the airflow physics enveloping the human body and its interaction with airflows generated by the air delivery systems.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

The PIV, complemented with PCV technique, is employed to investigate the resultant velocity vectors generated by the complex interaction between the CBL around human body and mutually opposing, transverse, and assisting flows. This airflow interaction depends on the direction and the magnitude of the invading flow and should be carefully considered in order to reduce personal exposure to ambient pollutants and thermal discomfort and achieve optimal ventilation design. Under the conditions studied, the downward flow at the velocity of 0.175 m/s is unable to modify airflow patterns in the breathing zone, which is very different from assisting flow that alters the CBL at the same velocity. Increasing the velocity of downward flow to 0.30 m/s reduces the peak velocity in the breathing zone from 0.185 to 0.10 m/s, while downward flow of 0.425 m/s destroys the thermal plume. In the case of transverse flow supplied from the front, the CBL is strongly disturbed already at the velocity of 0.175 m/s. The CBL of a standing manikin exposed to the uniform upward airflow does not cause extensive mixing of flow patterns, such as in case of opposing flow from above that is disturbed by the occupant's thermal plume. Thus, a higher velocity of the assisting flow increases the velocity in front of the standing manikin. In case of the sitting manikin, however, increasing the upward velocity above 0.175 m/s reduces the peak velocity in the breathing zone and changes the flow direction, attributable to the blocking effect of the chair. These findings imply that furniture plays an important role in formation of airflow patterns and should be carefully considered in numerical predictions and for an optimal ventilation design.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

The authors would like acknowledge Dr. Jovan Pantelic for his comments to the discussion section.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information
  • ASHRAE (2012) Method of Testing General Ventilation Air-Cleaning Devices for Removal Efficiency by Particle Size, ASHRAE Standard 55–2012, Atlanta, Georgia, American Society of Heating, Refrigerating and Air-Conditioning Engineers.
  • ASHRAE (2013) HVAC Design Manual for Hospitals and Clinics, Atlanta, Georgia, American Society of Heating, Refrigerating and Air-Conditioning Engineers.
  • Baldwin, P.E.J. and Maynard, A.D. (1998) A survey of wind speeds in indoor workspaces, Ann. Occup. Hyg., 42, 303313.
  • Bjørn, E. and Nielsen, P. (2002) Dispersal of exhaled air and personal exposure in displacement ventilated rooms, Indoor Air, 12, 147164.
  • Bolashikov, Z., Nagano, H., Melikov, A., Velte, C. and Meyer, K.E. (2011a) Airflow characteristics at the breathing zone of a seated person: passive control over the interaction of the free convection flow and locally applied airflow from front for personalized ventilation application. Proceedings of Roomvent.
  • Bolashikov, Z., Melikov, A., Velte, C. and Meyer, K.E. (2011b) Airflow characteristics at the breathing zone of a seated person: interaction of the free convection flow and an assisting locally supplied flow from below for personalized ventilation application. Proceedings of Roomvent.
  • Bolashikov, Z., Melikov, A., Kierat, W., Popiołek, Z. and Brand, M. (2012) Exposure of health care workers and occupants to coughed airborne pathogens in a double-bed hospital room with overhead mixing ventilation, HVAC&R Res., 18, 602615.
  • CDC (2003) Guidelines for Environmental Infection Control in Health-Care facilities, USA, U.S. Department of Health and Human Services Centers for Disease Control and Prevention.
  • Clark, R.P. and Toy, N. (1975) Natural convection around the human head, J. Physiol., 244, 283293.
  • Craven, B.A. and Settles, G.S. (2006) A computational and experimental investigation of the human thermal plume, J. Fluids Eng.-T. ASME., 128, 12511258.
  • Etheridge, D. and Sandberg, M. (1996) Building Ventilation – Theory and Measurement, Chichester, UK, Wiley.
  • Hanzawa, H., Melikov, A. and Fanger, P.O. (1987) Airflow characteristics in the occupied zone of ventilated spaces, ASHRAE Trans., 93, 524539.
  • Heist, D.K., Eisner, A.D., Mitchell, W. and Wiener, R. (2003) Airflow around a child-size manikin in a low-speed wind environment, Aerosol. Sci. Technol., 37, 303314.
  • Homma, H. and Yakiyama, M. (1988) Examination of free convection around occupant's body caused by its metabolic heat, ASHRAE Trans., 94, 104124.
  • Insight 3G™ (2004) TSI Software Platform – Instruction Manual.
  • ISO (2005) Ergonomics of the Thermal Environment – Analytical Determination and Interpretation of Thermal Comfort Using Calculation of the PMV and PPD Indices and Local Thermal Comfort Criteria, ISO Standard 7730, Geneva, Switzerland, International Organization for Standardization.
  • Johnson, A.E., Fletcher, B. and Saunders, C.J. (1996) Air movement around a worker in a low-speed flow field, Ann. Occup. Hyg., 40, 5764.
  • Lewis, H.E., Foster, A.R., Mullan, B.J., Cox, R.N. and Clark, R.P. (1969) Aerodynamics of the human microenvironment, Lancet, 322, 12731277.
  • Licina, D., Melikov, A., Pantelic, J., Sekhar, C. and Tham, K.W. (2014) Experimental investigation of the human convective boundary layer in a quiescent indoor environment, Build. Environ., 75, 7991.
  • Melikov, A. and Dzhartov, V. (2013) Advanced air distribution for minimizing airborne cross-infection in aircraft cabins, HVAC&R Res., 19, 926933.
  • Melikov, A. and Kaczmarczyk, J. (2012) Air movement and perceived air quality, Build. Environ., 47, 400409.
  • Melikov, A. and Zhou, G. (1996) Air movement at the neck of the human body, Proceedings of Indoor Air, Nagoya, Japan, 1, 209214.
  • Melikov, A., Cermak, R., Kovar, O. and Forejt, L. (2003) Impact of airflow interaction on inhaled air quality and transport of contaminants in rooms with personalized and total volume ventilation, Proceedings of Healthy Buildings, Singapore, 2, 592597.
  • Murakami, S., Kato, S. and Zeng, J. (2000) Combined simulation of airflow, radiation and moisture transport for heat release from a human body, Build. Environ., 35, 489500.
  • Nielsen, P.V., Hyldgaard, C.E., Melikov, A., Andersen, H. and Soennichsen, M. (2007) Personal exposure between people in a room ventilated by textile terminals: with and without personalized ventilation, HVAC&R Res., 13, 635643.
  • Pantelic, J. and Tham, K.W. (2013) Adequacy of air change rate as the sole indicator of an air distribution system's effectiveness to mitigate airborne infectious disease transmission caused by a cough release in the room with overhead mixing ventilation: a case study, HVAC&R Res., 19, 947961.
  • Popiołek, Z., Bolashikov, Z., Kostadinov, K., Kierat, W. and Melikov, A. (2012) Exposure of health care workers and occupants to coughed air in a hospital room with displacement air distribution: impact of ventilation rate and distance from coughing patient, Proceedings of Healthy Buildings, Brisbane, Australia.
  • Spengler, J.D. and Sexton, K. (1983) Indoor air pollution: a public health perspective, Science, 221, 917.
  • Yang, B., Melikov, A. and Sekhar, C. (2009) Performance evaluation of ceiling mounted personalized ventilation system, ASHRAE Trans., 115, 395406.
  • Yang, J., Sekhar, C., Cheong, D.K.W. and Raphael, B. (2013) Computational fluid dynamics study and evaluation of different personalized exhaust devices, HVAC&R Res., 19, 934946.
  • Zukowska, D., Melikov, A. and Popiolek, Z. (2012) Impact of personal factors and furniture arrangement on the thermal plume above a sitting occupant, Build. Environ., 49, 104116.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
ina12120-sup-0001-VideoS1.avivideo/avi2176KVideo S1.Human CBL in a calm indoor environment.
ina12120-sup-0002-VideoS2.avivideo/avi2553KVideo S2. Interaction between the CBL and assisting flow at low velocity.
ina12120-sup-0003-VideoS3.avivideo/avi1988KVideo S3. Interaction between the CBL and assisting flow at medium velocity.
ina12120-sup-0004-VideoS4.avivideo/avi2482KVideo S4. Interaction between the CBL and assisting flow at high velocity.
ina12120-sup-0005-VideoS5.avivideo/avi1478KVideo S5. Human thermal plume in a calm indoor environment.
ina12120-sup-0006-VideoS6.avivideo/avi2169KVideo S6. Interaction between the CBL and opposing flow at low velocity.
ina12120-sup-0007-VideoS7.avivideo/avi2168KVideo S7. Interaction between the CBL and opposing flow at medium velocity.
ina12120-sup-0008-VideoS8.avivideo/avi2042KVideo S8. Interaction between the CBL and opposing flow at high velocity.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.