Indoor and Built Environment

http://ibe.sagepub.com CFD Study of the Thermal Environment around a Human Body: A Review
N. P. Gao and J. L. Niu Indoor and Built Environment 2005; 14; 5 DOI: 10.1177/1420326X05050132 The online version of this article can be found at: http://ibe.sagepub.com/cgi/content/abstract/14/1/5

Published by:
http://www.sagepublications.com

On behalf of:

International Society of the Built Environment

Additional services and information for Indoor and Built Environment can be found at: Email Alerts: http://ibe.sagepub.com/cgi/alerts Subscriptions: http://ibe.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

Review Paper

Environment

Indoor Built Environ 2005 14;1:516

Accepted: June 27, 2004

CFD Study of the Thermal Environment around a Human Body: A Review

N. P. Gao J. L. Niu
Department of Building Services Engineering, The Hong Kong Polytechnic University

Key Words Computational thermal manikin Turbulence models Thermal plume Heat transfer Contaminant distribution

Introduction An important goal in indoor air quality (IAQ) and thermal comfort research is an enhanced understanding of the relationship between the human body and the surrounding environment. Human bodies, in this respect, play a central role in indoor environments. To facilitate studies, thermal manikins were introduced at least a half century ago. Thermal manikins are heated dummies which were originally developed to measure the thermal insulation of clothing. In 1977, Mihira et al. [1] developed a thermal manikin not only for clothing insulation measurement but also for the evaluation of thermal environments. Later more complex thermal manikins were used to study thermal environments of human bodies in the indoor environment [24]. Although experimental researches using thermal manikins provide valuable information about airow velocities, temperature and pollutant concentration, some other detailed information, such as the air ow eld in the close vicinity of a person and the ratio between the amount of radiative heat transfer and convective heat transfer between human body and surroundings, cannot be obtained by experiments. During the last few years CFD technology has progressed and made it possible to analyse the micro-climate around a human body. A

Abstract
Computational uid dynamics (CFD) simulations of thermal manikins can provide detailed information which it is impossible to obtain from experiments. In this review several issues from CFD studies of thermal manikins are discussed, such as geometric complexity of computational thermal manikins (CTMs), selection of turbulence models, grid generation and boundary conditions. Calculation of CTMs with simple geometry requires less computational resource but a more anthropomorphic geometry can provide more accurate results. In the simulations using complex geometric representations of the human body the computational domain is discretised by body-tted co-ordinates or unstructured grids. The total number of cells can reach the quantitative magnitude of 1 million. The simulation results about airow eld, radiative and convective heat transfer and contaminant distribution from different researchers are compared. Further developments in the use of these simulations are proposed.

2005 Sage Publications DOI: 10.1177/1420326X05050132 Accessible online at: www.sagepublications.com

Dr N.P. Gao Department of Building Services Engineering The Hong Kong Polytechnic University, Hung Hom, Kowloon Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 Hong Not Kong, China use or unauthorized distribution. 2005 International Society of the Built Environment. All rights reserved. for commercial Tel. 852 2766 4698. E-mail 02902469r@polyu.edu.hk

growing number of computational thermal manikins (CTMs) have been proposed with the purpose of determining indices that are not able, or at least very difcult, to be obtained from experiments. A lot of work on CFD calculation of CTMs has been done in Japan (University of Tokyo) and Denmark (Aalborg University and the Technical University of Denmark). The intentions behind CFD simulations of modelled human bodies is various. Hayashi et al. [56], Murakami et al. [7], Bjrn and Nielsen [8], Xing et al. [9] simulated the contaminant distribution around modelled human bodies and its impact on inhaled air quality. Srensen [10] tabulated view factors between different body segments and between the outer surfaces and the body segments by numerical method. Murakami et al. [1113], Srensen and Voigt [14] investigated the airow velocity eld and convective and radiative heat loss from a human body through CFD. However, very few researches have been carried out to evaluate thermal comfort using CTMs. At the present time such work has reached the stage where the CFD simulation rst generates the environment data to input into a thermal comfort model and then the thermal comfort model calculates the whole body thermal sensation and local sensation. Such thermal comfort simulations are regarded as preliminary because they do not reach the degree of combined simulation which bridges analysis of the indoor air environment with thermoregulation of the human body. The development of CTMs is step by step. Twodimensional simplied geometry was used in earlier studies to simulate the air ow eld around a human body without consideration of temperature difference. Then a temperature eld and radiation was calculated through simplied 3-dimensional geometry. More complicated simulations take account of inhalation and exhalation processes and an unsteady state may be modelled [15]. This paper summarises some issues in CFD simulations of thermal manikins. Geometric complexity, turbulence models, boundary conditions, grid generation and simulation results from various studies were reviewed. Finally we give some suggestions and discuss the future outlook for these simulations.

Geometry of CTMs CTMs differ in size, posture and their level of geometric complexity. Although there are no standards for the size of CTMs the skin surface area and body volume are

based on those of real human bodies. The height of most CTMs is about 1.65 m and the surface area ranges from 1.594 m2 [14] to 1.688 m2 [11]. According to different situations that are selected, 3 postures of CTMs are normally applied: sleeping, seated and standing. The level of geometric complexity of CTMs depends on the purpose of the simulations involving them. Dunnett [16] proposed an elliptical cylinder to simulate a standing person and Niwa et al. [17] applied a heated box representing a seated person. Davidson and Nielsen [18] simulated the air ow in facial regions and the nasal cavity by means of 2-dimensional simulations with a detailed model of a human head. Heinsohn [19] used a cuboid (parallelepiped) to model the exposure of a standing person. Tjelaat and Knott [20] proposed a cuboid suspended from the oor to calculate thermal comfort indices. Other researchers have used a 3-dimensional cuboid in rectangular geometry to model either a seated person [21] or a standing person [2227]. The geometries discussed above are based on computational aided designs. The latest and most complicated CTM geometry is obtained using the laser scanning technique. All the ne geometric characteristics and extremities of the human body can be represented by using this technique (Figure 1) [28]. The hair and clothes are usually not included since correct description of them is a difcult task. Srensen and Voigt [14] extracted the geometrical details of a thermal manikin in a seated position using a laser scanning technique. In their simulation the surface of a human body was represented by approximately 125,000, 250,000, or 500,000 triangular elements for coarse, medium and ne surface descriptions (Figure 2). This model is especially useful in some cases, such as in calculation of segment-to-segment view factors and inhaled air quality under the inuence of facial geometry characteristics. As in the case, say, of modelling an air diffuser, the key problem is the large difference in the scale between the detail of body extremities and the room size ranging in magnitude from centimetres to metres. Calculation of CTMs with simple geometry requires fewer computational resources and less time for modelling and grid generation. However, using a detailed representation of a human body is closer to practical situations and so able to provide better local results. From the view point of computational expense, the geometry of CTMs should be as simple as possible. From the view point of simulation accuracy, it should be as real as possible. Therefore, it is necessary to know how complex a shape for the CTMs is required to obtain the desired CFD results. Topp et al.

Indoor Built Environ 2005;14:516

Gao/Niu

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

Fig. 1. Using laser scanning technique to obtain a CTM with the real geometry of the human body [28] (left: thermal manikin set-up to obtain the geometry of a CTM; right: thermal manikin converted to a CTM in the laser scanning workshop).

[2930] investigated numerically the difference between different shapes of CTMs with respect to both local and global air distributions as well as convective transfer. Two seated CTMs with different levels of geometric complexity were placed in a unidirectional ow eld. The simulated human body geometries are illustrated in Figure 3. The one on the left is a simple rectangular model of a seated person and the others (middle and right) are accurate representations of a human body. Topp et al. found that the geometry of the CTM had an inuence only in the region very close to where it was located. It had no inuence on the airow at some distance from it. An accurate geometric model of CTMs is necessary for the numerical studies of thermal plume around the human body, thermal comfort and inhaled air quality since the geometry of the human body plays a role in the measurement of draft and contaminant transportation. In their studies of heat transfer from a human body to its surroundings in which the convective heat

ow rate was xed at 38.0 W, the convective heat transfer coefcient of the complicated CTM model (7.4 Wm2C1) was higher than that of the simple CTM model (5.0 Wm2C1). It was calculated from the following equation. qc hc ts tin (1)

where hc convective heat transfer coefcient, Wm2C1; qc heat ux, Wm2; ts human body mean skin temperature, C; tin room air supply temperature, C; The higher value for the complicated CTM model was the result of the higher local air velocity due to the geometric conguration of the gures. From this study we can see that when numerical research focuses on global airow patterns in a ventilated room a less detailed geometry is enough, however an

CFD Study of the Environment Around a Human Body

Indoor Built Environ 2005;14:516

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

expected at the head region and the inhaled air quality is very sensitive to the interaction of different airows, such as the warm rising airow caused by natural convection, the personalised airow, the respiration airow and the room ambient airow [28].

Turbulence Models Although there are many turbulence models it is an unfortunate fact that no single turbulence model is universally accepted as being superior to the others for all classes of problems. The choice of a turbulence model is a compromise which depends on considerations such as; the physics encompassed in the ow, the established practice for a specic class of problems, the level of accuracy required, the available computational resources, and the amount of time available for the simulation. In the simulations of airow around human body, the most frequently used turbulence models are the RNG k- model and the low Reynolds number k- model. Murakami et al. [11] pointed out that since no wall functions had been applicable for the standard k- turbulence model to simulate the airow around thermal manikins, the low Reynolds number k- model with locally-ne grids at the human body surface is the only way to calculate the convective heat transfer. However in the simulations using the low Reynolds number k- model and an accurate geometric model of the human body, it is difcult to achieve convergence due to the extremely ne grids near the human body [31]. In the simulation by Srensen and Voigt [14] using the low Reynolds number k- model, in order to tackle the boundary layer around the CTM, 20 layers of extruded triangular prisms were created at the surface of the manikin with an initial height of 0.2 mm and a growth rate of 1.13 between layers. Gao and Niu [28] predicted the micro-environment around human body with personalised ventilation using a standard k- model. In order to describe the complicated geometry of the human body unstructured grids with small scale (about 48 mm) was applied. However accurate simulation of convective heat transfer by the standard k- model requires that at the wall boundary the value of y should be in a reasonable range. It was found that the mean convective heat transfer coefcient changed from 4.95 to 4.38 Wm2C1 when they coarsened the grids around the human body and increased the value of y into the range from 15 to 29 [28]. Therefore ne grids lead to smaller y, and further they over-predict the convective heat transfer.

Fig. 2. Close-up of the CTMs face (from the website of the International Center for Indoor Environment and Energy, Denmark Technical University, www.ie.dtu.dk/manikin/).

accurate geometry of the human body is necessary when we are interested in the local ow pattern around the human body. Two examples are given here. One case is in the vehicle HVAC industry. In the simulation of airow and to evaluate the thermal comfort condition in a car, a CTM with detailed body surface characteristics is needed since the length scale of a car compartment is close to the boundary layer thickness of air around the human body. The other is in the numerical simulations of the inhaled air quality provided by personalised ventilation systems in which the air supply jet is located very close to the face. All facial extremities should be addressed by the CTM since a large eld gradient is

Fig. 3. Geometry of the CTMs from [29].

Indoor Built Environ 2005;14:516

Gao/Niu

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

Due to the temperature difference between the human body and its surroundings, the airow pattern close to the body can be a mixed convection ow or a natural convection ow which depends on whether the human body is placed in stagnant air or not. The importance of buoyancy forces in a mixed convection ow can be measured by the ratio of the Grashof and Reynolds numbers: Gr gTL Re2 v2 (2)

When this number approaches or exceeds unity, this means strong buoyancy contributions to the ow. Conversely, if it is very small, buoyancy forces may be ignored in simulations. In pure natural convection, the strength of the buoyancy-induced ow is measured by the Rayleigh number: gTL Ra
3

Since there may be many typical ow elements, such as jet ow, buoyancy driven ow, laminar ow and potential ow, in the simulation of air ow in a ventilated room with a human body, no turbulence model which is now available can deal perfectly with this complex ow pattern. The k- model is the best choice for now among various turbulence models available in view of its simulation accuracy and computational expense. The low Reynolds number k- model performs better in the prediction of heat loss from human body while the standard k- model and RNG k- model are sufcient if the emphasis is on airow eld.

Grid Generation Since the human body has an extremely complex geometry, grid generation for its simulation has some special requirements. The correctness and accuracy of a simulation result depends very much on grid quality. Ordinarily, for complex congurations the grid generation takes up most of the time and human labour in the whole process of CFD calculation. The accuracy is inuenced by a number of grid factors, such as grid size, grid topology, grid shape, consistency of the grids with the geometry, etc. The space can be discretised into Cartesian co-ordinates, cylinder co-ordinates, body-tted co-ordinates and unstructured grids. In the simulation of airow around complicated geometries, body-tted co-ordinates and unstructured grids are often used to represent the complex boundaries. Body-tted co-ordinates are suitable for internal or external ows with smoothly varying irregular boundaries. An unstructured grids method is more exible in grid distribution than the body-tted coordinates. Because structured grids have higher quality than unstructured grids, the ventilated room is usually divided into 2 parts: a cuboid enclosing the human body and the remaining room space. The cuboid is broken with unstructured grids due to the complex geometry while the remaining room space is discretised with structured grids (Figure 4). The number of cells is very large especially in a simulation using an accurate representation of the human body. The 3-dimensional computational domain is discretised into 125,568 cells in Murakami et al. [11], 292,421 in Topp et al. [29] and more than 1 million in Srensen and Voigt [14]. This is because to discretise the region close to the complicated boundary of human body requires very ne cells.

(3)

where is the thermal expansion coefcient: 1 T and is the thermal diffusivity: k cp (5)

(4)

Rayleigh numbers less than 108 indicate a buoyancyinduced laminar ow, with transition to turbulence occurring over the range of 108 Ra 1010. If we assume the human body to be a cylinder with the height of 1.65 m and radius of 0.15 m, the Reynolds number will exceed 2500 based on the equivalent diameter of 0.3 m as the characteristic length of a human body if the air velocity is more than 0.13 ms1. When a human body is sitting in stagnant air, the Ra number will reach about 4.1 109 at the head level, assuming the temperature difference is 9C. Therefore in most cases the airow around the human body is turbulent or is in the transition zone from laminar ow to turbulent ow. However, strictly speaking, the k- model is only applicable in the fully developed turbulent ow. It is difcult to reproduce the transition from laminar to turbulent with the k- model. Furthermore the k- model is not accurate enough for simulating the air ow eld around a bluff body. Murakami et al. [13] used a contrived method on a k- model to simulate this transition ow by adding a slight turbulence generation term to the k-equation and -equation throughout the whole computational domain.

CFD Study of the Environment Around a Human Body

Indoor Built Environ 2005;14:516

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

Qs Qt Es Qt Qc Qr

(7) (8)

In most simulations without consideration of moisture transfer from the skin surface, we can obtain a sensible heat transfer Qt from the skin surface to the environment by excluding the heat loss caused by respiration and evaporation according to the equations from Fanger [32] and ASHRAE 1989 [33]: Qres 1.7 105M(5867 Pa) 0.0014M(34 ta) (9) Es 3.05 103(5733 6.99M Pa) 0.42(M 58.15)

(10)

Fig. 4. Grid conguration of the computational domain from [14].

If the CTM is obtained from a laser scanning technique, the surface is composed of a huge number of elements. For example, the number of triangular elements of the body surface reached 500,000 in [14]. A large number of surface elements has its advantages for the view factors calculation of body segments. However this number, which has the quantitative magnitude of 100,000, means it is almost impossible to use in calculations of the view factor using commercial CFD codes. Thus the elements on a body surface should be coarsened to calculate the view factors.

where Pa vapour pressure, pa; ta ambient air temperature, C. In the situation where there is a xed surface temperature, the mean skin temperature ts under thermal neutrality may be estimated by the following equation according to Fanger [32]: ts 35.77 0.028Qm (11)

Boundary Conditions Two methods are usually applied to describe the boundary condition of the human body: xed surface temperature and xed surface heat ux. The body boundary conditions are based on human heat exchange with the environment. Under steady-state conditions, the metabolic heat production M is equal to the heat loss from a human body Qm which is divided into heat loss from skin surface Qs and that caused by respiration Qres. Qs is composed of sensible heat loss Qt and latent heat loss Es. Sensible heat loss Qt consists of convective Qc and radiative Qr heat loss. M Qm Qs Qres (6)

In numerical studies, the body temperature is usually set in the range from 31C [14] to 33.7C [34]. And the convective heat transfer rate is usually set in the range from 20 Wm2 [5] to 25 Wm2 [26]. In other experimental studies, a third kind of boundary condition for the human body was used. Tanabe et al. [35] assumed the vapour pressure Pa to be 1.5 kpa, which was equivalent to typical indoor air conditions at 24C and 50% relative humidity. Then the following equation could be derived from Equations (6) to (10): Qm 1.96Qt 21.56 (12)

By replacing the term Qm in Equation (11) with Equation (12), the following equation was obtained: ts 36.4 0.054Qt (13)

The skin surface of a thermal manikin was controlled to satisfy Equation (13), which implied that the skin temperature and heat loss at different segments of the human body were coupled and neither were set at constant values. However this kind of skin surface boundary condition has still not been realised in CFD studies. The

10

Indoor Built Environ 2005;14:516

Gao/Niu

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

inherent fault of control Equation (13) when it is applied to each body part is that this equation is deduced from the heat balance of the whole body. The thermal regulation of the human body and the physiological difference of individual segments are not taken into account for studying the local thermal comfort under a non-uniform thermal environment. To do this some multi-node heatbalance models would be required. If the respiration process is taken into account in the simulations, the ow rate and respiration cycle under light physical work should be studied. Different researchers [5,15,36,37] have produced different curves to approximate the change of the inhaled and exhaled air ow rate with time (Figures 5a, 5b, 5c, 5d). The real respiration characteristics of humans may vary a little according to age, race, gender, etc. The exhaled air condition is not completely constant. The mean exhaled air temperature depends mainly on inhaled air condition [38]. At an ambient temperature of 20C, the mean temperature of exhaled air through the nose is approximately 32C. The molecular composition of exhaled air is different from that of atmospheric air. Therefore the density of exhaled air is not the same as the density of the atmosphere, which will impact on the buoyancy function.

Bjrn [39] has given the density of exhaled air at different exhalation temperatures. In some cases, such as numerical prediction of thermal comfort, the body should be divided into many segments with each segment having its own boundary condition. For example, we may expect that the skin temperature of the feet and the breasts is not the same. Usually the temperature of the feet is much lower than that of the breasts. As mentioned above the control Equation (13) has shortcomings if applied to body parts. In the numerical simulations iteration methods could be envisaged due to the rapid development of multi-node thermal regulation models, such as Berkeley thermal comfort model [40]. Initial surface temperatures Ti are assumed as the boundary conditions to calculate the airow eld, room temperature eld and humidity ratio, etc. Then using the resultant air conditions near the human body we can predict the skin surface temperatures Tc at each segment with the help of thermal regulation models and replace Ti with Tc. The nal skin temperatures can be obtained when the difference between Ti and Tc is small enough. This method has been applied in practise by Currle [41].

0.25 0.20 0.15 0.10 0.05 0 0.05 0.10 0.15 0.20 0.25

Respiration rate (L/s)

Respiration rate (L/s)

0.3

10

15

20

25

30

0.5

1.5

0.5 0.5 0.5 1.5 0.5 0.5

Time (s)

0.3

Time (s)

one cycle (6 s)

Fig. 5a. The sinusoidal curve of human respiration according to Zuo [36].

Fig. 5c. Modelling of respiration by Murakami [15].

Respiration rate (L/s)

Actual pattern (0.6 L for each inhaling and exhaling) Air flow velocity at mouth entrance during breathing 0.6 L
2.5 0.5 2.5 0.5 2.5 0.5 2.5

Velocity (m/s)

0.2 0.1 0 0.1 0.2

3.871

Time (s)

Approximation (steady inhaling at 14.4 L/min) Time (s)

Fig. 5b. Assumed model for steady inhalation by Hayashi et al. [5].

Fig. 5d. Breathing waveform by Hyun and Kleinstreuer [37].

CFD Study of the Environment Around a Human Body

Indoor Built Environ 2005;14:516

11

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

Radiative Heat Transfer A goal for thermal comfort and heat stress researchers is to construct a comprehensive model of human thermoregulation and predict thermal comfort by numerical methods. The completion of such a model requires radiative and convective heat transfer coefcient of individual body segments to be known. Especially in the recently developed personalised air supply systems, local heat transfer at specic body regions varies greatly because the human body is set in a highly asymmetric and nonisothermal environment. Experimental research on local radiative and convective heat transfer has been done by de Dear et al. [42]. Recently, some numerical research has been carried out (Table 1) [10,14]. In order to calculate the radiative heat transfer from a human body to a room wall and between body segments, a CTM with accurate and detailed representation of the human body is necessary. Furthermore, radiation is affected by the posture of the human body. Innite combinations of human body postures will lead to different values of the radiative heat transfer rate. Usually the seated and standing postures are simulated. In the calculation of radiation heat transfer, the emissivity of the human body has been set about 0.95. The value of 4.7 Wm2C1 has been widely accepted as a reasonable whole body estimate of the radiative heat transfer coefcient for general purposes [43] and is used as a reference value to evaluate calculation results. The radiation from a human body to room walls is surface to surface radiation. The calculation of view factors is computationally very expensive when there are a large number of radiating surfaces. In models

Table 2. Radiative heat transfer coefcient for different segments of a standing human body Body segments Experiment by Ichihara et al. [44] (Wm2C) 7.3 4.8 4.2 3.7 4.0 4.0 3.9 4.3 3.8 3.6 4.3 Experiment by de Dear et al. [42] (Wm2C) 3.9 5.3 4.3 4.1 4.9 5.2 4.2 4.1 4.5 4.4 4.5

Foot Leg Thigh Hand Arm Shoulder Pelvis Head Chest Back Whole body

with detailed representation of the human body the skin surface is constructed from thousands or even tens of thousands of patches. The calculation of a view factor needs considerable computational time as well as high storage capacity. Experimental data for the radiative heat transfer coefcient for individual segments of standing human body have been measured (Table 2) [42,44]. However no corresponding numerical data are reported. Murakami et al. [45] simulated the radiative heat transfer using a simplied geometry of the human body in a standing posture. In their research no radiative heat transfer coefcient for individual segments was reported: their mean value for the radiative heat transfer rate was 38.3 Wm2.

Convective Heat Transfer

Table 1. Radiative heat transfer coefcient for different segments of a seated human body Body segments Calculation by Srensen et al. [14] (Wm2C) 5.36 5.12 4.61 4.12 4.30 4.63 5.01 5.22 4.73 5.07 4.83 Experiment by de Dear et al. [42] (Wm2C) 4.2 5.4 4.6 3.9 5.2 4.8 4.8 3.9 3.4 4.6 4.5

Foot Leg Thigh Hand Arm Shoulder Pelvis Head Chest Back Whole body

The convective heat transfer coefcient reported by different researchers [11,14,29,31,42,46] varies largely because the convective heat transfer from a human body to its surroundings is affected by air velocity, air ow direction and turbulence intensity (Table 3). The widely accepted expression for the natural convective heat transfer coefcient is the following equation [32]: hc 2.38(ts ta)0.25 Wm2C1 (14)

In still air the human body is enclosed by a rising thermal plume due to the temperature difference between the body surface and ambient air. The thickness of this boundary layer around the human body increases with height and is similar to the ow around a vertical

12

Indoor Built Environ 2005;14:516

Gao/Niu

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

Table 3. Comparison of the mean convective heat transfer coefcients hc of a human body in still air Researchers Murakami et al. [11] Murakami et al. [11] Srensen and Voigt [14] Topp et al. [29] Voigt [31] Brohus [46] de Dear et al. [42] de Dear et al. [42] Method CFD CFD CFD CFD CFD Experiment Experiment Experiment Posture standing standing seated seated seated standing standing seated Ambient air speed

(ms21)
0.12 0.12 stagnant 0.05 0.025 0.05 0.1 0.1

(Wm22C)
3.9 4.3 3.13 7.4 6.1 3.86 3.4 3.3

hc

Table 4. Convective heat transfer coefcients for different segments of a seated human body in a stagnant air ow eld Body segments Calculation by Srensen and Voigt [14] (Wm2C) 4.66 3.04 3.18 4.50 3.82 2.71 2.80 3.62 2.38 2.23 3.13 Experiment by de Dear et al. [42] (Wm2C) 4.2 4.0 3.7 4.5 3.8 3.4 2.8 3.7 3.0 2.6 3.3 Experiment by Silva and Coelho [48] (Wm2C) 6.2 5.7 4.4 4.4 6.0 4.6 3.4 0.6 2.7 2.1 3.9 Experiment by Yang et al. [49] (Wm2C) 5.6 6.2 3.9 5.9 6.3 5.9 3.1 6.2 3.8 2.4 4.4

Foot Leg Thigh Hand Arm Shoulder Pelvis Head Chest Back Whole body

heated at plane. Its thickness is about 8 cm at foot level and 15 cm at head height [13]. The maximum air velocity above the head given by different researchers differs greatly. A value of 0.23 ms1 has been given for a standing human body [13,34] and more than 0.5 ms1 for a seated human body [14]. The value of 0.23 ms1 is consistent with that from earlier experiments [47]. A value greater than 0.5 ms1 seems unreasonable since the temperature difference between the human body and ambient air used in this work was similar to that for other simulations and experiments. This rising warm thermal plume around the human body has a great impact on convective heat transfer and inhaled air quality. If the thermal plume is broken due to a high velocity of airow impinging the human body the heat transfer coefcient increases greatly. Murakami et al. [45] simulated a standing human body with metabolic heat production of 100.4 Wm2. In their simulation 29.1 Wm2 was released to the surroundings by convection, 38.3 Wm2, by radiation, 24.3 Wm2 by evaporation and 8.7 Wm2 by respiration. That is, 43%

of sensible heat was lost by convection and 53% due to radiation. In the simulation by Srensen and Voigt [14], 40% of sensible heat was lost by convection and 60% by radiation. The results from Murakami et al. [45] and Srensen and Voigt [14] are in agreement. In still air the convective heat transfer coefcients of different body segments varies greatly [48,49] (Table 4). The feet and hands have large convective heat transfer coefcients since they are exposed to a cold rising airow where the thermal plume is not fully developed. The convective heat transfer coefcient of the back is small because it is enclosed by a warm thermal plume.

Contaminant Distribution Contaminant distribution around the human body has an important impact on inhaled air quality. Since the human body is enclosed by a warm rising thermal plume, air quality at the breathing zone is affected by this rising

CFD Study of the Environment Around a Human Body

Indoor Built Environ 2005;14:516

13

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

airow. The essential question is whether inhalation and exhalation are taking place within this rising airow. If we assume that: (1) (2) (3) (4) a person has a respiration rate of 0.6 L per inhalation, each inhalation lasts about 2.5 s, air is inhaled from a hemisphere, the air velocity at a distance of 8 cm from the nose is about 0.006 ms1.

It is important to evaluate the effect of entrainment of air from the lower zone to the breathing zone since inhaled air comes from the rising thermal plume. Brohus and Nielsen [51] proposed the effective entrainment ratio e of lower air to the inhaled air: Cref Cb e Cref Cf (15)

Then the effect of the inhalation on the surrounding air will be very small. Since the thickness of the thermal plume at the height level of the nose can reach 15 cm the inhaled air will consist mostly of the air from the rising thermal plume (Figure 6). However from the results of an experimental study on a living person and a thermal manikin [50], the exhalation air jet can always penetrate through the thermal plume around the human body.

Fig. 6. Path lines showing origin of inhaled air in a stagnant ambient environment.

where Cref contaminant concentration at the breathing height apart from the body; Cb contaminant concentration of inhaled air; Cf contaminant concentration at the oor level. The rising thermal plume may have a positive or negative inuence on inhaled air quality which depends on where the contaminant source is so long as the envelope of the warm rising airow is not broken by the ambient room air movement or other invading ows. For example, if the contaminant is emitted from the oor, the rising airow will draw up the contaminant to the breathing zone. Contrarily, if the contaminant source is above the head in a displacement ventilated room, the rising airow will entrain the fresh air to the breathing zone. In the simulation of contaminant distribution around the human body, attention should be paid to the geometric model of the CTM and the posture of the body. If detailed information is required in the breathing zone such as in the evaluation of the performance of personalised ventilation systems [28], the presence of mouth, nose and ears, etc. is required. Models with these facial extremities will require an increased number of grids and be computationally expensive. Brohus and Nielsen [26] found great difference in the contaminant concentration at the breathing level of 2 simplied models of the human body with different geometry. If the human body faces a horizontal airow, the gap between the legs will allow air to ow across the lower body and affect the entrainment of the rising airow which therefore affects different contaminant concentrations at the breathing level. Simplied CTMs obtained by reducing legs to a solid extension of the torso may not predict the transportation of pollutants accurately. The inhaled air quality is also affected by the posture of the body. Hayashi et al. [5] considered the human body in different postures: standing, sitting and sleeping and analysed the inhaled air quality numerically. In the standing position the volume from which air will be inhaled extends to the oor. The occupant inhales air from the volume that was close to the oor at the front of the body. The scenario is similar for a seated occupant. When an occupant is lying on the

14

Indoor Built Environ 2005;14:516

Gao/Niu

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

oor, air is drawn horizontally from both sides to the head.

Summary This paper discusses several questions on CFD simulations of airow around a human body and heat transfer from the body to the environment. Since a gurative representation of a human body is geometrically complex and it is heated the CFD study of the micro-environments around the body has some special characteristics: The geometry of a CTM has an impact on local airow around the body but almost no inuence on global ow. Simplied geometric models of a human body have their advantages through simplifying grid generation, reducing computational cost and the need for memory storage for view factor calculations, etc. However, models that accurately represent the human body can yield more detailed information on local areas around the body. k- turbulence models are usually used in calculations with CTMs. When a human body is located in still air, the airow around it is buoyancy-driven ow which transits from laminar ow to turbulent ow. The standard k- turbulence model has an inherent fault in modelling transitional ow and the low Reynolds k- turbulence model with a buoyancy term always has problems in convergence. Body-tted co-ordinates or unstructured grids should be used to discretise the computational domain.

The results of a number of simulations are reviewed in this paper. Due to the variable nature of the parameters used for simulations the results show differences in airow velocity, radiative heat transfer coefcient, convective heat transfer coefcient and contaminant distribution. The techniques are still developing and we can expect further advances in CTM simulations. These may include: At the present time it is difcult, even impossible, to do a combined simulation which connects the heat and mass transfer from the human body to the environment with human thermo-regulation. This specialised kind of simulation would have use in the prediction of local thermal comfort. Latent and sensible heat transfer from body segments needs to be taken into account. Coupling of mass and heat transfer from the inner-body to the outer-body at different body segments is the key to solving this problem. Simulations of transient inhalation and exhalation processes with accurate geometric representation of the human body would be useful in the study of personal ventilation. No numerical study has so far been reported in this eld. At the present time the respiration cycle is reduced to a steady inhalation process by many researchers. Including exhalation makes the calculations more complex since it affects the inhaled air quality because some exhaled air will penetrate the thermal plume and the rest will mix with the rising warm air around the body and be re-inhaled.

References
1 Mihira K, Toda H, Arai, H: Study on thermal manikin. Japanese J Human Factor 1977; 13(2):4753. 2 Fanger PO, Banhidi L, Olesen BW, Langkilde G: Comfort limits for asymmetric thermal radiation. ASHRAE Trans 1980;86(2): 141156. 3 Fanger PO, Ipsen BM, Langkilde G, Olesen BW, Christensen NK, Tanabe S: Comfort limits for asymmetric thermal radiation. Energy Build 1986;8(3):225236. 4 Tanabe S, Kimura K, Inoue U: Proposal of evaluation method with thermal manikin. Annual Meeting of Architecture Institute of Japan, 1989;1:875876. 5 Hayashi T, Ishizu Y, Kato S, Murakami S: CFD analysis on characteristics of contaminated indoor air ventilation and its application in the evaluation of the effects of contaminant inhalation by a human occupant. Build Environ 2002;37:219230. 6 Hayashi T, Kato S, Murakami S, Yang JH: CFD analysis on rising stream around a human body and its effect on inhalation air quality. ASHRAE Trans 2002;108(2): 11731178. Murakami S, Kato S, Zeng J: Numerical simulation of contaminant distribution around a modeled human body: CFD study on computational thermal manikin part II. ASHARE Trans 1998;104:226233. Bjrn E, Nielsen, PV: CFD simulation of contaminant transport between two breathing persons. ROOMVENT 98 1998;1:133140. Xing H, Hatton A, Awbi HB: A study of the air quality in the breathing zone in a room with displacement ventilation. Build Environ 2001;36:809820. Srensen DN: Radiation between segments of the seated human body. ROOMVENT 2002 2002;1:317320. Murakami S, Kato S, Zeng J: Development of a computational thermal manikin-CFD analysis of thermal environment around human body. Tsinghua-HVAC 95 1995;2:349354. Murakami S, Kato S, Zeng J: CFD analysis of thermal environment around human body. INDOOR AIR 96 1996;2:479484. Murakami S, Kato S, Zeng J: Flow and temperature elds around the human body with various room air distribution, CFD study on computational thermal manikin part I. ASHRAE Trans 1997;103:315. Srensen DN, Voigt LK: Modeling ow and heat transfer around a seated human body by computational dynamics. Build Environ 2003; 38:753762. Murakami S: CFD study on the micro-climate around the human body with inhalation and exhalation. ROOMVENT 2002 2002;1:2335. Dunnett SJ: A numerical study of the factors affecting worker exposure to contaminant. J Aerosol Sci 1994;25(Suppl.1):481482. Niwa K, Murakami S, Kato S, Kondo Y, Kitamura N: Numerical analysis of ow and temperature eld with local air conditioning by supply jets from the seats in hall. ROOMVENT 96 1996;2:307314. Davidson L, Nielsen PV: Calculation of the

13

14

15

16

10

17

11

12

18

CFD Study of the Environment Around a Human Body

Indoor Built Environ 2005;14:516

15

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.

19

20

21

22

23

24

25

26

27

28

29

two-dimensional airow in facial regions and nasal cavity using an unstructured nite volume solver. ISSN 1395-7953 R9539, Department of Building Technology and Structure Engineering, Aalborg University, Denmark, 1995. Heinsohn RJ: Industrial ventilation: engineering principle. ISBN 0-471-63703-3, New York, John Wiley & Sons, 1991. Tjelaat PO, Knott R: A simulation model for thermal comfort of a person in a large enclosure. INDOOR AIR 96 1996;2:10151020. Iwamoto S: A study on numerical prediction methods of thermal environment around occupants. INDOOR AIR 96 1996;1:299303. Bluyssen PM, Lemaire T: The distribution of perceived air quality in an ofce space computer simulations and sensory evaluations. ROOMVENT 92 1992;3:195211. Gan G, Awbi HB, Croome DJ: CFD simulation of the indoor environment for ventilation design. Presented at the ASME 1993 Winter Meeting Paper No. 93-WA/HT-49, New Orleans, USA, 1993. Gan G: Numerical method for full assessment of indoor thermal comfort. Indoor Air 1994;4:154168. Brohus H, Nielsen PV: Personal exposure to contaminant sources in uniform velocity eld. Healthy Buildings 95 1995;3:15551560. Brohus H, Nielsen PV: CFD models of persons evaluated by full-scale wind channel experiments. ROOMVENT 96 1996;2: 137144. Awbi HB: A CFD study of the air quality at the breathing zone. INDOOR AIR 96 1996;2:10091014. Gao NP, Niu JL: CFD study on microenvironment around human body and personalized ventilation. Build Environ 2004;39: 795805. Topp C, Nielsen PV, Srensen DN: Application of computer simulated persons in indoor environmental modeling. ASHRAE Trans 2002;108(2):10841089.

30 Topp C: Inuence of geometry of a computer simulated person on contaminant distribution and personal exposure. ROOMVENT 2002 2002;1:265268. 31 Voigt LK: NavierStokes simulations of airow in rooms and around a human body. PhD thesis, Technical University of Denmark, 2001. 32 Fanger PO: Thermal comfort: analysis and application in environmental engineering. Copenhagen, Danish Technical Press, 1970. 33 ASHRAE: Thermal comfort. ASHRAE handbook of fundamentals, Atalanta, American Society of Heating, Refrigerating and AirConditioning Engineers, Inc., 1989. 34 Murakami S, Zeng J, Hayashi T: CFD analysis of wind environment around a human body. J Wind Eng Ind Aerodyn 1999;83:393408. 35 Tanabe S, Arens EA, Bauman FS, Zhang H, Madsen TL: Evaluating thermal environments by a thermal manikin with controlled skin surface temperature. ASHRAE Trans 1994; 103:3949. 36 Zuo HG: Experimental study of personalized air system for the reduction of pollutant exposure. Master thesis, Hong Kong Polytechnic University, 2003. 37 Hyun S, Kleinstreuer C: Numerical simulation of mixed convection heat and mass transfer in a human inhalation test chamber. Int J Heat Mass Transf 2001;44:22472260. 38 Hppe P: Temperature of expired air under varying climatic conditions. Intl J Biometeorology 1981;25(2):127132. 39 Bjrn E: Simulation of human respiration with breathing thermal manikin. Proc of Third International Meeting on Thermal Manikin Testing, Stockholm, Sweden, 1999;1:7881. 40 Huizenga C, Zhang H, Arens E: A model of human physiology and comfort for assessing complex thermal environments. Build Environ 2001;36:691699. 41 Currle, J: Numerical simulation of the ow in a passenger compartment and evaluation of the thermal comfort of the occupants. SAE Technical Paper Serials 970529, 1997.

42 de Dear RJ, Arens E, Hui Z, Oguro M: Convective and radiative heat transfer coefcient for individual human body segments. Intl J Biometeorol 1997;40(3):141156. 43 ASHRAE: Physiological principles and thermal comfort. ASHRAE handbook of fundamentals. Atlanta, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. Atalanta, 1993. 44 Ichihara M, Saitou M, Tanabe S, Nishimura M: Measurement of convective heat transfer coefcient and radiative heat transfer coefcient of standing human body by using thermal manikin. Proc Annual Meeting of the Architectural Institute of Japan, 1995;1:379380 (in Japanese). 45 Murakami S, Kato S, Zeng J: Combined simulation of airow, radiation and moisture transport for heat release from a human body. Build Environ 2000;35:489500. 46 Brohus H: Personal exposure to contaminant sources in ventilated rooms. PhD thesis, Aalborg University, Denmark, 1997. 47 Homma H, Yakiyama M: Examination of free convection around occupants body caused by its metabolic heat. ASHRAE Trans 1988;94: 104124. 48 Silva MCG, Coelho JA: Convection coefcients for the human body parts determined with a thermal mannequin. ROOMVENT 2002 2002;1:277280. 49 Yang JH, Kato S, Hayashi T, Murakami S: Measurement of local convective heat transfer coefcients of the human body in outdoor and indoor environments. ROOMVENT 2002 2002;1:281284. 50 Hyldgaard CE: Human as a source of heat and air pollution. ROOMVENT 94 1994;1:413433. 51 Brohus H, Nielsen PV: Contaminant distribution around persons in rooms ventilated by displacement ventilation. ROOMVENT 94 1994;1:294312.

16

Indoor Built Environ 2005;14:516

Gao/Niu

Downloaded from http://ibe.sagepub.com at PENNSYLVANIA STATE UNIV on February 6, 2008 2005 International Society of the Built Environment. All rights reserved. Not for commercial use or unauthorized distribution.