Examination of thermal adaptive effect of postural and positional adjustment of a seated human body exposed to spot airflow.

INTRODUCTION

The most widespread standards for the environmental evaluation of thermal comfort, such as the PMV and SET* indices, which were developed based on the human thermal models of Fanger and Gagge, respectively, by focusing on the heat balance on the human body, consider the human body as a static heat source (Fanger 1973; ISO 1994; Gagge et al. 1971; ASHRAE 1992). However, when people are located in an uncomfortable thermal environment, they can adapt to the surrounding environment to restore comfort. Since a person constantly and actively adjusts the heat exchange with his or her surroundings, he or she provides an active existence to the environment. Therefore, Richard de Dear proposed a new thermal comfort model termed the "adaptive model," which was based on statistical analysis of the results of field experiments performed in over 20,000 office buildings (Dear and Brager 1998; Brager and de Dear 1998). Results of their analysis clarified that people working in a building provided with natural ventilation feel comfort in environments with a wider temperature range than the results calculated using the PMV index. It is possible to reduce the environmental load by decreasing the energy consumption for air-conditioning, which can be assessed using the adaptive model to give looser environmental control than that due to the heat balance models. On the other hand, research on the development of a task-ambient air-conditioning system with a personal airconditioner (PAC) has increased in recent years to aid energy conservation (Yang et al. 2004; Cermak et al. 2006). When a worker is in a task region air-conditioned by a spot PAC, his/ her thermal adaptive behaviors, such as the adjustments of posture and position, etc., which are frequently done in daily life, may contribute greatly to thermal comfort because of the prominent nonuniform velocity and temperature distributions formed by the cooled spot airflow. Therefore, to ensure that energy conservation and thermal comfort are maintained simultaneously, it is necessary to adapt the current adaptive model to incorporate the influence of a person's thermal adaptive behaviors on thermal comfort.

In this study, we examined the thermal effects of postural and positional adjustments of a seated human body exposed to a cooled high-speed spot airflow by simulating personal thermal characteristics in various postures and positions. First, the velocity distribution characteristics were examined in the working space of a seated office worker, which was covered by spot airflow. Then, experiments using an experimental thermal manikin were conducted to measure the local skin temperature and sensible heat transfer rate when changing the leaning posture of the manikin. Also, the coupled simulation of convection, radiation, moisture transport, and Fanger's neutral model was adopted to simulate heat exchange between a seated human body and the surrounding environment, by changing the body's orientation and position relative to the spot airflow. The examination was based on the results of the experiment and simulation.

EXAMINATION OF VELOCITY DISTRIBUTION CHARACTERISTICS IN TASK REGION COVERED BY SPOT AIRFLOW

Outline of Measurement

The experiment was conducted in a climate chamber as shown in Figure 1a. A desk equipped with the spot PAC was placed at the chamber's center, with a styrene board (1.2 m high) to form the surroundings. The ambient air-conditioning system used in the room supplied fresh air of 28[degrees]C from the whole floor surface at a speed of 0.05 m/s; air was exhausted through the whole ceiling surface. The PAC supplied a spot airflow of 28[degrees]C at a speed of 2.5 m/s. The velocities at the positions shown in Figure 1b were measured using an ultrasonic anemometer (KAJIO). During the experiment, the velocities at the positions in section a-a' were measured first, to confirm if the velocities were distributed symmetrically in the flow field. The velocities at the positions enclosed by the dotted line in section b-b' and c-c' were measured subsequently.

[FIGURE 1 OMITTED]

As shown in Figure 1c, the spot PAC is intended for use on a desk. It has two supply openings in the frontal panel, which have an effective diameter of 7 cm (actual size of openings is 8 x 5 cm) and can have grilles to allow adjustment of the airflow's direction. In the experiment, the spot airflow supplied by a spot cooler (Daikin) was spouted out after its airflow rate and temperature were successively adjusted by a branch chamber and a proportional/integral/differential (PID) controller (ASWAN).

Experimental Results

As shown in Figure 2, the velocities were distributed symmetrically left and right in the horizontal sections at each height, with the maximum value at the center of the Y dimension. At the positions within a distance of 15 cm from the center of the Y dimension, the velocities were obviously larger than the ambient velocity of 0.05 m/s. Moreover, for the positions with the same distance to the center of the Y dimension, the velocities became weaker at positions farther away from the desk at X dimension. At the positions over 20 cm away from the center of the Y dimension, the velocities were very small, as the spot airflow exerted little influence. Therefore, the velocities were predominantly influenced by the airflow supplied from the floor. In addition, at the positions at a height of 25 cm and within a distance of 5 cm from the center of the Y dimension, the velocities were larger than at the corresponding position at a height of 15 cm.

[FIGURE 2 OMITTED]

Discussion

According to the results described above, the spot airflow decreased with a steep velocity gradient from its center to the periphery in the Y and Z dimensions, while weakening relatively slowly in the X dimension. This affords the possibility for an office worker to adjust the velocity of the airflow impacting his or her body surface by changing posture or position when the task region is covered by spot airflow. As the change of the impacting velocities will affect heat exchange due to convection on the body surface, it is possible to improve thermal sensation and comfort by postural and positional adjustment. Hence, postural and positional adjustment may be effective thermal adaptive behaviors for a seated human body exposed to the spot airflow.

EXPERIMENTAL ANALYSIS OF HEAT EXCHANGE CHARACTERISTICS OF A THERMAL MANIKIN BY ADJUSTMENT OF ITS LEANING POSTURE

Outline of Experiments

This experiment was conducted in the same climate chamber as mentioned above, using an identical ambient airconditioning system as used in the previous experiments (Figure 3a). However, the upward airflow from the floor was set at 26[degrees]C with humidity of about 60% RH. The experimental thermal manikin was seated in a chair, and the desk, fixed with a PAC, was placed at the chamber's center. The styrene board used in the above experiment was removed here for easier measurement of the velocity distribution around the face using a particle image velocimetry (PIV) system.

[FIGURE 3 OMITTED]

A PAC, as shown in Figure 3b, was used in this experiment. It is composed of a refrigerant circuit, an input device for temperature setting, and a control device to adjust airflow rate and direction. As this PAC is designed to be portable, it has no ducts connecting the outdoor environment but has four openings as shown in Figure 3b. During the measurement, this PAC was used to supply a cooled spot airflow of 17[degrees]C at an airflow rate of 50 [m.sup.3]/h.

The thermal manikin was controlled using the "comfort mode" based on Fanger's thermal neutral equation (Tanabe et al. 1994); local skin temperature and sensible heat transfer rate were measured. As shown in Figure 3c, the manikin was dressed as an office worker in summer with a total clothing thermal insulation of about 0.63 clo.

As mentioned above, the velocity distribution around the manikin's face was measured using a PIV system (Raffel et al. 1998; Zhu et al. 2004). The PIV system's settings used in the experiment are shown in Tables 1 and 2. In this system, the pattern correlation method was adopted to calculate wind velocity vectors; erroneous vectors were removed by velocity-range validation, peak validation, and moving average validation. The final results shown here were obtained by statistical analysis from 500 measurement results of the velocity vectors.

Table 1. Components of PIV System

Processor  PIV 2100 processor of Dantec

CCD        Flow Sense M2 10 bit camera (1600x1600) with
Camera     Nikon Micro lens of 60 mm (35 mm, F = 5.6)
Laser      Plused Nd: YAG laser of New Wave 50 mJ/pulse, 15Hz)

Table 2. Parameters of PIV System

Size of Field  160 x 160 mm

Time between   1 s
Recordings

Time between   For measurement on flow field
Laser Pulses   around PAC's opening: 1 s For
               measurement on flow field
               around face: 6 ms

Interrogation  32 x 32 pixel
Area

Overlap        25% x 25%

Size of        1722 (41 x 42)
Vectors

Experimental Cases

As shown in Table 3, four cases were conducted in total. Except for the upright sitting posture, three postures with the body forward leaning or back leaning were examined. In Case E1, the minimum distance of 15 cm between the manikin and the desk's side was at its abdomen. Compared with Case E1, the manikin is closer to the PAC when it is leaning forward in E2 and farther away from the PAC when leaning backward in Cases E3 and E4.

Table 3. Experimental Cases

Case No              E1           E2          E3          E4

Leaning posture    Upright  15[degrees]  15[degrees]  30[degrees]
                               forward     backward

Distance between    0.67m        0.46m         0.88m        1.12m
the face and the
PAC opening for
cooled air

Experimental Results

Velocity Distribution around PAC's Supply Opening for Cooled Airflow. As shown in Figure 4a, the cooled air was horizontally spouted out toward the manikin in a spot type. The velocities were about 2 m/s, maximized at 2.5 m/s near the opening. However, the spot airflow became weaker due to diffusion. Furthermore, the spot airflow was falling by gravity, as its temperature was 9[degrees]C lower than the ambient temperature.

[FIGURE 4 OMITTED]

Velocity Distribution around Manikin's Face. The velocities near the face in the central section across the manikin in each case are shown in Figures 4b-4e. When the cooled spot airflow reached the body surface after traveling about 0.5 m, its maximum velocity was decreased to about 1/10 of that at the opening. However, as the velocity distribution was over 0.1 m/s near the face, the spot airflow showed a greater influence on the flow field in the measurement region, regardless of the manikin's leaning posture. Moreover, compared with Case E1, in which the manikin was seated upright, the velocity distribution near the face tended to increase somewhat when the manikin was leaned forward or backward.

Distribution of Skin Temperature. As shown in Figure 5a, as the spot airflow collided with different body surfaces of the manikin when its leaning posture was changed, the local temperatures also changed. In each case, at body parts such as the head, hands, and forearms, which were exposed to the spot airflow, the skin temperatures were lower compared with other body parts. In Case E1, as the chest was directly impacted by the spot airflow, its temperature was lower compared with other cases. In Case E2, since the manikin was made to lean forward, the back was closer to the PAC and the body surfaces collided with the spot airflow more than in other cases. Moreover, the forward-leaning posture made the spot airflow go downward after collision and caused temperature decrease at the thighs and legs. Therefore, the overall mean skin temperature was lowest in Case E2. In Case E3, the results were similar to those in Case E1, although the manikin was leaned backward at an angle of 15[degrees]. When the manikin was leaned backward at an angle of 30[degrees] in Case E4, the skin temperature increased at the back, which was farther away from the PAC, and decreased at the head and hands, which were more exposed to the spot airflow.

[FIGURE 5 OMITTED]

Distribution of Sensible Heat Transfer Rate. As the manikin was controlled by Fanger's thermal neutral equation, its overall mean and local sensible heat transfer rates were inversely proportional to the corresponding surface temperatures, as shown in Figure 5b. Therefore, the sensible heat transfer rates changed with the inclination opposite to the skin temperatures when the manikin was made to lean forward or backward.

NUMERICAL ANALYSIS OF HEAT EXCHANGE CHARACTERISTICS OF A HUMAN BODY WITH ADJUSTMENT OF ITS ORIENTATION AND POSITION RELATIVE TO SPOT AIRFLOW

Cases Analyzed

Five cases were examined, as shown in Table 4 and Figure 6. According to Figure 7a, Case S1 was examined in conditions similar to Case E1, with a horizontal distance of 61 cm between the PAC and the abdomen of a seated human body model. In Cases S2 and S3, the human body was placed in the same position but was turned 45[degrees] and 90[degrees], respectively. In Cases S4 and S5, the human body was moved 20 cm farther away from the position in Case S1 along the vertical (Y) and parallel (X) directions of the table, respectively. Therefore, comparison of Cases S1, S2, and S3 will help to examine the thermal adaptive effects of the adjustment of the body's orientation relative to the spot airflow; comparison of Cases S1, S4, and S5 will help to examine the thermal adaptive effects of the body's positional adjustment.

Table 4. Cases Analyzed

Case No.                 S1          S2          S3

Angle between the    0[degrees]  45[degrees]  90[degrees]
human body and
spot airflow

Position of the      P1          P1           P1
human body
(see Figure 8)

Case No.                 S4          S5

Angle between the    0[degrees]  0[degrees]
human body and
spot airflow

Position of the      P2          P3
human body
(see Figure 8)

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

Simulation Method

A low-Reynolds-number type k-[epsilon] turbulence model (Lien et al. 1996) was used for calculation of the flow field together with the SIMPLE algorithm and the UD difference scheme (Table 5). As show in Figure 7, a seated human body model, which was similar to the experimental manikin with a total area of about 1.46 [m.sup.2], was utilized in the simulation. The model had 7338 surface meshes. The first.five cell layers placed over the human body surfaces are prism-shaped fluid cells, while the remainder of the flow field is filled with tetrahedral meshes (Omori et al. 2004). In this research, the nondimensionalized normal distance (i.e., the wall coordinate) measured from the body surface to the center of the first fluid cell is represented by [y.sup.+], which is smaller than 4 over almost the entire body surface in each case. Incidentally, there were about 310,000, 340,000, 320,000, 310,000, 320,000 spatial grids for Cases S1, S2, S3, S4,and S5, respectively. Moreover, the PAC was modeled to be similar to the real machine used for the experiment described above. The boundary conditions for each opening of the PAC were measured in Case E1. The detailed boundary conditions for calculation of the flow field are described in Table 6.

Table 5. CFD Method

Turbulence Model   low-Reynolds-number type k-[eta] turbulence model
Algorithm          SIMPLE
Difference Scheme  First Order Up Wind
Grid System        Surface meshes of human body model: 7338
                   Spatial cells: about 320,000

Table 6. Boundary Conditions

Floor     Velocity: 0.05 m/s; temperature:
          26[degrees]C; RH: 60%. Turbulence
          intensity: 33%; turbulence scale:
          0.005 m

Ceiling   Velocity, temperature: free-slip

PAC       Size:6 x 6 cm; airflow rate: 0.017
Supply    [m.sup.3]/s; temperature:
Opening   44.5[degrees]C RH: 60%; turbulence
for       intensity: 10%; turbulence scale:
Exhaust   0.06 m
Heat

PAC       Size: 5 x 33 cm; airflow rate: 0.016
Supply    [m.sup.3]/s; temperature:
Opening   17[degrees]C RH: 60%; turbulence
for       intensity: 10%; turbulence scale:
Cooled    0.01 m
Air

PAC       Size: 9.6 cm; airflow rate: 0.023
Suction   [m.sup.3]/s; temperature:
Opening   26[degrees]C
for
Cooled
Air

Surface   Skin temperature and sweat rate were
of the    set at values as calculated by
Human     Fanger's model
Body

Other     Humidity insulation, temperatures
Wall      are fixed on the values due to
Surfaces  radiation simulation

Thermal radiation was calculated using Gebhart's absorption factor method (Gebhart 1959); configuration factors over the complex geometry were accurately calculated using a Monte-Carlo method (Howell and Perlmutter 1964) incorporating a symmetrization procedure (Omori et al. 1998). The emissivity was uniformly set at 0.95 for all the wall surfaces in the calculation. The floor temperature was fixed at 26[degrees]C, the temperature for the outlet boundary was equal to that of the outflow air, and the remaining wall surface temperatures were obtained by computational fluid dynamics (CFD).

Fanger's thermal neutral model was used for the calculation of heat transfer inside the human body with a metabolic heat production of 1.1 met (Nishi et al. 1997). Here, Fanger's model was applied to simulate the temperature and sweat rate at each body surface mesh. Furthermore, to account for the influence of hair and clothes, the proper effective clothing thermal resistance was applied for each body part, as shown in Table 7.

Table 7. Effective Clothing Thermal Resistance for Body
Parts Covered by Clothes

Body Part        Thermal Resistance

Head         0.089 [m.sup.2]*[degrees]C/W
Chest        0.183 [m.sup.2]*[degrees]C/W
Back         0.196 [m.sup.2]*[degrees]C/W
Waist        0.214 [m.sup.2]*[degrees]C/W
Upper arms   0.120 [m.sup.2]*[degrees]C/W
Thighs       0.113 [m.sup.2]*[degrees]C/W
Legs         0.067 [m.sup.2]*[degrees]C/W
Feet         0.105 [m.sup.2]*[degrees]C/W

The Coupled Simulation Process

The coupled simulation was composed of four steps as follows: (1) Airflow, temperature, and moisture fields were calculated by CFD to determine the temperature, convective heat transfer rate, and the moisture condition for each surface mesh. (2) The radiant heat transfer rate of each surface mesh was calculated by the radiation calculation using the surface temperatures as determined by CFD. (3) The convective and radiant heat transfer rates, surface temperature, and the moisture condition of each human body surface mesh were collected and transmitted to Fanger's model. (4) The surface temperature and sweating were calculated using Fanger's model and transmitted to the CFD simulation as the boundary conditions for each body surface mesh at the next step. These four steps were repeated until the CFD simulation was completed convergently.

Simulation Results

Flow Fields around the Human Body. As shown in Figure 8a, in each case, except for the high-speed spot airflow present between the human body and the PAC, a rising airflow of about 0.05 m/s formed uniformly over the indoor flow field. In addition, a weak descending airflow over the head was observed. The cooled spot airflow fell by gravity in its travel toward the human body, and was divided into an upward and a downward flow over the body surface, after colliding with the body. In Case S1, the cooled spot airflow impacted the body at the abdomen. In Case S2, the frontal body surface below the neck was covered by airflow exceeding 0.3 m/s. In Case S3, a descending airflow with a maximum velocity of 0.5 m/s at the chest formed in front of the torso after the spot airflow had collided with the body at its side. In Case S4, the spot airflow became weaker and impacted the lower body surface compared with Case S1. In Case S5, the weakened spot airflow showed less influence on the flow field around the torso and was relatively up-down uniform at a velocity of about 0.3 m/s.

[FIGURE 8 OMITTED]

Temperature Field around Human Body. As shown in Figure 8b, except for the field cooled by the spot airflow, almost the whole room was about 26[degrees]C in each case. A distribution exceeding 26.5[degrees]C was clearly observed around the face in Case S1. The high temperature range became narrower when the body's orientation relative to the spot airflow was changed in Cases S2 and S3 and was greater in the upper space when the body was moved farther away from the PAC in Cases S4 and S5. Moreover, the body surface covered by airflow lower than 24.5[degrees]C involved the majority of the frontal torso below the neck in Cases S1 and S2 but was limited to the chest and abdomen in Case S3. In Case S4, almost the whole frontal surface of the torso was covered by the cooled airflow lower than 22.5[degrees]C. In Case S5, it was almost uniformly distributed around the frontal body surface of the torso at a temperature of 23[degrees]C.

Skin Temperature Distribution of Human Body. As shown in Figure 8c, the overall mean skin temperature was similar in each case at a value of about 34[degrees]C. At the face, neck, forearms, and hands, which were exposed to the air, the skin temperatures were lower compared with the body parts in clothes. In addition, among all the clothed body parts, the lowest temperature distribution appeared at the upper thighs, which had less clothing and were impacted by the branch airflow produced by collision of the spot airflow and the body surface. Compared with Case S1, in Case S2 the skin temperature was lower at the left upper arm, which is close to the PAC, and higher at the right upper arm, which is far away from the PAC. Moreover, a left-right temperature gradient appeared at the face and neck, and the temperature decreased on the side close to the PAC. In Case S3, the temperature gradient also appeared at the torso, where the temperature was much lower on the side close to the PAC and higher on the opposite side. Unsurprisingly, the temperature also increased at the crotch. In Case S4, the temperature increased at the jaw, neck, chest, upper arms, and legs, as the spot airflow impacted the human body at lower velocities. The temperature increased more at the surfaces above the chest and was over 35[degrees]C at the shoulder. In Case S5, as the human body was moved farther away from the spot airflow in the direction parallel to the table, a left-right temperature distribution also appeared at the head, neck, chest, abdomen, and upper limbs, due to the temperature increase at the body surface out of the spot airflow. However, the temperature increased uniformly at the lower limbs.

Sensible Heat Transfer Rate Distribution of Human Body. As shown in Figure 8d, the overall mean value of the sensible (convection + radiation) heat transfer rate was similar in each case with a value of about 45 W/[m.sup.2]. The value exceeded 60 W/[m.sup.2] at the body surfaces exposed to the air, which was greater than that of the body surfaces in clothes. In Case S1, the sensible heat transfer rate showed an almost symmetrical distribution. The value was 35-55 W/[m.sup.2] in the front of the torso and relatively greater at the chest. Compared with Case S1, in Case S2, together with the change of the body's orientation relative to the spot airflow, the sensible heat transfer rate became greater on the side close to the PAC and smaller on the opposite side. In addition, the rate decreased at the legs. The left-right nonuniform distribution of the sensible heat transfer rate was more obvious in Case S3. The increase was greatest at the right upper arm where the maximum value was over 65 W/[m.sup.2]. Conversely, it decreased to less than 30 W/[m.sup.2] on the left side of the body. Moreover, it was less than 25 W/[m.sup.2] at the surfaces of the crotch. In Case S4, the sensible heat transfer rate decreased at the chest, upper arms, neck, and upper thighs, etc., which collided with the spot airflow or its branch airflow owing to collisions with the body surface. However, there was little change at the abdomen, which was directly impacted by the spot airflow, with its value ranging at 40-45 W/[m.sup.2]. In Case S5, the sensible heat transfer rate showed a left-right asymmetrical distribution at the surfaces of the face, neck, and frontal torso, as the heat release decreased at surfaces farther away from the spot airflow. At the legs and feet, the sensible heat transfer rate was distributed almost symmetrically, as in Case S1.

Convective Heat Transfer Rate Distribution of Human Body. As shown Figure 8e, the overall mean value of the convective heat transfer rate was approximately 28 W/[m.sup.2] for each case. At the body parts covered by the spot airflow or its branches produced by collision with the body surface, the convective heat transfer rate exceeded 35 W/[m.sup.2], greater than that at other body surfaces. Moreover, it maintained the left-right symmetrical distribution at the legs and feet in all cases, regardless of the postural and positional adjustment of the human body. In Case S1, the distribution was almost leftright symmetrical over the whole body surface. At the frontal torso, frontal upper arms, and upper thighs, although the body surfaces were inside the clothes, the convective heat transfer rate exceeded 40 W/[m.sup.2] under the influence of the cooled spot airflow. With the exception of the surfaces of the above body parts and the exposed body parts, the value was less than 25 W/[m.sup.2]. In Cases S2 and S3, owing to the change of the body's orientation relative to the spot airflow, the convective heat transfer rate increased at the face, neck, frontal torso, arms, hands, and upper thighs, which were moved closer to the spot airflow, and decreased at the body parts that were moved farther away from the spot airflow. This resulted in a left-right asymmetrical distribution. In Case S3, the values were greater than 80 W/[m.sup.2] at the left upper arm and left side of the neck and less than 20 W/[m.sup.2] at the right side of the torso. In Case S4, the convective heat transfer rate decreased across the whole frontal torso surface as the human body was moved farther away from the PAC in the Y direction. Furthermore, it was less then 15 W/[m.sup.2] at the shoulder. In Case S5, a left-right gradient formed at the face, neck, and torso owing to the nonuniform velocity distribution around their surfaces. As a result, the convective heat transfer rate was greater at the surfaces that collided with the stronger wind.

Radiant Heat Transfer Rate Distribution of Human Body. As shown in Figure 8f, the overall mean value of the radiant heat transfer rate was approximately 17 W/[m.sup.2] in each case. The radiant heat transfer rate was relatively low at the frontal torso, frontal upper arms, and upper thighs, which were covered by the spot airflow or its branches owing to collisions with the body surface, with the value less than 15 W/[m.sup.2]. However, it exceeded 30 W/[m.sup.2] at the face, forearms, and hands. The change of posture and position exerted little influence on the radiant heat transfer rate at the legs and feet. In Case S1, a left-right symmetrical distribution appeared almost over all of the body surfaces. However, in Cases S2 and S3, it was smaller on the side closer to the spot airflow and larger on the opposite side at the neck, frontal torso, upper arms, and upper thighs, owing to the change of the body's orientation relative to the spot airflow. This resulted in a left-right asymmetrical distribution. In Case S3, the radiant heat transfer rate was less than -10 W/[m.sup.2] at the left upper arm. In Case S4, the radiant heat transfer rate increased at the neck, shoulder, chest, and upper arms; a distribution of less than 0 W/[m.sup.2] was limited to the abdomen and part of the upper thighs. In Case S5, a leftright gradient was generated at the neck and frontal torso, etc. Moreover, the distribution of less than 0 W/[m.sup.2] was centralized on the body's left side.

DISCUSSION

Validation of Numerical Method's Prediction Precision

The results of experimental case E1 and simulation case S1, which were conducted in similar conditions, are compared in terms of skin temperature and sensible heat transfer rate in Figure 9. According to the comparison of the skin temperature, except for the head, hands, and forearms, etc., which were exposed to the air, the experimental and analytical results correlated well, with a temperature difference within 0.2[degrees]C. However, as the human body model has longer upper limbs compared with the experimental thermal manikin, more heat was released from the hands and forearms by convection as they are located closer to the PAC's supply opening for the cooled air in the simulation. This resulted in a greater difference at the hands, with a maximum value of 1.05[degrees]C at the right hand. Similar results were obtained by comparison in terms of the sensible heat transfer rate. The maximum difference of 20.26 W/[m.sup.2] also appeared at the right hand.

[FIGURE 9 OMITTED]

Examination of Thermal Adaptive Effect of Postural and Positional Adjustment of a Human Body Covered by Cooled Spot Airflow

Both in the experiments and the simulations described above, the differences of the local thermal characteristics were disregarded since Fanger's human thermal model was used, which considers the whole human body as a simple node and determines whole-body heat balance rather than local balance. Moreover, as the body parts that should be most influenced by the spot airflow were assumed to be covered in clothes, changes in thermal performance owing to changes in the surrounding flow field caused by postural and positional adjustment were not fully characterized. Therefore, the interpretation of the results obtained from the experiment and simulation was evaluated in light of these two limitations. Moreover, for easier comparisons, we assigned cases E1 and S1 as control cases, and calculated other cases' deflection to the basic cases in terms of skin temperature and sensible heat transfer rate, as shown in Figures 10 and 11, respectively.

According to the comparison of the experimental results, the differences in skin temperature varied by [+ or -]0.6[degrees]C and the differences in sensible heat transfer rate varied by [+ or -]10 W/[m.sup.2] for the whole body and each local body part when the leaning posture was changed. The results indicate that the adjustment of the body's leaning posture from -15[degrees] to +30 [degrees] (-: backward, +: forward) had a very small effect on the human's thermal adaptability, since the adaptive behavior had less influence on the heat exchange characteristics over the body surface in the experimental conditions.

In cases S1, S2, and S3, the change of the sensible heat transfer rate was limited to [+ or -]5 W/[m.sup.2] at the body parts except for the hands, left arm, and left thigh. Moreover, the change of the skin temperature was small and did not exceed [+ or -]0.5[degrees]C, except for the left arm. On the other hand, the temperature decrease at the left arm in case S3 was over 1.0[degrees]C compared with case S1. This indicates that local thermal conditions can be improved by adjusting the body's orientation relative to the spot airflow. However, as the skin temperature and the sensible heat transfer rate becomes leftright asymmetrical when the angle between the body and the spot airflow is enlarged, it may make people feel less comfortable.

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

According to the comparison of cases S1, S4, and S5, there was almost no change in skin temperature and sensible heat transfer rate for the overall body and for the head, back, legs, and feet, etc., which were not covered by the spot airflow. When the human body was moved farther away from the table in the Y direction (Figure 6), the changes of skin temperature and sensible heat transfer rate were greater at the face, neck, breast, right upper arm, and right hand, which were exposed to the air. Moreover, the influence of the spot airflow became smaller due to the skin temperature increase at the face and neck. However, when the human body was moved in the X direction parallel to the desk (Figure 6), the sensible heat transfer increased and the skin temperature decreased at the left upper limbs. An opposite phenomenon was evident at the right upper limbs. Therefore, such positional adjustment may improve local thermal comfort.

CONCLUSIONS

The velocity distribution formed in the task region when a spot PAC was used was measured using an ultrasonic anemometer. According to the results, a greater velocity gradient was formed in the flow field examined. This suggests that the wind conditions around a human body can be adjusted by changing its posture and position.

The thermal adaptive effect of the behavior to change the body's leaning posture was examined experimentally using a thermal manikin exposed to the spot airflow supplied by a spot PAC. According to the results, it can be concluded that the adjustment of the lean posture exerts a comparatively small effect on the improvement of the thermal comfort, although it can influence the flow field around the face.

The coupled simulation of convection, radiation, moisture transport, and Fanger's human thermal model was used to simulate the heat transfer characteristics of a seated human body exposed to the spot airflow. Based on the simulation results, the thermal adaptive effect of adjustments of the body's orientation and position relative to the spot airflow were examined. The results indicate that people can improve their local thermal comfort by adjusting their orientation and position relative to the spot airflow when using a spot PAC with a prominent effect on cooling the human body.

ACKNOWLEDGMENTS

This research was done as part of the "Revolutionary Simulation Software for the 21st Century (RSS21)" project supported by Research and Development for Next-Generation Information Technology of the Ministry of Education, Culture, Sports, Science and Technology. Part of this study was supported by NEDO (New Energy and Industrial Technology Development Organization, Japan) and the Hans Christian [empty set]rsted Postdoc Programme of Technical University of Denmark. The authors would like to express sincere thanks to all those who supported our study.

REFERENCES

ASHRAE. 1992. ANSI/ASHRAE Standard 55-1992, Thermal Environmental Conditions for Human Occupancy. Atlanta: American Society for Heating, Refrigerating and Air-Conditioning Engineers, Inc.

Brager, G.S., and R.J. de Dear. 1998. Thermal adaptation in the built environment: A literature review. Energy and Buildings 27:83-96.

Cermak, R., A.K. Melikov, L. Forejt, and O. Kovar. 2006. Performance of personalized ventilation in conjunction with mixing and displacement ventilation. HVAC&R Research 12(2):295-312.

de Dear, R.J., and G. Brager. 1998. Developing an adaptive model of thermal comfort and preference. ASHRAE Transactions 104(1):145-67.

Fanger, P.O. 1973. Thermal Comfort--Analysis and Applications in Environmental Engineering. New York: McGraw-Hill Book Company.

Gagge, A.P., J.A.J. Stolwijk, and Y. Nishi. 1971. An effective temperature scale based on a simple model of human physiological regulatory response. ASHRAE Transactions 77(1):247-62.

Gebhart, B. 1959. A new method for calculating radiant exchanges. ASHRAE Transactions 65:321-32.

Howell, J.R., and M. Perlmutter. 1964. Monte carlo solution of thermal transfer through radiation media between gray walls. Transactions of ASME 86:116-22.

ISO. 1994. ISO-7730, Moderate Thermal Environments-- Determination of PMV and PPD Indices and Specification of the Conditions for Thermal Comfort. International Standards Organization.

Lien, F.S., W.L. Chen, and M.A. Leschziner. 1996. Low- Reynolds-number eddy-viscosity modelling based on non-linear stress-strain/vorticity relations. Proceedings of 3rd Symposium on Engineering Turbulence Modelling and Measurement, pp.1-10.

Nishi, Y., et al. 1997. Mechanism of Thermal Comfortable Environment--Aim at a Plentiful Living Space, Chapter 2, p. 30. Society of Heating, Air-Conditioning and Sanitary Engineers of Japan.

Omori, T., J.H. Yang, S. Kato, and S. Murakami. 2004. Coupled simulation of convection and radiation on thermal environment around an accurately shaped human body. Proceedings of the 9th International Conference on Air Distribution in Rooms, Coimbra, Portugal, September 5-8. CD-ROM.

Omori, T., S. Yamaguchi, and H. Taniguchi. 1998. Accurate Monte Carlo simulation of radiative heat transfer with unstructured grid systems. 11th International Symposium on Transport Phenomena, Hsinchu, Taiwan, November 29-December 3, pp. 567-73.

Raffel, M., C.E. Willert, and J. Kompenhans. 1998. Particle Image Velocimetry. Berlin, Heideberg: Springer.

Tanabe, S., E.A. Arens, F.S. Bauman, H. Zhang, and T.L. Madsen. 1994. Evaluating thermal environments by using a thermal manikin with controlled skin surface temperature. ASHRAE Transactions 100(1):39-48.

Yang, J.H., S. Kato, and T. Chikamoto. 2004. Airflow measurement around human body with wide-cover personal air-conditioning. Proceedings of the 9th International Conference on Air Distribution in Rooms, Coimbra, Portugal, September 5-8. CD-ROM.

Zhu, S.W., S. Kato, and J.H. Yang. 2004. Investigation of flow field in human's respiration area in a calm environment by visualization experiment and numerical analysis. Journal of Environmental Engineering, Transactions of AIJ 583:37-42.

Shengwei Zhu, PhD Shinsuke Kato, PhD Ling Yang

Member ASHRAE

Shengwei Zhu is a postdoctoral research fellow at the International Centre for Indoor Environment and Energy, Department of Mechanical Engineering, Technical University of Denmark, Lyngby, Denmark. Shinsuke Kato is a professor at the Institute of Industrial Science, University of Tokyo. Ling Yang is a graduate student in the Department of Architecture, School of Engineering, University of Tokyo.