- Email: [email protected]
An individualized human thermoregulation model for Chinese adults
Building and Environment 70 (2013) 257e265
Contents lists available at ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
An individualized human thermoregulation model for Chinese adults Xin Zhou, Zhiwei Lian*, Li Lan School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 19 May 2013 Received in revised form 23 August 2013 Accepted 30 August 2013
An individualized human thermoregulation model for the prediction of skin temperature was established for Chinese adults. Considering the differences in body size and composition between western and Chinese people, a standard Chinese model was built based on the anthropometric and physiological data of Chinese people, and then it was individualized with four parameters including height, weight, age and sex. Sensitivity analysis revealed that the difference in height and weight could result in a variance of 1.2 C in calculated mean skin temperature. Both the standard and the individualized Chinese model were tested with the experimental data of Chinese adults under different ambient conditions. Significant improvements were found in mean and local skin temperatures predicted by the standard Chinese model compared to the standard Fiala model. The maximum bias of the predicted mean skin temperature decreased from 0.79 C to 0.48 C, and that of local skin temperature changed from 2.11 C to 1.46 C. Further significant improvements were found when comparing the individualized Chinese model with the standard Chinese model. For the individualized model, the mean bias of mean skin temperature between prediction and measurement ranged from 0.20 C to 0.38 C, and the mean bias as well as its standard deviation of most local skin temperature was less than 1 C. Prediction accuracy was also validated in the extensive comparison with other researchers’ experiments on Chinese subjects. Prediction accuracy of Chinese adults’ skin temperature could be improved via the modification and individualization of thermoregulation model with Chinese physiological characteristics. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Thermoregulation modeling Skin temperature Chinese people Physiological parameter
1. Introduction Thermal comfort, associated with thermal sensation is one of the most important performance indicators for Heating, Ventilation and Air Conditioning (HVAC) systems. Currently thermal comfort criteria imbedded in standards ASHRAE Standard 55 and ISO Standard 7730 are based on the predicted mean vote (PMV) model [1], focusing on minimizing the percentage of people dissatisfied [2]. However, good agreement between PMV model and experiment was particularly found for uniform and steady-state environmental conditions [3]. On the other hand, thermal comfort may be achieved more energy-efficiently in non-uniform thermal environments. The distributions of skin temperature in such environments are different according to the region of human body, so developing a model to evaluate thermal comfort in asymmetrical environments or transient conditions has being a hotspot of recent study [4]. Furthermore, accurate models for predicting thermal comfort not only can be beneficial in avoiding malperformance in the use phase of a building [5], but also especially valuable for
* Corresponding author. Tel.: þ86 21 34204263; fax: þ86 21 34206814. E-mail address: [email protected] (Z. Lian). 0360-1323/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.buildenv.2013.08.031
predicting responses under conditions that cannot be tested ethically using human volunteers [6]. Thermal comfort model consist of physiological (thermoregulation in human body) and psychological model. Thermoregulation model, for the purpose of calculating skin temperature, is the first step of developing thermal comfort model. For practical utilization in design and evaluation of building environment, the calculated results of the model must agree well with the experimental measurements on human body. However, the thermoregulatory responses of different individuals are different, which makes the comparison impossible without a methodology to express individual difference in the model itself [7]. There were two approaches to deal with the problem of individual differences in thermoregulation model; one is from the passive systems of the body [8], such as thermal capacitance, thermal resistance, or surface area related to heat transfer, and the other is from the controlling systems of the body, such as regulatory sweating or skin blood flow. A populationbased dynamic human thermoregulation model built by Havenith [9] expanded with control equations incorporating several individual characteristics such as body surface area, mass, and body fat percentage. Although the prediction of individual heat strain had been improved by the model, a substantial part of differences in individual responses remain unexplained. Zhang et al. [10]
258
X. Zhou et al. / Building and Environment 70 (2013) 257e265
developed a model called “body builder” to translate descriptive data of an individual into a set of physiological parameters, which were then incorporated into a human thermoregulation model. Their simulated result only showed that the normal body fat group had lower skin temperature but higher rectal temperature than the lean group, yet a future work including more validations with experiments needs to be done. Study carried out by van Marken Lichtenbelt et al. [11] shows that on a group level predictions of skin temperatures can be improved when adopting individualized body characteristics and measured metabolic rate, but the predictions on an individual level were not improved. Fiala model [12] was individualized by Lichtenbelt et al. with respect to anthropometrics, body fat, and metabolic rate using the physiological data of Dutchman [13]; predictions of the individualized model were improved but a significant error remained. Furthermore this adapted model was used to predict thermal sensation and comfort in building environment [5]. Individual difference has been considered in human thermoregulation models, but ethnic difference which is closely related to the establishment of domestic building standard has received few concerns, although study indicates there are significant differences in body size between western and eastern people [14]. In China building energy consumption that largely depends on the operation of HVAC system is increasing rapidly [15,16], to balance energy consuming and thermal comfort of Chinese people, a precise model capable to predict thermal response of Chinese people is in urgent need. Thus in this study, a thermoregulation model was established for Chinese people and then individualized with four descriptive parameters of human body: height, weight, sex and age. 2. Establishment of Chinese model 2.1. Standard Chinese model Fiala thermoregulation model [12], which has been proved to be reasonable accurate [17,18], was adopted in this study after
extensive evaluation. Fiala model consists of two interacting systems: the controlled passive system and the controlling active system. In the passive system, as shown in Fig. 1 [12], the human body was subdivided into 14 segments, using cylinders stand for the trunk and extremities and a cylinder combined with sphere to represent for the face and head. Most of the body elements were divided into four layers (core, muscle, fat and skin from the inside to outside) with three sectors (anterior, posterior and inferior). Convection, radiation (long-wave and short-wave) and evaporation were considered in the heat exchange between human and environment, and heat changes via respiration was also included [20]. In the active system, central nervous system (CNS) accounts for overall changes in muscle metabolism by shivering, skin blood flow by vasodilatation and vasoconstriction, and skin moisture excretion by sweating. Local autonomic regulation was employed to modify local sweat rates, local blood flows, and tissue metabolic rates [21]. Former studies indicated that major physiological difference between ethnicity and individuals lies in the passive system [11,22]. A standard Chinese model was established based on the Fiala model with modifications in body size and composition. Human dimension of Chinese adults specified in the national standard of China [19] were employed for the establishment of the standard Chinese model. Considering the difference on body size between sex, independent model for man and woman were developed. Skin surface area (Ask) of standard Chinese model was derived from former experiment researches [23,24]. Thermophysiological parameters such as body fat percent (BF%) value for the standard Chinese model were obtained from the research of Deurenberg [14], basal metabolic rate (BMR) value from the research of Liu [25] and cardiac output (CO) was from the physiological research on Chinese [26]. Comparison on human dimension and thermophysiological parameters between Chinese and Fiala model were shown in Table 1. The weight of standard Chinese model was 19.1% lower than standard Fiala model, which leads to relatively lower value in Ask and BMR of a standard Chinese man. Body size of Chinese woman was even smaller than that specified in the
Fig. 1. Schematic diagram of the passive system in Fiala model.
X. Zhou et al. / Building and Environment 70 (2013) 257e265
259
Table 1 Comparison of basal body parameter between standard Fiala and Chinese model. Standard model
Height (m)
Weight (kg)
BF%
Ask(m2)
BMR (W)
CO (L/min)
Age
Fiala (man) Chinese man (difference) Chinese woman (difference)
1.71 1.70 (0.6%) 1.60 (6.4%)
73.5 59.5 (19.1%) 51.5 (29.9%)
14.4 16.8 (16.7%) 25.2 (75.0%)
1.86 1.71 (8.1%) 1.55 (16.7%)
87.1 80.1 (8.0%) 64.3 (26.2%)
4.9 4.8 (2.0%) 4.3 (12.2%)
/ 30/ 30/
standard Fiala model, weight, Ask and BMR of standard Chinese woman was about 30% lower than that of standard Fiala model. The most significant difference between sexes lies in body fat; BF% value of standard Chinese woman was 75% larger than that of standard Fiala model. CO value of Chinese man model was nearly the same as Fiala model, but for Chinese woman it was about 10% lower. As age is an important factor for the evaluation of thermal sensation [27], the effect of age on body composition and BMR was introduced. Chinese adults were divided into three age groups [28], i.e., 18e25 years, 26e35 years and 36e60 (for male) 55 (for female), and age of 30 years as the mean value of middle age group was chosen for the standard Chinese model. Proportional change in human dimension of the standard Fiala model was not proper for the establishment of Chinese model. As we can see in Fig. 2, the difference between models was not only in the overall skin surface area, but also in the skin surface area distribution of segments mainly the thorax, hand, upper leg and foot. Moreover, the distribution of body fat in Chinese adults was significantly different from the western people according to a former physiological research [29]. Considering the difference in skin surface area and body fat distribution between Chinese and western people, the passive system was rebuilt according to the physiological data of Chinese people (Table 2). Other thermophysiological parameters such as heat exchange coefficients, heat conductivity and capacitance, also the physical parameters of tissue materials like density were kept to be the same as the standard Fiala model [12]. 2.2. Model individualization After extensive comparison and evaluation, four easily obtained body characteristics including height, weight, sex and age were
selected as the input parameters for the individualization of Chinese model. Other parameters like BF%, one of the most important thermophysiological differences between individuals, was usually considered as an independent input parameter [9,10,13]. However, considering the practical utilization in current research, BF% and BMR value, which can be predicated by height and weight of human in a subgroup of people [14], was not chosen as an input parameter. Fig. 3 shows the flow chart of model individualization with the four input parameters; the left side shows how to calculate CO and BMR and the right side demonstrates the individualization of passive system. BMR was calculated by the formula described by Liu et al. [25].
BMR ¼ 58 W þ 1741 H 14 Age 470 Sex þ 227
(1)
CO was calculated by equation (2) [26].
CO ¼ 0:024 W 0:057 Age 0:305 Sex þ 4:544
(2)
Where H is the height (m) and W is the weight (kg) of individuals. Here Sex was coded as a dummy variable (man is 0 and woman is 1). BMR value in original formula was in kilojoules (KJ) per day, thus a unit transfer (changed to W by adding 3.6*24) was applied. Basal metabolic rate of a certain segment (BMRi) was calculated by summing up the heat production of all materials (mainly muscle, brain and viscera) in the segment. Materials’ heat production per unit volume was initialized with the value given by Fiala model (e.g. for muscle it is 684 W/m3), and then they were changed with the ratio of the calculated BMR of individualized model to the BMR of the standard Chinese model. Blood perfusion (BP) in materials was changed with the calculated CO value in a similar way as BMRi with calculated BMR.
Fig. 2. Difference between the distribution of Ask between standard Chinese model and Fiala model.
260
X. Zhou et al. / Building and Environment 70 (2013) 257e265
Table 2 Geometric parameters of the standard Chinese man model. Body segments L (m)a r (m) Core Head Face Neck Shoulder Thorax Abdomen Upper arm Lower arm Hand Upper Leg Lower leg Foot a
0.120 0.081 0.309 0.297 0.524 0.595 0.478 0.365 0.705 0.749 0.464
Muscle Fat
0.067 0.073 0.040 0.052 0.068 0.020 0.044 0.021 0.024 0.070 0.076 0.123 0.071 0.076 0.101 0.015 0.028 0.013 0.026 0.011 0.029 0.018 0.058 0.016 0.042 0.020 0.031
0.081 0.077 0.051 0.035 0.128 0.115 0.036 0.034 0.035 0.063 0.046 0.035
Inner skin Outer skin 0.0830 0.0775 0.0515 0.0355 0.1285 0.1150 0.0365 0.0345 0.0355 0.0640 0.0465 0.0355
0.084 0.078 0.052 0.036 0.129 0.116 0.037 0.035 0.036 0.065 0.047 0.037
BF% ¼ 1:38 BMI þ 0:25 Age 12:1 Sex 8:1
Here L is the length of segment; r is the radius of segment.
Anthropometrical parameters such as length of segments (Lseg) and Ask, were calculated from three of the input parameters [30].
Lseg ¼ B0 þ B1 H þ B2 W
(3)
Where Lseg is the length of body segment, B0, B1, B2 is the regression coefficients, most R values of the regression formulas listed in Table 3 were over 0.99 when compared to the measurement’s data [19]. The length of segment was changed in a separately different ratio with the height and weight, and a more precise solution for the building of passive system than proportional change was achieved with these regression formulas. Ask can be calculated by the following formulas: For man
Ask ¼ 0:00607 H þ 0:0127 W 0:0698
(4.1)
and for woman
Ask ¼ 0:00586 H þ 0:0126 W 0:0461 AskðiÞ ¼ Ask fa
segments were simplified to cylinders (head was sphere), with the length and surface area of a segment (Lseg and Ask(i)), then the outer radius of a segment (also is the outer radius of skin) can be calculated, as shown in Fig. 3. Typically there are four layers skin, fat, muscle and core from the outside to the inside in a segment, and the length of body layers are the same as the segment. After the outer radius of skin were calculated by Lseg and Ask(i), inner radius of skin which is also the outer radius of fat can be obtained as the thickness of the skin layer (1e3 mm) of all body parts was kept unchanged. Radius of layer other than skin can be calculated by the weight distribution and density of materials, such as the thickness of fat layer can be draw from the BF%. BF% value was calculated according to the method described in Deurenberg et al. [14]:
(4.2)
(6)
Where BMI was body mass index, equals to weight divided by square of height; Sex here was coded as a dummy variable (man is 1 woman is 0). For a certain segment, volume of fat layer was calculated by the BF% value with fat distribution coefficients (fb, see in Table 3) and the density of fat; then together with the inner radius of skin layer (also the outer radius of fat layer), the inner radius of fat (rfat) which was also the outer radius of muscle can be calculated. According to physiological measurements [28,31], weight of core layers (Wcore) including bones, brain (in head), lung (in thorax) and viscera (in abdomen) were given, and also weight distribution coefficient of core layer (fc) for man and woman were listed in Table 3. With Wcore, fc and the density of core materials, the core radius (rcore) for a standard Chinese were calculated, and rcore for an individual was changed with height in a linear relationship with a correlation coefficient of 0.06. Radius of muscle (rmuscle) was determined by the thickness of fat and core, since the outer radius of muscle was the inner radius of fat layer and the inner radius of muscle was the same as rcore. With the length of segments and radius of all material layers, the passive system for Chinese individual was established.
(5) 2.3. Sensitivity analysis
Where fa is the distribution factor for segment (see in Table 3) derived from experimental measurement [23,24]. Since the body
Fig. 3. Flow chart of Chinese model individualization.
Sensitivity analysis was performed to evaluate how body temperatures were affected by the four input parameters employed in the individualized Chinese model. A typical mildly cold environment was set up for this analysis, in which an equal ambient and radiant temperature of 20 C, a relative humidity of 50% and an air velocity of 0.05 m/s were maintained. Subjects kept sedentary (1 met) in the environment for 60 min, wearing clothes of 0.6 clo. Body height and weight of 5th and 95th percentile of Chinese adults were chosen according to the national standard of China [19], and percentiles of age was defined as 5% (20 years old), 50% (30 years old) and 95% (60 for male and 55 for female). When one of the parameters was changed, the other three input parameters were kept to be the same value as the standard Chinese model. Skin and rectal temperature changes between individuals were demonstrated in Fig. 4. Compared with the standard model (50th percentile), the mean skin temperature (Tskm) of man with the same weight, age but differ in height have a difference of 1.2 C. Rectal temperature (Trectal) and local skin temperature of a shorter man was relative lower than the standard man; on the contrary, the man of 95th percentile in height has a relatively higher value both in skin and rectal temperature. A taller adult, who has the same weight and age with the short one, would have a relative larger value in Ask but lower in BF%. Since the body fat with lower in
X. Zhou et al. / Building and Environment 70 (2013) 257e265
261
Table 3 Regression parameters and distribution coefficients for the individualization of Chinese model. Body segments
Man
Woman
Length
Head Face Neck Shoulder Thorax Abdomen Upper arm Lower arm Hand Upper Leg Lower leg Foot
Ask
(i)
BF(i)
Wcore
Length
Ask
(i)
BF(i)
Wcore
B0
B1
B2
R
fa
fb
fc
B0
B1
B2
R
fa
fb
fc
0.00569 0.00799 0.00542 0.03041 0.02934 0.06005 0.07866 0.1631 0.01231 0.25105 0.30984 0.00155
0.043 0.089 0.060 0.152 0.145 0.246 0.362 0.354 0.186 0.548 0.598 0.252
0.00026 0.00038 0.00026 0.00033 0.00035 0.00076 0.00097 0.00066 0.00062 0.00041 0.00071 0.00058
0.968 0.968 0.968 0.999 0.999 0.999 0.999 0.999 0.999 0.999 0.999 0.999
4.05% 2.06% 1.48% 1.61% 13.63% 21.17% 8.09% 6.41% 4.93% 16.33% 13.29% 6.95%
5.0% 4.0% 1.0% 1.0% 9.0% 42.5% 8.5% 5.5% 3.0% 11.3% 6.6% 2.6%
8.0% 0.5% 0.4% 7.1% 12.3% 23.6% 7.5% 4.5% 2.5% 12.6% 10.6% 10.4%
0.00569 0.00860 0.00542 0.03041 0.00593 0.01214 0.19539 0.22079 0.01224 0.41748 0.40474 0.02318
0.043 0.086 0.060 0.132 0.127 0.280 0.468 0.394 0.206 0.769 0.680 0.288
0.00026 0.00038 0.00026 0.00031 0.00016 0.00032 0.00048 0.00051 0.00057 0.00026 0.00037 0.00047
0.967 0.967 0.967 0.994 0.994 0.994 0.993 0.994 0.995 0.993 0.993 0.994
4.07% 2.26% 1.05% 2.94% 8.21% 22.31% 8.29% 5.71% 4.52% 21.16% 12.83% 6.65%
4.0% 1.5% 1.0% 1.0% 11.0% 44.0% 8.0% 5.0% 3.0% 13.5% 6.0% 2.0%
7.3% 7.8% 0.4% 5.6% 10.6% 24.8% 6.4% 3.7% 2.0% 13.5% 9.1% 8.9%
heat conductivity (38% as to muscle) plays an important role in preventing heat conduction through human body, thus heat conduction value of a taller person from the core layer was higher and eventually result in a relatively higher skin temperature. A lager BMR value was another reason for the higher rectal temperature of the taller person, which also made contribution to the higher value of skin temperature. Mean and rectal temperature also increased with the height of woman model. Differ in weight had another effect on skin and rectal temperature. A lean person was with a relatively higher value in skin temperature but lower in rectal temperature, which agreed well with the physiological measurements of skin temperature on overweight and lean man [10,32]. Lower in BF% and BMR could explain the difference on body temperature of lean people. A lean person with less body fat cannot resist heat transfer from the core layer as effectively as normal person, and lower BMR value which means less heat generation in human body leads to a relatively lower rectal temperature. Compared to height and weight, differ in age caused less difference in skin temperature, but the rectal temperature decreases with the age was in a considerable gradient, which was mainly caused by the negative correlation between BMR value and age [27]. Compared with the man model of the same percentile, relatively lower in mean skin temperature but higher in rectal temperature was found for woman model. Calculation results of woman model also agreed well with the experimental measurement [33], which can be explained by the lower BMR and CO value together with the higher BF% value of woman.
Fig. 4. Effect of input parameters on mean skin temperature (Tskm) and rectal temperature (Trectal).
3. Experimental validation 3.1. Experiment protocol The established models were validated with experiment research on Chinese adults. Twenty two healthy non-smokers (11 men and 11 women) from local university participated in this experiment. Fig. 5 shows the height and weight distribution of the subjects, which were located in the range of 5% and 95% percentile of Chinese adults, and most of them were fitted well with the regression curve [19]. Thus the subjects recruited were qualified to stand for typical Chinese adults. Subjects were asked to avoid caffeine, alcohol, smoking, and intense physical activity at least 12 h prior to each experimental session. Summer clothing (vest and shorts) with a total resistance of about 0.3 clo was compulsory for every subject. All experiment protocols were approved by the university’s ethics committee and conformed to the guidelines contained within the Declaration of Helsinki. The whole experiment consists of two parts: “temperature-rise” phase and “temperature-fall” phase. Both phases followed the same schedule which was demonstrated in Fig. 6. After 30 min sitting and resting in a thermal-neutral environment as a preparation period, the subjects were then divided into two groups and in each group a different combination of ambient temperatures was designed. Group 1 (6 men and 5 women, G1) was exposed to an indoor
Fig. 5. Height and weight distribution of subjects.
262
X. Zhou et al. / Building and Environment 70 (2013) 257e265
Fig. 6. Schedule of the validation experiment.
environment with the temperature order of 21 C, 24 C, 26 C and 29 C (from S1 to S4), for group 2 (6 men and 5 women, G2) the order from S1 to S4 was 29 C, 26 C, 24 C and 21 C. In this study, the measurements were made after that the subject had stayed in a steady thermal environment for at least 40 min, which was proved to be a steady state for mean skin temperature according to the early study [34]. The experiment was performed in a climate chamber during May. There was only a window in the climate chamber and no direct solar radiation entered. The air temperature was controlled using a wall-mounted air conditioner. The indoor air velocity near the subjects was kept under 0.05 m/s, and the relative humidity of air was not dependently controlled but recorded by a hygrometer. In order to obtain the stable skin temperature in the environment, we tested the skin temperature for 5 min before the end of every stage. Under each ambient temperature, a subject’s local skin temperatures at 7 points were measured according to the research of Hardy et al. [35]. 3.2. Validation with the experiment When calculating the skin temperatures for each subject with the individualized model, the simulation follows the same schedule and environmental set up as the experiments. Table 5 shows the mean difference with standard deviation (SD) between predicted by thermoregulation model and experiment measurement. In order to avoid difference may be covered in the average calculation process, absolute value of the skin temperature difference between experiment and simulation was employed. In the temperature-rise phase, the mean difference between standard Fiala model and experiment in mean skin temperature was about 0.5 C in all stages. The prediction of head temperature (Thead anterior) was the most accurate; however, in other parts of human body like lower arm, hand back and foot back, the mean difference was exceed 1.5 C in most conditions, also with a relative higher SD value (exceeding 1 C in most conditions). Mean difference of standard Chinese model was lower than standard Fiala model in most conditions, paired sample T test results showed that there were significant improvements in both mean and local skin temperature, especially for the skin temperature of thorax, hand and foot. The difference mainly derived from the difference in body composition and skin surface area between Chinese and Fiala model. Although mean difference has been improved by
modification of the body parameter, standard deviation value of the prediction remained larger than 1 C in some conditions. Compared with standard Chinese model, prediction accuracy of individualized Chinese model was significantly improved, both in mean value and standard deviation. Mean difference of mean skin temperature was between 0.20 and 0.33 C, and the standard deviation was less than 0.37 C. Paired sample T test also showed a significant difference between the standard and individualized Chinese model. In temperature-fall phase, significant differences also can be found in mean and local skin temperature between standard Fiala model and Chinese model. The skin temperature difference of Fiala model in the last stage was increasing rapidly, while the prediction of Chinese model kept nearly the same accuracy in all the stage of experiments. Significant differences in skin temperature were also found between individualized Chinese model and standard Chinese model. In most local skin temperature evaluations, the differences between models were more significant in the last stage of experiment when the ambient temperature was 29 C. In both temperature-rise and temperature-fall experiments, the prediction of Chinese model kept in a relative stable level, whereas the mean difference of Fiala model in most conditions ascended with the experiment time. 4. Extensive validations The individualized Chinese model was also validated with the experiment data performed by other researchers on Chinese subjects. Simulations were carried out with the same ambient condition as well as the subject’s activity level and clothing with the experiments. As the detail information of each subject was often not provided by other researchers, but only with the mean value of height, weight, and age of the subjects. Thus mean values of the subjects’ parameters were employed in the validation of individualized Chinese model. Some experiments did not report subject’s information and thereby standard Chinese model was applied. Validation with a typical experiment was shown as follows [36]. After preparation and acclimation for 30 min, the subjects were exposed to an indoor environment with certain ambient temperature for 60 min, the ambient temperature ranges from 20 C to 34 C (cold to warm for subjects), and relative humidity was recorded and applied into the simulation. 15 healthy male Chinese subjects were recruited for the experiments study, every subject was sitting on the chairs and kept at rest (1 met) dressing in a summer cloth (0.6 clo). Fig. 7 shows the results of mean skin temperature and rectal temperature in experiments. Simulation results of mean skin temperature agreed well with the experiments under all of the ambient temperatures, and most of the rectal temperature value which increasing very slowly with the ambient temperature was also well agreed with the experiment results. The comparison of local skin temperature was demonstrated in Fig. 8, most of differences between the simulation and mean value of experiment results were in the range of 1 C. Well agreement
Table 4 Subjects’ information and environmental conditions and of experiments for extensive validation. Subjects’ information
Environmental conditions
Reference
Number
Age (year)
Height (m)
Weight (kg)
Thermal resistance (clo)
Activity level (met)
Ambient temperature ( C)
Relatively humidity (%)
Background air velocity (m/s)
30
20.1
1.67
56.9
0.6
1
24 16
23.2 /
1.68 /
60.2 /
0.6 0.6
1 1
40
/
/
/
0.6
1
20, 29, 28 20, 25, 30 20,
50 70 50 80 30 50 50
0.05 0.05 0.05 0.05 0.05 0.05 0.05
23, 29, 32 32 30 30 25, 28
[37] [38] [39]
[40]
Table 5 Mean difference ( C) standard deviation between model and measurements for mean and local skin temperature in different ambient temperature (AT).
Temperature-rise phase
At ( C)
Tskm
Standard Fiala
21 24 26 29 21 24 26 29 21 24 26 29 29 26 24 21 29 26 24 21 29 26 24 21
0.45 0.42 0.57 0.60 0.40 0.42 0.45 0.48 0.23 0.20 0.31 0.33 0.51 0.70 0.60 0.79 0.43 0.48 0.43 0.43 0.38 0.35 0.34 0.33
Standard Chinese
Individualized Chinese
Temperature-fall phase
Standard Fiala
Standard Chinese
Individualized Chinese
Thead
0.62 0.58 0.70 0.56 0.52: 0.45 0.57: 0.44:: 0.37* 0.22* 0.30** 0.20** 0.48 0.48 0.57 0.68 0.49 0.56:: 0.56::: 0.60:: 0.40** 0.37** 0.44** 0.32*
0.51 0.43 0.50 0.31 0.52 0.41 0.48 0.29 0.56 0.37 0.41 0.39 0.67 0.68 0.32 0.55 0.69 0.41 0.33 0.61 0.63 0.37 0.49 0.33
anterior
0.54 0.27 0.33 0.46 0.52 0.25 0.30 0.46 0.27* 0.23** 0.26 0.41 1.06 0.50 0.37 0.67 1.06 0.45:: 0.39 0.71::: 0.58** 0.44** 0.38*** 0.42***
Tthorax 0.77 0.98 0.96 0.68 0.71 0.87 0.64 0.54 0.72 0.79 0.78 0.57 0.88 1.68 1.37 1.68 0.92 1.44 1.21 0.98 0.77 1.00 0.92 0.80
anterior
0.85 0.93 0.85 0.66 0.89 0.87:: 0.79:: 0.61::: 0.47 0.63** 0.60** 0.39** 1.00 1.27 1.24 1.16 1.07 1.31::: 1.26::: 0.74::: 0.72** 0.93** 0.93*** 0.88***
Tlower 1.64 1.53 1.84 1.92 1.20 1.24 1.43 1.50 1.11 1.01 1.07 1.08 0.89 0.55 0.69 1.15 0.65 0.71 0.69 1.08 0.52 0.57 0.76 0.69
arm anterior
1.52 1.85 2.11 2.39 1.24:: 1.58 1.83 2.14: 0.73 1.01 0.84* 0.84** 0.67 0.74 0.96 1.33 0.64:: 0.73 0.99 1.03 0.47*** 0.59* 0.68 0.94
Thand 1.91 1.85 1.80 1.03 1.46 1.41 1.41 0.85 1.21 0.43 0.57 0.80 1.06 1.19 1.27 1.89 0.59 0.63 0.94 1.12 0.59 0.84 0.76 0.94
back
1.65 1.98 2.04 1.01 1.48:: 1.81:: 1.90::: 1.04::: 0.87 0.84** 1.18 0.59** 1.30 0.99 1.38 2.22 0.85::: 0.85::: 1.32::: 1.66::: 0.71** 0.62* 0.80* 1.12
:
Tupper 1.23 0.88 0.89 0.92 1.17 0.81 0.75 0.67 0.88 0.80 0.79 0.78 0.93 1.50 1.10 1.17 1.07 1.38 0.89 0.89 0.34 0.63 0.72 0.53
leg anterior
1.81 1.19 1.05 0.85 1.73 1.09 0.89 0.79:: 0.77* 0.55 0.39* 0.66* 0.71 1.18 0.88 1.20 0.70 1.14 0.84:: 0.89:: 0.62*** 0.89* 0.83* 0.76*
Tlower 0.98 0.77 0.89 0.99 0.68 0.78 0.85 0.72 0.87 0.88 0.70 0.61 0.49 0.60 0.54 0.98 0.48 0.60 0.56 0.56 0.46 0.71 0.69 0.78
leg anterior
1.17 0.97 0.99 0.66 1.16:: 0.93:: 0.97:: 0.56 0.61* 0.81* 0.62** 0.56** 0.59 0.83 0.89 1.12 0.61 0.80 0.86 0.96 0.44** 0.48* 0.60* 0.45
Tfoot
back
1.75 1.59 2.11 1.72 1.02 0.88 1.41 1.04 0.47 0.54 0.59 0.23 0.87 0.76 0.79 1.36 0.80 0.75 0.63 0.63 0.72 0.84 0.36 0.47
1.40 1.08 1.92 1.90 1.40: 1.09: 1.93: 1.90::: 0.56** 0.60** 0.75** 0.36*** 0.87 0.93 0.78 1.31 0.88 0.93:: 0.77::: 0.83::: 0.45** 0.51** 0.58** 0.56
P < 0.05 :: P < 0.01 and ::: P < 0.001 when standard Fiala model compared with standard Chinese model; *P < 0.05, **P < 0.01 and***P < 0.001 when standard Chinese model compared with Individualized Chinese model.
263
Fig. 7. Comparison of mean skin temperature and rectal temperature between experiment and simulation.
could be found in forehead, thorax, arm and leg, with difference no larger than 1 C in all of environmental conditions; but the difference of foot and hand in the mild cold and neutral environment exceeded 2 C. Extensive validations of individualized Chinese model with other researchers’ experiments were shown in Fig. 9. Environmental conditions and subjects’ information of experiments were shown in Table 4 [37e40]. Difference in mean skin temperature was mainly in the range of 0.5 C, and the difference between experiment and simulation decreased with the ambient temperature, especially when the ambient temperature exceeds 30 C. As the ambient temperature rises, differences in skin temperature between segments of human body were narrowed down, thus the prediction of skin temperature was improved. Local skin temperature value was divided into two parts: proximal (including head, thorax, abdomen and upper arms/legs), and distal (lower arms/legs, feet and hands). As the skin temperature in distal part was more easily affected by the ambient environment, and in proximal part was mainly controlled by the thermoregulation mechanism. Prediction results of the two parts were different. For skin temperature in proximal part, though the difference was mainly within the
Fig. 8. Comparison of local skin temperature between experiment and simulation.
X. Zhou et al. / Building and Environment 70 (2013) 257e265
Models
264
X. Zhou et al. / Building and Environment 70 (2013) 257e265
achieved. The four input parameters in this model may be not adequate to predict the body information of each individual Chinese adult very precisely, but in a reasonable range, and also more practical in design and evaluation of indoor environment. Moreover, this method of model individualization was not limited to Chinese, but also can be extensive used for other people based on the local physiological research and validated by human experiments. 5.2. Potential applications
Fig. 9. Comparison of mean skin temperature (Tskm), skin temperature of proximal (Tproximal) and distal (Tdistal) between experiment and simulation.
range of 1 C, most simulation results was larger than the experiment measurement, indicating a slight modification in controlling active system was needed. For distal part, the skin temperature of simulation was slightly lower than experiment (most differences lies in the range of 1 C), especially when the ambient temperature was larger than 30 C. Through extensive validations, the prediction accuracy of individualized Chinese model was proved to be adequate according to former research [41], the difference for mean skin temperature was mainly less than 0.5 C, and for local skin temperature was less than 1 C. 5. Discussion 5.1. Input parameters Accuracy in predication and practicability in use are the major principles for the selection of input individual parameters. There are a few studies which discussed the individualization with proper input parameters. Sensitive analysis of height, mass, BF% and resting metabolic rate (RMR) was involved in the research of Lichtenbelt [42]. In their analysis, when a parameter is changed to the 5th or 95th percentile value, all other parameters are kept at the average value, which are not practical for real people. For instance, two ordinary persons with different weight (60 kg and 80 kg) but same height (1.70 m), their BF% should not be the same. Also it is the same with the RMR. The four parameters (height, mass, BF% and RMR) have inherent relationship with each other which should not be considered as independent input parameters. Input parameters used in the individualization of UC Berkley model are the height, weight, gender, body fat (or dimensional information) and skin color [10]. Body fat was calculated through the measurement of skinfold caliper, which needs the measurement of circumference at levels of umbilicus, minimal abdominal width at approximately midway between xiphoid and umbilicus. Metabolic rate (MR) and body composition (BC) were selected as input parameters for the Fiala model individualization [11]. MR was calculated from the O2 consumption and CO2 production measured by indirect calorimetry, using a ventilated hood system, and the model achieved the best accuracy with the input of MR and BC. The involvement of BF or MR measurements in the former model may be too complex for practical prediction of skin temperature for individuals, and the individualization of thermoregulation model should consider more about practical application when an acceptable accuracy had been
The incorporation of thermoregulation model and thermal psychological model, which could provide a practical link between building environment and human comfort, is the next step of work on Chinese model. Combination of computation fluid dynamics (CFD), thermal thermoregulation model and thermal sensation model were carried out by many researchers recently [4,5]; these studies indicated that the thermoregulation models were capable for the prediction of thermal sensation based on skin temperature. Fig. 10 shows one possible application of this Chinese model for predicting thermal comfort based on experiment carried out by Yu [43] and Liu [44]. The subjects in Yu’s experiments were exposed in the cold environment with ambient temperature of 20 C (mildly cool), 16 C (cool) and 12 C (cold). A simulation was carried out with the same environmental condition and subjects’ information, and a good agreement has been found in mean skin temperature changes between simulation results and experiment investigation. Meanwhile, a comfort zone (Tskm between 32.72 C and 33.74 C) for Chinese adults in uniform environment was given by experiment research form Liu [44], then the calculated mean skin temperature could be used to roughly estimate whether Chinese people is comfortable, eventually this model can be used for the design and evaluation of uniform indoor environment, for guaranteeing the thermal comfort for Chinese adults. Through the modification of standard Fiala model and individualization of standard Chinese model, prediction accuracy was significantly improved in most conditions, but still left some discrepancies in local skin temperature prediction which remained unexplained. As our knowledge about how inter-individual difference affects the human thermoregulation system was still scant, further experiments to cover wider population and environmental conditions are necessary. Further refinements of the model may include the individual difference between thermal response,
Fig. 10. Potential application of the model on prediction of thermal comfort in uniform environment.
X. Zhou et al. / Building and Environment 70 (2013) 257e265
especially when accuracy prediction on an individual level or prediction in the clinic under special thermal conditions. 6. Conclusion 1. A standard Chinese model was established based on the anthropometric and physiological study on Chinese adults. This model was then individualized with four input parameters including height, weight age and sex. Sensitivity analysis revealed that the four input parameters could have great impacts on mean skin temperature (up to 1.2 C). 2. The standard Chinese model was tested with experimental data on Chinese people. Results indicated that compared with the standard Fiala model, the standard Chinese model significantly improved the prediction accuracy in mean and most local skin temperature. The maximum difference in mean skin temperature between the Fiala model prediction and experimental data was reduced from 0.79 C to 0.48 C by the standard Chinese model, and the maximum difference in local skin temperature decreased from 2.11 C to 1.46 C. 3. The prediction accuracy was further improved by the individualized Chinese model, which greatly decreased the standard deviations of prediction for individuals. For the individualized model, the mean difference of mean skin temperature between model prediction and experimental data ranged from 0.20 C to 0.38 C; the standard deviation of most predicted local skin temperature was less than 1 C, compared to 2.39 C when predicted by the standard Chinese model.
[14]
[15] [16] [17]
[18]
[19] [20]
[21]
[22]
[23] [24] [25] [26] [27] [28] [29]
Acknowledgments
[30]
This work is financially supported by the Key project and General Project of National Natural Science Foundation of China (No. 51238005 and 51108260).
[31]
References
[33]
[1] Fanger PO. Thermal comfort. Copenhagen: Danish Technical University; 1970. [2] van Marken Lichtenbelt WD, Kingma BR. Building and occupant energetics: a physiological hypothesis. Archit Sci Rev 2013:1e6. [3] Van Hoof J. Forty years of Fanger’s model of thermal comfort: comfort for all? Indoor Air 2008;18(3):182e201. [4] Cheng Y, Niu J, Gao N. Thermal comfort models: a review and numerical investigation. Build Environ 2012;47(0):13e22. [5] Schellen L, Loomans MGLC, Kingma BRM, de Wit MH, Frijns AJH, van Marken Lichtenbelt WD. The use of a thermophysiological model in the built environment to predict thermal sensation. Build Environ 2013;59:10e22. [6] Xu X, Tikuisis P, Gonzalez R, Giesbrecht G. Thermoregulatory model for prediction of long-term cold exposure. Comput Biol Med 2005;35(4):287e98. [7] Takada S, Kobayashi H, Matsushita T. Thermal model of human body fitted with individual characteristics of body temperature regulation. Build Environ 2009;44(3):463e70. [8] Gagge A. An effective temperature scale based on a simple model of human physiological regulatory response. ASHRAE Trans 1971;77:247e62. [9] Havenith G. Individualized model of human thermoregulation for the simulation of heat stress response. J Appl Physiol 2001;90(5):1943e54. [10] Zhang H, Huizenga C, Arens E, Yu T. Considering individual physiological differences in a human thermal model. J Therm Biol 2001;26(4e5):401e8. [11] van Marken Lichtenbelt WD, Frijns AJH, Fiala D, Janssen FEM, van Ooijen AMJ, van Steenhoven AA. Effect of individual characteristics on a mathematical model of human thermoregulation. J Therm Biol 2004;29(7e8):577e81. [12] Fiala D, Lomas KJ, Stohrer M. A computer model of human thermoregulation for a wide range of environmental conditions: the passive system. J Appl Physiol 1999;87(5):1957e72. [13] Marken Lichtenbelt WD, Frijns AJH, Ooijen MJ, Fiala D, Kester AM, Steenhoven AA. Validation of an individualised model of human thermo
[32]
[34] [35] [36]
[37]
[38]
[39] [40] [41] [42]
[43]
[44]
265
regulation for predicting responses to cold air. Int J Biometeorol 2006;51(3): 169e79. Deurenberg P, Ge K, Hautvast JGAJ, Wang J. Body mass index as predictor for body fat: comparison between Chinese and Dutch adult subjects. Asia Pac J Clin Nutr 1997;6(2):102e5. Pérez-Lombard L, Ortiz J, Pout C. A review on buildings energy consumption information. Energy Build 2008;40(3):394e8. Hong T. A close look at the China design standard for energy efficiency of public buildings. Energy Build 2009;41(4):426e35. Fiala D, Lomas KJ, Stohrer M. Computer prediction of human thermoregulatory and temperature responses to a wide range of environmental conditions. Int J Biometeorol 2001;45(3):143e59. Psikuta A, Fiala D, Laschewski G, Jendritzky G, Richards M, Blazejczyk K, et al. Validation of the Fiala multi-node thermophysiological model for UTCI application. Int J Biometeorol 2012;56(3):443e60. Human dimension of Chinese adults. Beijing: CNIS; 1988. Fiala D, Psikuta A, Jendritzky G, Paulke S, Nelson DA, van Marken Lichtenbelt WD, et al. Physiological modeling for technical, clinical and research applications. Front Biosci 2010;2:939e68. Fiala D, Havenith G, Brode P, Kampmann B, Jendritzky G. UTCI-Fiala multinode model of human heat transfer and temperature regulation. Int J Biometeorol 2012;56(3):429e41. van Marken Lichtenbelt WD, Westerterp-Plantenga MS, van Hoydonck P. Individual variation in the relation between body temperature and energy expenditure in response to elevated ambient temperature. Physiol Behav 2001;73(1e2):235e42. Zhao S, Liu Y, Yao J, Yang Z, Liang Y, Zhang S. The estimation of body surface area of adults Chinese females. Acta Nutrimenta Sinica 1987;9(3):200e7. Zhao S, Liu Y, Yao J, Gao S, Zhang S. The estimation of body surface area of adults Chinese males. Acta Nutrimenta Sinica 1984;6(2):87e95. Liu HY, Lu YF, Chen WJ. Predictive equations for basal metabolic rate in Chinese adults: a cross-validation study. J Am Diet Assoc 1995;95(12):1403e8. Zhen Y. Physiology. Beijing: Higer Education Press; 2010. Guergova S, Dufour A. Thermal sensitivity in the elderly: a review. Ageing Res Rev 2011;10(1):80e92. Inertial parameters of adult human body. Beijing: CNIS; 2004. Lu H, Fu X, Ma X, Wu Z, He W, Wang Z, et al. Relationships of percent body fat and percent trunk fat with bone mineral density among Chinese, black, and white subjects. Osteoporos Int 2011;22(12):3029e35. Zhang C, Lin D. Parameters research on the adult anthropomorphic phantom of china. Machinery 2009;36(7):1e3. Rusong C, Wang Z, Ouyang B. Reference values of skeleton mass for normal Chinese adults. Chin J Radiological Med Prot 1995;15(2):317e20. Claessens-van Ooijen AM, Westerterp KR, Wouters L, Schoffelen PF, Van Steenhoven AA, Van Marken Lichtenbelt WD. Heat production and body temperature during cooling and rewarming in overweight and lean men*. Obesity 2006;14(11):1914e20. van Ooijen AMJ, van Marken Lichtenbelt WD, Westerterp KR. Individual differences in body temperature and the relation to energy expenditure: the influence of mild cold. J Therm Biol 2001;26(4e5):455e9. Dear R, Ring J, Fanger P. Thermal sensations resulting from sudden ambient temperature changes. Indoor Air 1993;3(3):181e92. Hardy JD, Du Bois EF, Soderstrom G. The technic of measuring radiation and convection one figure. J Nutr 1938;15(5):461e75. Yu J. Studies on the effects of physiological acclimatization on thermal responses of people accustomed to different indoor thermal environments [Ph.D Thesis]. Shanghai: Donghua Unviersity; 2011. Chen H, Zhang Y, Wang J, Meng Q. Human thermal comfort research in hothumid area of China (1): experimental methods and results. HV&AC 2010;40(3):83e8. Wang Y, Lian Z, Lan L. The effect of turbulence intensity on local skin temperature and subjective responses to draft. Energy Build 2011;43(10): 2678e83. Peng Y, Lian Z, Pan L. Skin temperature measurement method. J Southeast Univ 2010;26(2):258e61. Sun Y. Experimental study on human thermal sensation based on skin temperature. Dalian: Dalian University of Technology; 2009. Huizenga C, Hui Z, Arens E. A model of human physiology and comfort for assessing complex thermal environments. Build Environ 2001;36(6):691e9. van Marken Lichtenbelt WD, Schrauwen P, van de Kerckhove S, WesterterpPlantenga MS. Individual variation in body temperature and energy expenditure in response to mild cold. Am J Physiol Endocrinol Metab 2002;282(5): E1077e83. Yu J, Ouyang Q, Zhu Y, Shen H, Cao G, Cui W. A comparison of the thermal adaptability of people accustomed to air-conditioned environments and naturally ventilated environments. Indoor Air 2012;22(2):110e8. Liu W. Study on objective evaluation index of human thermal comfort [Ph.D Thesis]. Shanghai: Shanghai Jiao Tong University; 2008.
Contents lists available at ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
An individualized human thermoregulation model for Chinese adults Xin Zhou, Zhiwei Lian*, Li Lan School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 19 May 2013 Received in revised form 23 August 2013 Accepted 30 August 2013
An individualized human thermoregulation model for the prediction of skin temperature was established for Chinese adults. Considering the differences in body size and composition between western and Chinese people, a standard Chinese model was built based on the anthropometric and physiological data of Chinese people, and then it was individualized with four parameters including height, weight, age and sex. Sensitivity analysis revealed that the difference in height and weight could result in a variance of 1.2 C in calculated mean skin temperature. Both the standard and the individualized Chinese model were tested with the experimental data of Chinese adults under different ambient conditions. Significant improvements were found in mean and local skin temperatures predicted by the standard Chinese model compared to the standard Fiala model. The maximum bias of the predicted mean skin temperature decreased from 0.79 C to 0.48 C, and that of local skin temperature changed from 2.11 C to 1.46 C. Further significant improvements were found when comparing the individualized Chinese model with the standard Chinese model. For the individualized model, the mean bias of mean skin temperature between prediction and measurement ranged from 0.20 C to 0.38 C, and the mean bias as well as its standard deviation of most local skin temperature was less than 1 C. Prediction accuracy was also validated in the extensive comparison with other researchers’ experiments on Chinese subjects. Prediction accuracy of Chinese adults’ skin temperature could be improved via the modification and individualization of thermoregulation model with Chinese physiological characteristics. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Thermoregulation modeling Skin temperature Chinese people Physiological parameter
1. Introduction Thermal comfort, associated with thermal sensation is one of the most important performance indicators for Heating, Ventilation and Air Conditioning (HVAC) systems. Currently thermal comfort criteria imbedded in standards ASHRAE Standard 55 and ISO Standard 7730 are based on the predicted mean vote (PMV) model [1], focusing on minimizing the percentage of people dissatisfied [2]. However, good agreement between PMV model and experiment was particularly found for uniform and steady-state environmental conditions [3]. On the other hand, thermal comfort may be achieved more energy-efficiently in non-uniform thermal environments. The distributions of skin temperature in such environments are different according to the region of human body, so developing a model to evaluate thermal comfort in asymmetrical environments or transient conditions has being a hotspot of recent study [4]. Furthermore, accurate models for predicting thermal comfort not only can be beneficial in avoiding malperformance in the use phase of a building [5], but also especially valuable for
* Corresponding author. Tel.: þ86 21 34204263; fax: þ86 21 34206814. E-mail address: [email protected] (Z. Lian). 0360-1323/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.buildenv.2013.08.031
predicting responses under conditions that cannot be tested ethically using human volunteers [6]. Thermal comfort model consist of physiological (thermoregulation in human body) and psychological model. Thermoregulation model, for the purpose of calculating skin temperature, is the first step of developing thermal comfort model. For practical utilization in design and evaluation of building environment, the calculated results of the model must agree well with the experimental measurements on human body. However, the thermoregulatory responses of different individuals are different, which makes the comparison impossible without a methodology to express individual difference in the model itself [7]. There were two approaches to deal with the problem of individual differences in thermoregulation model; one is from the passive systems of the body [8], such as thermal capacitance, thermal resistance, or surface area related to heat transfer, and the other is from the controlling systems of the body, such as regulatory sweating or skin blood flow. A populationbased dynamic human thermoregulation model built by Havenith [9] expanded with control equations incorporating several individual characteristics such as body surface area, mass, and body fat percentage. Although the prediction of individual heat strain had been improved by the model, a substantial part of differences in individual responses remain unexplained. Zhang et al. [10]
258
X. Zhou et al. / Building and Environment 70 (2013) 257e265
developed a model called “body builder” to translate descriptive data of an individual into a set of physiological parameters, which were then incorporated into a human thermoregulation model. Their simulated result only showed that the normal body fat group had lower skin temperature but higher rectal temperature than the lean group, yet a future work including more validations with experiments needs to be done. Study carried out by van Marken Lichtenbelt et al. [11] shows that on a group level predictions of skin temperatures can be improved when adopting individualized body characteristics and measured metabolic rate, but the predictions on an individual level were not improved. Fiala model [12] was individualized by Lichtenbelt et al. with respect to anthropometrics, body fat, and metabolic rate using the physiological data of Dutchman [13]; predictions of the individualized model were improved but a significant error remained. Furthermore this adapted model was used to predict thermal sensation and comfort in building environment [5]. Individual difference has been considered in human thermoregulation models, but ethnic difference which is closely related to the establishment of domestic building standard has received few concerns, although study indicates there are significant differences in body size between western and eastern people [14]. In China building energy consumption that largely depends on the operation of HVAC system is increasing rapidly [15,16], to balance energy consuming and thermal comfort of Chinese people, a precise model capable to predict thermal response of Chinese people is in urgent need. Thus in this study, a thermoregulation model was established for Chinese people and then individualized with four descriptive parameters of human body: height, weight, sex and age. 2. Establishment of Chinese model 2.1. Standard Chinese model Fiala thermoregulation model [12], which has been proved to be reasonable accurate [17,18], was adopted in this study after
extensive evaluation. Fiala model consists of two interacting systems: the controlled passive system and the controlling active system. In the passive system, as shown in Fig. 1 [12], the human body was subdivided into 14 segments, using cylinders stand for the trunk and extremities and a cylinder combined with sphere to represent for the face and head. Most of the body elements were divided into four layers (core, muscle, fat and skin from the inside to outside) with three sectors (anterior, posterior and inferior). Convection, radiation (long-wave and short-wave) and evaporation were considered in the heat exchange between human and environment, and heat changes via respiration was also included [20]. In the active system, central nervous system (CNS) accounts for overall changes in muscle metabolism by shivering, skin blood flow by vasodilatation and vasoconstriction, and skin moisture excretion by sweating. Local autonomic regulation was employed to modify local sweat rates, local blood flows, and tissue metabolic rates [21]. Former studies indicated that major physiological difference between ethnicity and individuals lies in the passive system [11,22]. A standard Chinese model was established based on the Fiala model with modifications in body size and composition. Human dimension of Chinese adults specified in the national standard of China [19] were employed for the establishment of the standard Chinese model. Considering the difference on body size between sex, independent model for man and woman were developed. Skin surface area (Ask) of standard Chinese model was derived from former experiment researches [23,24]. Thermophysiological parameters such as body fat percent (BF%) value for the standard Chinese model were obtained from the research of Deurenberg [14], basal metabolic rate (BMR) value from the research of Liu [25] and cardiac output (CO) was from the physiological research on Chinese [26]. Comparison on human dimension and thermophysiological parameters between Chinese and Fiala model were shown in Table 1. The weight of standard Chinese model was 19.1% lower than standard Fiala model, which leads to relatively lower value in Ask and BMR of a standard Chinese man. Body size of Chinese woman was even smaller than that specified in the
Fig. 1. Schematic diagram of the passive system in Fiala model.
X. Zhou et al. / Building and Environment 70 (2013) 257e265
259
Table 1 Comparison of basal body parameter between standard Fiala and Chinese model. Standard model
Height (m)
Weight (kg)
BF%
Ask(m2)
BMR (W)
CO (L/min)
Age
Fiala (man) Chinese man (difference) Chinese woman (difference)
1.71 1.70 (0.6%) 1.60 (6.4%)
73.5 59.5 (19.1%) 51.5 (29.9%)
14.4 16.8 (16.7%) 25.2 (75.0%)
1.86 1.71 (8.1%) 1.55 (16.7%)
87.1 80.1 (8.0%) 64.3 (26.2%)
4.9 4.8 (2.0%) 4.3 (12.2%)
/ 30/ 30/
standard Fiala model, weight, Ask and BMR of standard Chinese woman was about 30% lower than that of standard Fiala model. The most significant difference between sexes lies in body fat; BF% value of standard Chinese woman was 75% larger than that of standard Fiala model. CO value of Chinese man model was nearly the same as Fiala model, but for Chinese woman it was about 10% lower. As age is an important factor for the evaluation of thermal sensation [27], the effect of age on body composition and BMR was introduced. Chinese adults were divided into three age groups [28], i.e., 18e25 years, 26e35 years and 36e60 (for male) 55 (for female), and age of 30 years as the mean value of middle age group was chosen for the standard Chinese model. Proportional change in human dimension of the standard Fiala model was not proper for the establishment of Chinese model. As we can see in Fig. 2, the difference between models was not only in the overall skin surface area, but also in the skin surface area distribution of segments mainly the thorax, hand, upper leg and foot. Moreover, the distribution of body fat in Chinese adults was significantly different from the western people according to a former physiological research [29]. Considering the difference in skin surface area and body fat distribution between Chinese and western people, the passive system was rebuilt according to the physiological data of Chinese people (Table 2). Other thermophysiological parameters such as heat exchange coefficients, heat conductivity and capacitance, also the physical parameters of tissue materials like density were kept to be the same as the standard Fiala model [12]. 2.2. Model individualization After extensive comparison and evaluation, four easily obtained body characteristics including height, weight, sex and age were
selected as the input parameters for the individualization of Chinese model. Other parameters like BF%, one of the most important thermophysiological differences between individuals, was usually considered as an independent input parameter [9,10,13]. However, considering the practical utilization in current research, BF% and BMR value, which can be predicated by height and weight of human in a subgroup of people [14], was not chosen as an input parameter. Fig. 3 shows the flow chart of model individualization with the four input parameters; the left side shows how to calculate CO and BMR and the right side demonstrates the individualization of passive system. BMR was calculated by the formula described by Liu et al. [25].
BMR ¼ 58 W þ 1741 H 14 Age 470 Sex þ 227
(1)
CO was calculated by equation (2) [26].
CO ¼ 0:024 W 0:057 Age 0:305 Sex þ 4:544
(2)
Where H is the height (m) and W is the weight (kg) of individuals. Here Sex was coded as a dummy variable (man is 0 and woman is 1). BMR value in original formula was in kilojoules (KJ) per day, thus a unit transfer (changed to W by adding 3.6*24) was applied. Basal metabolic rate of a certain segment (BMRi) was calculated by summing up the heat production of all materials (mainly muscle, brain and viscera) in the segment. Materials’ heat production per unit volume was initialized with the value given by Fiala model (e.g. for muscle it is 684 W/m3), and then they were changed with the ratio of the calculated BMR of individualized model to the BMR of the standard Chinese model. Blood perfusion (BP) in materials was changed with the calculated CO value in a similar way as BMRi with calculated BMR.
Fig. 2. Difference between the distribution of Ask between standard Chinese model and Fiala model.
260
X. Zhou et al. / Building and Environment 70 (2013) 257e265
Table 2 Geometric parameters of the standard Chinese man model. Body segments L (m)a r (m) Core Head Face Neck Shoulder Thorax Abdomen Upper arm Lower arm Hand Upper Leg Lower leg Foot a
0.120 0.081 0.309 0.297 0.524 0.595 0.478 0.365 0.705 0.749 0.464
Muscle Fat
0.067 0.073 0.040 0.052 0.068 0.020 0.044 0.021 0.024 0.070 0.076 0.123 0.071 0.076 0.101 0.015 0.028 0.013 0.026 0.011 0.029 0.018 0.058 0.016 0.042 0.020 0.031
0.081 0.077 0.051 0.035 0.128 0.115 0.036 0.034 0.035 0.063 0.046 0.035
Inner skin Outer skin 0.0830 0.0775 0.0515 0.0355 0.1285 0.1150 0.0365 0.0345 0.0355 0.0640 0.0465 0.0355
0.084 0.078 0.052 0.036 0.129 0.116 0.037 0.035 0.036 0.065 0.047 0.037
BF% ¼ 1:38 BMI þ 0:25 Age 12:1 Sex 8:1
Here L is the length of segment; r is the radius of segment.
Anthropometrical parameters such as length of segments (Lseg) and Ask, were calculated from three of the input parameters [30].
Lseg ¼ B0 þ B1 H þ B2 W
(3)
Where Lseg is the length of body segment, B0, B1, B2 is the regression coefficients, most R values of the regression formulas listed in Table 3 were over 0.99 when compared to the measurement’s data [19]. The length of segment was changed in a separately different ratio with the height and weight, and a more precise solution for the building of passive system than proportional change was achieved with these regression formulas. Ask can be calculated by the following formulas: For man
Ask ¼ 0:00607 H þ 0:0127 W 0:0698
(4.1)
and for woman
Ask ¼ 0:00586 H þ 0:0126 W 0:0461 AskðiÞ ¼ Ask fa
segments were simplified to cylinders (head was sphere), with the length and surface area of a segment (Lseg and Ask(i)), then the outer radius of a segment (also is the outer radius of skin) can be calculated, as shown in Fig. 3. Typically there are four layers skin, fat, muscle and core from the outside to the inside in a segment, and the length of body layers are the same as the segment. After the outer radius of skin were calculated by Lseg and Ask(i), inner radius of skin which is also the outer radius of fat can be obtained as the thickness of the skin layer (1e3 mm) of all body parts was kept unchanged. Radius of layer other than skin can be calculated by the weight distribution and density of materials, such as the thickness of fat layer can be draw from the BF%. BF% value was calculated according to the method described in Deurenberg et al. [14]:
(4.2)
(6)
Where BMI was body mass index, equals to weight divided by square of height; Sex here was coded as a dummy variable (man is 1 woman is 0). For a certain segment, volume of fat layer was calculated by the BF% value with fat distribution coefficients (fb, see in Table 3) and the density of fat; then together with the inner radius of skin layer (also the outer radius of fat layer), the inner radius of fat (rfat) which was also the outer radius of muscle can be calculated. According to physiological measurements [28,31], weight of core layers (Wcore) including bones, brain (in head), lung (in thorax) and viscera (in abdomen) were given, and also weight distribution coefficient of core layer (fc) for man and woman were listed in Table 3. With Wcore, fc and the density of core materials, the core radius (rcore) for a standard Chinese were calculated, and rcore for an individual was changed with height in a linear relationship with a correlation coefficient of 0.06. Radius of muscle (rmuscle) was determined by the thickness of fat and core, since the outer radius of muscle was the inner radius of fat layer and the inner radius of muscle was the same as rcore. With the length of segments and radius of all material layers, the passive system for Chinese individual was established.
(5) 2.3. Sensitivity analysis
Where fa is the distribution factor for segment (see in Table 3) derived from experimental measurement [23,24]. Since the body
Fig. 3. Flow chart of Chinese model individualization.
Sensitivity analysis was performed to evaluate how body temperatures were affected by the four input parameters employed in the individualized Chinese model. A typical mildly cold environment was set up for this analysis, in which an equal ambient and radiant temperature of 20 C, a relative humidity of 50% and an air velocity of 0.05 m/s were maintained. Subjects kept sedentary (1 met) in the environment for 60 min, wearing clothes of 0.6 clo. Body height and weight of 5th and 95th percentile of Chinese adults were chosen according to the national standard of China [19], and percentiles of age was defined as 5% (20 years old), 50% (30 years old) and 95% (60 for male and 55 for female). When one of the parameters was changed, the other three input parameters were kept to be the same value as the standard Chinese model. Skin and rectal temperature changes between individuals were demonstrated in Fig. 4. Compared with the standard model (50th percentile), the mean skin temperature (Tskm) of man with the same weight, age but differ in height have a difference of 1.2 C. Rectal temperature (Trectal) and local skin temperature of a shorter man was relative lower than the standard man; on the contrary, the man of 95th percentile in height has a relatively higher value both in skin and rectal temperature. A taller adult, who has the same weight and age with the short one, would have a relative larger value in Ask but lower in BF%. Since the body fat with lower in
X. Zhou et al. / Building and Environment 70 (2013) 257e265
261
Table 3 Regression parameters and distribution coefficients for the individualization of Chinese model. Body segments
Man
Woman
Length
Head Face Neck Shoulder Thorax Abdomen Upper arm Lower arm Hand Upper Leg Lower leg Foot
Ask
(i)
BF(i)
Wcore
Length
Ask
(i)
BF(i)
Wcore
B0
B1
B2
R
fa
fb
fc
B0
B1
B2
R
fa
fb
fc
0.00569 0.00799 0.00542 0.03041 0.02934 0.06005 0.07866 0.1631 0.01231 0.25105 0.30984 0.00155
0.043 0.089 0.060 0.152 0.145 0.246 0.362 0.354 0.186 0.548 0.598 0.252
0.00026 0.00038 0.00026 0.00033 0.00035 0.00076 0.00097 0.00066 0.00062 0.00041 0.00071 0.00058
0.968 0.968 0.968 0.999 0.999 0.999 0.999 0.999 0.999 0.999 0.999 0.999
4.05% 2.06% 1.48% 1.61% 13.63% 21.17% 8.09% 6.41% 4.93% 16.33% 13.29% 6.95%
5.0% 4.0% 1.0% 1.0% 9.0% 42.5% 8.5% 5.5% 3.0% 11.3% 6.6% 2.6%
8.0% 0.5% 0.4% 7.1% 12.3% 23.6% 7.5% 4.5% 2.5% 12.6% 10.6% 10.4%
0.00569 0.00860 0.00542 0.03041 0.00593 0.01214 0.19539 0.22079 0.01224 0.41748 0.40474 0.02318
0.043 0.086 0.060 0.132 0.127 0.280 0.468 0.394 0.206 0.769 0.680 0.288
0.00026 0.00038 0.00026 0.00031 0.00016 0.00032 0.00048 0.00051 0.00057 0.00026 0.00037 0.00047
0.967 0.967 0.967 0.994 0.994 0.994 0.993 0.994 0.995 0.993 0.993 0.994
4.07% 2.26% 1.05% 2.94% 8.21% 22.31% 8.29% 5.71% 4.52% 21.16% 12.83% 6.65%
4.0% 1.5% 1.0% 1.0% 11.0% 44.0% 8.0% 5.0% 3.0% 13.5% 6.0% 2.0%
7.3% 7.8% 0.4% 5.6% 10.6% 24.8% 6.4% 3.7% 2.0% 13.5% 9.1% 8.9%
heat conductivity (38% as to muscle) plays an important role in preventing heat conduction through human body, thus heat conduction value of a taller person from the core layer was higher and eventually result in a relatively higher skin temperature. A lager BMR value was another reason for the higher rectal temperature of the taller person, which also made contribution to the higher value of skin temperature. Mean and rectal temperature also increased with the height of woman model. Differ in weight had another effect on skin and rectal temperature. A lean person was with a relatively higher value in skin temperature but lower in rectal temperature, which agreed well with the physiological measurements of skin temperature on overweight and lean man [10,32]. Lower in BF% and BMR could explain the difference on body temperature of lean people. A lean person with less body fat cannot resist heat transfer from the core layer as effectively as normal person, and lower BMR value which means less heat generation in human body leads to a relatively lower rectal temperature. Compared to height and weight, differ in age caused less difference in skin temperature, but the rectal temperature decreases with the age was in a considerable gradient, which was mainly caused by the negative correlation between BMR value and age [27]. Compared with the man model of the same percentile, relatively lower in mean skin temperature but higher in rectal temperature was found for woman model. Calculation results of woman model also agreed well with the experimental measurement [33], which can be explained by the lower BMR and CO value together with the higher BF% value of woman.
Fig. 4. Effect of input parameters on mean skin temperature (Tskm) and rectal temperature (Trectal).
3. Experimental validation 3.1. Experiment protocol The established models were validated with experiment research on Chinese adults. Twenty two healthy non-smokers (11 men and 11 women) from local university participated in this experiment. Fig. 5 shows the height and weight distribution of the subjects, which were located in the range of 5% and 95% percentile of Chinese adults, and most of them were fitted well with the regression curve [19]. Thus the subjects recruited were qualified to stand for typical Chinese adults. Subjects were asked to avoid caffeine, alcohol, smoking, and intense physical activity at least 12 h prior to each experimental session. Summer clothing (vest and shorts) with a total resistance of about 0.3 clo was compulsory for every subject. All experiment protocols were approved by the university’s ethics committee and conformed to the guidelines contained within the Declaration of Helsinki. The whole experiment consists of two parts: “temperature-rise” phase and “temperature-fall” phase. Both phases followed the same schedule which was demonstrated in Fig. 6. After 30 min sitting and resting in a thermal-neutral environment as a preparation period, the subjects were then divided into two groups and in each group a different combination of ambient temperatures was designed. Group 1 (6 men and 5 women, G1) was exposed to an indoor
Fig. 5. Height and weight distribution of subjects.
262
X. Zhou et al. / Building and Environment 70 (2013) 257e265
Fig. 6. Schedule of the validation experiment.
environment with the temperature order of 21 C, 24 C, 26 C and 29 C (from S1 to S4), for group 2 (6 men and 5 women, G2) the order from S1 to S4 was 29 C, 26 C, 24 C and 21 C. In this study, the measurements were made after that the subject had stayed in a steady thermal environment for at least 40 min, which was proved to be a steady state for mean skin temperature according to the early study [34]. The experiment was performed in a climate chamber during May. There was only a window in the climate chamber and no direct solar radiation entered. The air temperature was controlled using a wall-mounted air conditioner. The indoor air velocity near the subjects was kept under 0.05 m/s, and the relative humidity of air was not dependently controlled but recorded by a hygrometer. In order to obtain the stable skin temperature in the environment, we tested the skin temperature for 5 min before the end of every stage. Under each ambient temperature, a subject’s local skin temperatures at 7 points were measured according to the research of Hardy et al. [35]. 3.2. Validation with the experiment When calculating the skin temperatures for each subject with the individualized model, the simulation follows the same schedule and environmental set up as the experiments. Table 5 shows the mean difference with standard deviation (SD) between predicted by thermoregulation model and experiment measurement. In order to avoid difference may be covered in the average calculation process, absolute value of the skin temperature difference between experiment and simulation was employed. In the temperature-rise phase, the mean difference between standard Fiala model and experiment in mean skin temperature was about 0.5 C in all stages. The prediction of head temperature (Thead anterior) was the most accurate; however, in other parts of human body like lower arm, hand back and foot back, the mean difference was exceed 1.5 C in most conditions, also with a relative higher SD value (exceeding 1 C in most conditions). Mean difference of standard Chinese model was lower than standard Fiala model in most conditions, paired sample T test results showed that there were significant improvements in both mean and local skin temperature, especially for the skin temperature of thorax, hand and foot. The difference mainly derived from the difference in body composition and skin surface area between Chinese and Fiala model. Although mean difference has been improved by
modification of the body parameter, standard deviation value of the prediction remained larger than 1 C in some conditions. Compared with standard Chinese model, prediction accuracy of individualized Chinese model was significantly improved, both in mean value and standard deviation. Mean difference of mean skin temperature was between 0.20 and 0.33 C, and the standard deviation was less than 0.37 C. Paired sample T test also showed a significant difference between the standard and individualized Chinese model. In temperature-fall phase, significant differences also can be found in mean and local skin temperature between standard Fiala model and Chinese model. The skin temperature difference of Fiala model in the last stage was increasing rapidly, while the prediction of Chinese model kept nearly the same accuracy in all the stage of experiments. Significant differences in skin temperature were also found between individualized Chinese model and standard Chinese model. In most local skin temperature evaluations, the differences between models were more significant in the last stage of experiment when the ambient temperature was 29 C. In both temperature-rise and temperature-fall experiments, the prediction of Chinese model kept in a relative stable level, whereas the mean difference of Fiala model in most conditions ascended with the experiment time. 4. Extensive validations The individualized Chinese model was also validated with the experiment data performed by other researchers on Chinese subjects. Simulations were carried out with the same ambient condition as well as the subject’s activity level and clothing with the experiments. As the detail information of each subject was often not provided by other researchers, but only with the mean value of height, weight, and age of the subjects. Thus mean values of the subjects’ parameters were employed in the validation of individualized Chinese model. Some experiments did not report subject’s information and thereby standard Chinese model was applied. Validation with a typical experiment was shown as follows [36]. After preparation and acclimation for 30 min, the subjects were exposed to an indoor environment with certain ambient temperature for 60 min, the ambient temperature ranges from 20 C to 34 C (cold to warm for subjects), and relative humidity was recorded and applied into the simulation. 15 healthy male Chinese subjects were recruited for the experiments study, every subject was sitting on the chairs and kept at rest (1 met) dressing in a summer cloth (0.6 clo). Fig. 7 shows the results of mean skin temperature and rectal temperature in experiments. Simulation results of mean skin temperature agreed well with the experiments under all of the ambient temperatures, and most of the rectal temperature value which increasing very slowly with the ambient temperature was also well agreed with the experiment results. The comparison of local skin temperature was demonstrated in Fig. 8, most of differences between the simulation and mean value of experiment results were in the range of 1 C. Well agreement
Table 4 Subjects’ information and environmental conditions and of experiments for extensive validation. Subjects’ information
Environmental conditions
Reference
Number
Age (year)
Height (m)
Weight (kg)
Thermal resistance (clo)
Activity level (met)
Ambient temperature ( C)
Relatively humidity (%)
Background air velocity (m/s)
30
20.1
1.67
56.9
0.6
1
24 16
23.2 /
1.68 /
60.2 /
0.6 0.6
1 1
40
/
/
/
0.6
1
20, 29, 28 20, 25, 30 20,
50 70 50 80 30 50 50
0.05 0.05 0.05 0.05 0.05 0.05 0.05
23, 29, 32 32 30 30 25, 28
[37] [38] [39]
[40]
Table 5 Mean difference ( C) standard deviation between model and measurements for mean and local skin temperature in different ambient temperature (AT).
Temperature-rise phase
At ( C)
Tskm
Standard Fiala
21 24 26 29 21 24 26 29 21 24 26 29 29 26 24 21 29 26 24 21 29 26 24 21
0.45 0.42 0.57 0.60 0.40 0.42 0.45 0.48 0.23 0.20 0.31 0.33 0.51 0.70 0.60 0.79 0.43 0.48 0.43 0.43 0.38 0.35 0.34 0.33
Standard Chinese
Individualized Chinese
Temperature-fall phase
Standard Fiala
Standard Chinese
Individualized Chinese
Thead
0.62 0.58 0.70 0.56 0.52: 0.45 0.57: 0.44:: 0.37* 0.22* 0.30** 0.20** 0.48 0.48 0.57 0.68 0.49 0.56:: 0.56::: 0.60:: 0.40** 0.37** 0.44** 0.32*
0.51 0.43 0.50 0.31 0.52 0.41 0.48 0.29 0.56 0.37 0.41 0.39 0.67 0.68 0.32 0.55 0.69 0.41 0.33 0.61 0.63 0.37 0.49 0.33
anterior
0.54 0.27 0.33 0.46 0.52 0.25 0.30 0.46 0.27* 0.23** 0.26 0.41 1.06 0.50 0.37 0.67 1.06 0.45:: 0.39 0.71::: 0.58** 0.44** 0.38*** 0.42***
Tthorax 0.77 0.98 0.96 0.68 0.71 0.87 0.64 0.54 0.72 0.79 0.78 0.57 0.88 1.68 1.37 1.68 0.92 1.44 1.21 0.98 0.77 1.00 0.92 0.80
anterior
0.85 0.93 0.85 0.66 0.89 0.87:: 0.79:: 0.61::: 0.47 0.63** 0.60** 0.39** 1.00 1.27 1.24 1.16 1.07 1.31::: 1.26::: 0.74::: 0.72** 0.93** 0.93*** 0.88***
Tlower 1.64 1.53 1.84 1.92 1.20 1.24 1.43 1.50 1.11 1.01 1.07 1.08 0.89 0.55 0.69 1.15 0.65 0.71 0.69 1.08 0.52 0.57 0.76 0.69
arm anterior
1.52 1.85 2.11 2.39 1.24:: 1.58 1.83 2.14: 0.73 1.01 0.84* 0.84** 0.67 0.74 0.96 1.33 0.64:: 0.73 0.99 1.03 0.47*** 0.59* 0.68 0.94
Thand 1.91 1.85 1.80 1.03 1.46 1.41 1.41 0.85 1.21 0.43 0.57 0.80 1.06 1.19 1.27 1.89 0.59 0.63 0.94 1.12 0.59 0.84 0.76 0.94
back
1.65 1.98 2.04 1.01 1.48:: 1.81:: 1.90::: 1.04::: 0.87 0.84** 1.18 0.59** 1.30 0.99 1.38 2.22 0.85::: 0.85::: 1.32::: 1.66::: 0.71** 0.62* 0.80* 1.12
:
Tupper 1.23 0.88 0.89 0.92 1.17 0.81 0.75 0.67 0.88 0.80 0.79 0.78 0.93 1.50 1.10 1.17 1.07 1.38 0.89 0.89 0.34 0.63 0.72 0.53
leg anterior
1.81 1.19 1.05 0.85 1.73 1.09 0.89 0.79:: 0.77* 0.55 0.39* 0.66* 0.71 1.18 0.88 1.20 0.70 1.14 0.84:: 0.89:: 0.62*** 0.89* 0.83* 0.76*
Tlower 0.98 0.77 0.89 0.99 0.68 0.78 0.85 0.72 0.87 0.88 0.70 0.61 0.49 0.60 0.54 0.98 0.48 0.60 0.56 0.56 0.46 0.71 0.69 0.78
leg anterior
1.17 0.97 0.99 0.66 1.16:: 0.93:: 0.97:: 0.56 0.61* 0.81* 0.62** 0.56** 0.59 0.83 0.89 1.12 0.61 0.80 0.86 0.96 0.44** 0.48* 0.60* 0.45
Tfoot
back
1.75 1.59 2.11 1.72 1.02 0.88 1.41 1.04 0.47 0.54 0.59 0.23 0.87 0.76 0.79 1.36 0.80 0.75 0.63 0.63 0.72 0.84 0.36 0.47
1.40 1.08 1.92 1.90 1.40: 1.09: 1.93: 1.90::: 0.56** 0.60** 0.75** 0.36*** 0.87 0.93 0.78 1.31 0.88 0.93:: 0.77::: 0.83::: 0.45** 0.51** 0.58** 0.56
P < 0.05 :: P < 0.01 and ::: P < 0.001 when standard Fiala model compared with standard Chinese model; *P < 0.05, **P < 0.01 and***P < 0.001 when standard Chinese model compared with Individualized Chinese model.
263
Fig. 7. Comparison of mean skin temperature and rectal temperature between experiment and simulation.
could be found in forehead, thorax, arm and leg, with difference no larger than 1 C in all of environmental conditions; but the difference of foot and hand in the mild cold and neutral environment exceeded 2 C. Extensive validations of individualized Chinese model with other researchers’ experiments were shown in Fig. 9. Environmental conditions and subjects’ information of experiments were shown in Table 4 [37e40]. Difference in mean skin temperature was mainly in the range of 0.5 C, and the difference between experiment and simulation decreased with the ambient temperature, especially when the ambient temperature exceeds 30 C. As the ambient temperature rises, differences in skin temperature between segments of human body were narrowed down, thus the prediction of skin temperature was improved. Local skin temperature value was divided into two parts: proximal (including head, thorax, abdomen and upper arms/legs), and distal (lower arms/legs, feet and hands). As the skin temperature in distal part was more easily affected by the ambient environment, and in proximal part was mainly controlled by the thermoregulation mechanism. Prediction results of the two parts were different. For skin temperature in proximal part, though the difference was mainly within the
Fig. 8. Comparison of local skin temperature between experiment and simulation.
X. Zhou et al. / Building and Environment 70 (2013) 257e265
Models
264
X. Zhou et al. / Building and Environment 70 (2013) 257e265
achieved. The four input parameters in this model may be not adequate to predict the body information of each individual Chinese adult very precisely, but in a reasonable range, and also more practical in design and evaluation of indoor environment. Moreover, this method of model individualization was not limited to Chinese, but also can be extensive used for other people based on the local physiological research and validated by human experiments. 5.2. Potential applications
Fig. 9. Comparison of mean skin temperature (Tskm), skin temperature of proximal (Tproximal) and distal (Tdistal) between experiment and simulation.
range of 1 C, most simulation results was larger than the experiment measurement, indicating a slight modification in controlling active system was needed. For distal part, the skin temperature of simulation was slightly lower than experiment (most differences lies in the range of 1 C), especially when the ambient temperature was larger than 30 C. Through extensive validations, the prediction accuracy of individualized Chinese model was proved to be adequate according to former research [41], the difference for mean skin temperature was mainly less than 0.5 C, and for local skin temperature was less than 1 C. 5. Discussion 5.1. Input parameters Accuracy in predication and practicability in use are the major principles for the selection of input individual parameters. There are a few studies which discussed the individualization with proper input parameters. Sensitive analysis of height, mass, BF% and resting metabolic rate (RMR) was involved in the research of Lichtenbelt [42]. In their analysis, when a parameter is changed to the 5th or 95th percentile value, all other parameters are kept at the average value, which are not practical for real people. For instance, two ordinary persons with different weight (60 kg and 80 kg) but same height (1.70 m), their BF% should not be the same. Also it is the same with the RMR. The four parameters (height, mass, BF% and RMR) have inherent relationship with each other which should not be considered as independent input parameters. Input parameters used in the individualization of UC Berkley model are the height, weight, gender, body fat (or dimensional information) and skin color [10]. Body fat was calculated through the measurement of skinfold caliper, which needs the measurement of circumference at levels of umbilicus, minimal abdominal width at approximately midway between xiphoid and umbilicus. Metabolic rate (MR) and body composition (BC) were selected as input parameters for the Fiala model individualization [11]. MR was calculated from the O2 consumption and CO2 production measured by indirect calorimetry, using a ventilated hood system, and the model achieved the best accuracy with the input of MR and BC. The involvement of BF or MR measurements in the former model may be too complex for practical prediction of skin temperature for individuals, and the individualization of thermoregulation model should consider more about practical application when an acceptable accuracy had been
The incorporation of thermoregulation model and thermal psychological model, which could provide a practical link between building environment and human comfort, is the next step of work on Chinese model. Combination of computation fluid dynamics (CFD), thermal thermoregulation model and thermal sensation model were carried out by many researchers recently [4,5]; these studies indicated that the thermoregulation models were capable for the prediction of thermal sensation based on skin temperature. Fig. 10 shows one possible application of this Chinese model for predicting thermal comfort based on experiment carried out by Yu [43] and Liu [44]. The subjects in Yu’s experiments were exposed in the cold environment with ambient temperature of 20 C (mildly cool), 16 C (cool) and 12 C (cold). A simulation was carried out with the same environmental condition and subjects’ information, and a good agreement has been found in mean skin temperature changes between simulation results and experiment investigation. Meanwhile, a comfort zone (Tskm between 32.72 C and 33.74 C) for Chinese adults in uniform environment was given by experiment research form Liu [44], then the calculated mean skin temperature could be used to roughly estimate whether Chinese people is comfortable, eventually this model can be used for the design and evaluation of uniform indoor environment, for guaranteeing the thermal comfort for Chinese adults. Through the modification of standard Fiala model and individualization of standard Chinese model, prediction accuracy was significantly improved in most conditions, but still left some discrepancies in local skin temperature prediction which remained unexplained. As our knowledge about how inter-individual difference affects the human thermoregulation system was still scant, further experiments to cover wider population and environmental conditions are necessary. Further refinements of the model may include the individual difference between thermal response,
Fig. 10. Potential application of the model on prediction of thermal comfort in uniform environment.
X. Zhou et al. / Building and Environment 70 (2013) 257e265
especially when accuracy prediction on an individual level or prediction in the clinic under special thermal conditions. 6. Conclusion 1. A standard Chinese model was established based on the anthropometric and physiological study on Chinese adults. This model was then individualized with four input parameters including height, weight age and sex. Sensitivity analysis revealed that the four input parameters could have great impacts on mean skin temperature (up to 1.2 C). 2. The standard Chinese model was tested with experimental data on Chinese people. Results indicated that compared with the standard Fiala model, the standard Chinese model significantly improved the prediction accuracy in mean and most local skin temperature. The maximum difference in mean skin temperature between the Fiala model prediction and experimental data was reduced from 0.79 C to 0.48 C by the standard Chinese model, and the maximum difference in local skin temperature decreased from 2.11 C to 1.46 C. 3. The prediction accuracy was further improved by the individualized Chinese model, which greatly decreased the standard deviations of prediction for individuals. For the individualized model, the mean difference of mean skin temperature between model prediction and experimental data ranged from 0.20 C to 0.38 C; the standard deviation of most predicted local skin temperature was less than 1 C, compared to 2.39 C when predicted by the standard Chinese model.
[14]
[15] [16] [17]
[18]
[19] [20]
[21]
[22]
[23] [24] [25] [26] [27] [28] [29]
Acknowledgments
[30]
This work is financially supported by the Key project and General Project of National Natural Science Foundation of China (No. 51238005 and 51108260).
[31]
References
[33]
[1] Fanger PO. Thermal comfort. Copenhagen: Danish Technical University; 1970. [2] van Marken Lichtenbelt WD, Kingma BR. Building and occupant energetics: a physiological hypothesis. Archit Sci Rev 2013:1e6. [3] Van Hoof J. Forty years of Fanger’s model of thermal comfort: comfort for all? Indoor Air 2008;18(3):182e201. [4] Cheng Y, Niu J, Gao N. Thermal comfort models: a review and numerical investigation. Build Environ 2012;47(0):13e22. [5] Schellen L, Loomans MGLC, Kingma BRM, de Wit MH, Frijns AJH, van Marken Lichtenbelt WD. The use of a thermophysiological model in the built environment to predict thermal sensation. Build Environ 2013;59:10e22. [6] Xu X, Tikuisis P, Gonzalez R, Giesbrecht G. Thermoregulatory model for prediction of long-term cold exposure. Comput Biol Med 2005;35(4):287e98. [7] Takada S, Kobayashi H, Matsushita T. Thermal model of human body fitted with individual characteristics of body temperature regulation. Build Environ 2009;44(3):463e70. [8] Gagge A. An effective temperature scale based on a simple model of human physiological regulatory response. ASHRAE Trans 1971;77:247e62. [9] Havenith G. Individualized model of human thermoregulation for the simulation of heat stress response. J Appl Physiol 2001;90(5):1943e54. [10] Zhang H, Huizenga C, Arens E, Yu T. Considering individual physiological differences in a human thermal model. J Therm Biol 2001;26(4e5):401e8. [11] van Marken Lichtenbelt WD, Frijns AJH, Fiala D, Janssen FEM, van Ooijen AMJ, van Steenhoven AA. Effect of individual characteristics on a mathematical model of human thermoregulation. J Therm Biol 2004;29(7e8):577e81. [12] Fiala D, Lomas KJ, Stohrer M. A computer model of human thermoregulation for a wide range of environmental conditions: the passive system. J Appl Physiol 1999;87(5):1957e72. [13] Marken Lichtenbelt WD, Frijns AJH, Ooijen MJ, Fiala D, Kester AM, Steenhoven AA. Validation of an individualised model of human thermo
[32]
[34] [35] [36]
[37]
[38]
[39] [40] [41] [42]
[43]
[44]
265
regulation for predicting responses to cold air. Int J Biometeorol 2006;51(3): 169e79. Deurenberg P, Ge K, Hautvast JGAJ, Wang J. Body mass index as predictor for body fat: comparison between Chinese and Dutch adult subjects. Asia Pac J Clin Nutr 1997;6(2):102e5. Pérez-Lombard L, Ortiz J, Pout C. A review on buildings energy consumption information. Energy Build 2008;40(3):394e8. Hong T. A close look at the China design standard for energy efficiency of public buildings. Energy Build 2009;41(4):426e35. Fiala D, Lomas KJ, Stohrer M. Computer prediction of human thermoregulatory and temperature responses to a wide range of environmental conditions. Int J Biometeorol 2001;45(3):143e59. Psikuta A, Fiala D, Laschewski G, Jendritzky G, Richards M, Blazejczyk K, et al. Validation of the Fiala multi-node thermophysiological model for UTCI application. Int J Biometeorol 2012;56(3):443e60. Human dimension of Chinese adults. Beijing: CNIS; 1988. Fiala D, Psikuta A, Jendritzky G, Paulke S, Nelson DA, van Marken Lichtenbelt WD, et al. Physiological modeling for technical, clinical and research applications. Front Biosci 2010;2:939e68. Fiala D, Havenith G, Brode P, Kampmann B, Jendritzky G. UTCI-Fiala multinode model of human heat transfer and temperature regulation. Int J Biometeorol 2012;56(3):429e41. van Marken Lichtenbelt WD, Westerterp-Plantenga MS, van Hoydonck P. Individual variation in the relation between body temperature and energy expenditure in response to elevated ambient temperature. Physiol Behav 2001;73(1e2):235e42. Zhao S, Liu Y, Yao J, Yang Z, Liang Y, Zhang S. The estimation of body surface area of adults Chinese females. Acta Nutrimenta Sinica 1987;9(3):200e7. Zhao S, Liu Y, Yao J, Gao S, Zhang S. The estimation of body surface area of adults Chinese males. Acta Nutrimenta Sinica 1984;6(2):87e95. Liu HY, Lu YF, Chen WJ. Predictive equations for basal metabolic rate in Chinese adults: a cross-validation study. J Am Diet Assoc 1995;95(12):1403e8. Zhen Y. Physiology. Beijing: Higer Education Press; 2010. Guergova S, Dufour A. Thermal sensitivity in the elderly: a review. Ageing Res Rev 2011;10(1):80e92. Inertial parameters of adult human body. Beijing: CNIS; 2004. Lu H, Fu X, Ma X, Wu Z, He W, Wang Z, et al. Relationships of percent body fat and percent trunk fat with bone mineral density among Chinese, black, and white subjects. Osteoporos Int 2011;22(12):3029e35. Zhang C, Lin D. Parameters research on the adult anthropomorphic phantom of china. Machinery 2009;36(7):1e3. Rusong C, Wang Z, Ouyang B. Reference values of skeleton mass for normal Chinese adults. Chin J Radiological Med Prot 1995;15(2):317e20. Claessens-van Ooijen AM, Westerterp KR, Wouters L, Schoffelen PF, Van Steenhoven AA, Van Marken Lichtenbelt WD. Heat production and body temperature during cooling and rewarming in overweight and lean men*. Obesity 2006;14(11):1914e20. van Ooijen AMJ, van Marken Lichtenbelt WD, Westerterp KR. Individual differences in body temperature and the relation to energy expenditure: the influence of mild cold. J Therm Biol 2001;26(4e5):455e9. Dear R, Ring J, Fanger P. Thermal sensations resulting from sudden ambient temperature changes. Indoor Air 1993;3(3):181e92. Hardy JD, Du Bois EF, Soderstrom G. The technic of measuring radiation and convection one figure. J Nutr 1938;15(5):461e75. Yu J. Studies on the effects of physiological acclimatization on thermal responses of people accustomed to different indoor thermal environments [Ph.D Thesis]. Shanghai: Donghua Unviersity; 2011. Chen H, Zhang Y, Wang J, Meng Q. Human thermal comfort research in hothumid area of China (1): experimental methods and results. HV&AC 2010;40(3):83e8. Wang Y, Lian Z, Lan L. The effect of turbulence intensity on local skin temperature and subjective responses to draft. Energy Build 2011;43(10): 2678e83. Peng Y, Lian Z, Pan L. Skin temperature measurement method. J Southeast Univ 2010;26(2):258e61. Sun Y. Experimental study on human thermal sensation based on skin temperature. Dalian: Dalian University of Technology; 2009. Huizenga C, Hui Z, Arens E. A model of human physiology and comfort for assessing complex thermal environments. Build Environ 2001;36(6):691e9. van Marken Lichtenbelt WD, Schrauwen P, van de Kerckhove S, WesterterpPlantenga MS. Individual variation in body temperature and energy expenditure in response to mild cold. Am J Physiol Endocrinol Metab 2002;282(5): E1077e83. Yu J, Ouyang Q, Zhu Y, Shen H, Cao G, Cui W. A comparison of the thermal adaptability of people accustomed to air-conditioned environments and naturally ventilated environments. Indoor Air 2012;22(2):110e8. Liu W. Study on objective evaluation index of human thermal comfort [Ph.D Thesis]. Shanghai: Shanghai Jiao Tong University; 2008.