Correction of the evaporative resistance of clothing by the temperature of skin fabric on a sweating and walking thermal manikin

Abstract

Determining the evaporative resistance of clothing by using a sweating thermal manikin requires the accurate skin temperature of the manikin. The skin temperature measured by an embedded wire sensor (EWS), which is widely used in existing manikins, is theoretically higher than that of the wet skin fabric on the manikin where water evaporates. Therefore, we directly measured the surface temperature of the skin fabric using multiple thermistors (MT) and compared it with that of EWS. Four different work clothing ensembles were tested in an isothermal condition. The mean skin temperatures of MT were lower than those of EWS by 0.49, 0.62, 0.75, and 0.89°C, for the manikin walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s, respectively. Concomitantly, the real evaporative resistances using MT were lower than those using EWS by 9.1, 10.9, 12.5, and 14.4%. These results clearly show that the skin temperature of EWS should be corrected. In our thermal manikin, the temperature difference (TD) (°C) between MT and EWS was calculated for each of five parts: Arm, Trunk, Hip, Thigh, and Calf. The area-weighted average TD of the five parts was expressed as follows: TD = Heat Flux (W/m²) × 0.0092. This equation provides an estimate of TD without measuring surface skin temperature directly and helps to correct the evaporative resistance of clothing ensembles.

Keywords

evaporative resistance skin temperature sweating thermal manikin

In hot working environments, evaporation is the main avenue for heat to dissipate from the human body to the environment.¹ It is well known that work clothing ensembles, such as protective clothing, prevent evaporation of sweat from human skin to the environment and decrease heat dissipation from the body, leading to elevated skin temperature, core temperature or sweat rate.² In previous studies, the relative humidity and the temperature in the microenvironment between the skin and the clothing was found to increase sharply when wearing impermeable protective clothing.³ Another study showed that subjects wearing clothing with higher air permeability were able to walk at higher speeds in five types of protective clothing.⁴ Therefore, the correct evaporative resistance of clothing, which is defined as the division of the vapor pressure difference between the skin and the environment by the evaporative heat flux per unit area, is needed to predict the heat strain of workers in hot working environments. The maximum cooling power of evaporation can be derived from this evaporative resistance, as well as from the actual ambient water vapor pressure and the skin water vapor pressure. In fact, the predicted heat strain (PHS) model⁵ incorporates the i_m value (= total thermal insulation (I_t)/total evaporative resistance (R_et)/Lewis number) as one of the input parameters to predict the core temperature or sweat rate. As such, evaporative resistance is an important consideration of reducing heat strain and should be measured with considerations of both standing and walking. Since the effect of walking on evaporative resistance of clothing is very complicated,⁶ measurement using a movable sweating thermal manikin is appropriate for evaluation.

The development of sweating thermal manikins, which have the same size and shape as humans, has enabled the reproducible measurement of R_et or thermal insulation (I_t) of clothing ensembles.^7,8 However, correctly estimating R_et in sweating thermal manikins requires accurate measurement of the vapor pressure at the manikin skin, the environmental conditions, and the heat generated from within the manikin shell.⁹ Since the correct measurement of the vapor pressure is very difficult, it is estimated using the skin temperature, assuming that the relative humidity is 100% on the fabric skin. Recently, most sweating thermal manikins have used a shell-embedded wire temperature sensor (EWS) to measure the skin temperature. Wire temperature sensors wound evenly throughout the whole manikin zones are embedded in a strong and heat-conducting material. Using the known temperature dependence of the resistance of EWS, temperature is calculated from the measured resistance. Although EWS has many advantages in measuring the skin temperature, it has a serious disadvantage in that it does not measure the temperature of wet skin fabric on the manikin where water evaporates, which is required for the calculation of the R_et for clothing. When measuring the R_et of clothing with a sweating thermal manikin, a skin fabric is usually placed on the manikin to wet the entire surface of the manikin shell. Thus, some thermal resistances are assumed to be present between the EWS and the surface of the skin fabric. According to Fourier’s law, the combination of thermal resistances and heat flux from a manikin results in a lower temperature on the surface of the skin fabric than as measured by the EWS. This skin temperature decrease results in a lower saturated vapor pressure at the skin surface than is estimated based on the embedded sensors, and thus the actual vapor resistance is lower than that calculated from EWS.

A previous paper¹⁰ showed that there was a temperature drop of the wet fabric skin compared with the EWS. Skin temperature drop was measured in a standing condition for three different ambient conditions. There were, however, some shortcomings in this experiment. First, the small number of temperature sensors in the study would make it difficult to measure the mean skin temperature correctly. A complicated form of manikin would lead to varying heat dissipation from the skin, which would cause skin temperature differences. During walking in particular, the air ventilation would be higher in the arms or the legs than in the trunk, leading to the lower skin temperature in the arms or the legs. It would be impossible to measure the mean skin temperature with only a small number of sensors, due to the skin temperature variation. Therefore, in our study we have used multiple thermistors (MT) to measure the temperature of wet fabric. MT enable measurement of the mean skin temperature during walking although the skin temperature may not be uniform. A second shortcoming of the earlier study was that the wet skin temperature was measured under non-isothermal conditions. The maximum temperature difference between the ambient temperature and manikin was large in their study. When the temperature difference is large, an accurate measurement of skin temperature is very difficult. In contrast, in the present experiment, isothermal conditions were used. Third, the earlier study did not use a sweating manikin, and thus the content of water in the wet fabric decreased. To address this, we used a sweating thermal manikin, which was able to provide water from inside the manikin at the same temperature as the skin. In our study, we investigated the temperature difference between existing EWS and MT on wet fabric with a sweating thermal manikin during both standing and walking conditions.

The objective of this study was to more accurately estimate the evaporative resistance of clothing during walking as well as in standing conditions. This is the first study to measure the mean skin temperature on wet fabric skin directly by MT during walking conditions using a movable sweating thermal manikin. To achieve the objective, we measured the skin temperature directly by fixing MT on to the skin fabric covering the sweating movable manikin and compared readings to that of EWS. We tested four kinds of clothing ensembles which are frequently worn in the workplace.

Methods

Thermal manikin

A movable sweating thermal manikin (Newton, Measurement Technology Northwest, Seattle, USA) was used to measure both the mean skin temperature and heat flux in each zone of the manikin, and to measure the evaporative resistance of clothing. Our multi-segmented sweating thermal manikin Newton had 26 zones, each of which could control its heat flux, skin temperature (measured using an EWS) and sweat rate with software (ThermDAC, Measurement Technology Northwest, Seattle, USA) installed on a laptop computer (Dell, Texas, USA). The Newton had the average shape of an Asian, 1.65 m in height, composed of an aluminum-filled carbon-epoxy shell with embedded heating and a temperature wire sensor. The skin temperature wire sensor was embedded 0.5 mm under the surface of the manikin shell. When measuring the evaporative resistance of clothing, the Newton was covered with a thin sweating skin fabric (about 0.1 mm in thickness) to evenly wet the entire manikin shell. Due to the tightness and elasticity of the skin fabric, there was no space between the manikin shell and the skin fabric, except at the joints of different zones. During the experiment, the temperature of the EWS was fixed at 34°C via PID-controlled heating within each zone using ThermDAC. The generated heat data and temperature measured by the EWS were stored every minute on the laptop computer.

Climate chamber

Experiments were conducted in a climate-controlled chamber using the sweating thermal manikin. During the experiment, room temperature was set at 34.0 ± 0.1°C, and humidity was set at 50 ± 5%. The room temperature was measured at two points of the Newton: the chest and the knee levels. The mean temperature of the two points was used as the room temperature of the climate chamber. The air flowed from the ceiling to the floor at a speed of about 0.2 m/s, which kept the room at constant temperature.

Skin temperature measurement

To measure the skin temperature directly on the skin fabric, we attached four thermistors (LT-ST08-12, Gram Corp., Japan) per zone by stringing them tightly to the skin fabric. The cup-shaped thermistor had only a 5-mm diameter. The bottom of each thermistor was a flat metal conductor; the upper side was an insulator. The locations for the thermistor in each zone of the skin fabric were determined as follows. First, we put 15–25 thermistors in each zone proportional to the zone’s area. Elastic bands were used to fix the thermistor to the skin fabric. Three or four elastic bands were placed at equal intervals in each zone, and eight or ten thermistors were attached under an elastic band at equal intervals. Thus, the candidate locations for the four points were distributed evenly within each zone. Second, we registered the temperature measured by these thermistors in a nude and wet isothermal condition. These measured temperatures were stored using a data logger (LT-8, Gram Corp., Japan). Finally, four points were selected in each zone such that the average of these four points best approximated the average of all points within the zone, excluding the nearest points. After these four points for thermistor locations were decided for each zone, a total of 80 thermistors (four sensors in each zones × 20 zones) were sewed on the skin fabric with the bottom firmly in contact with the fabric (Figure 1). Zones of the face, the head, and both hands and feet were excluded.

Figure 1.

Fixation of the thermistor on the fabric skin (side view). The thermistors were sewed on to the skin fabric with the flat bottom firmly in contact with the skin fabric via a string. The flat bottom of the thermistor was made of a conductor metal. Conversly, the top of the thermistor was an insulator. The evaporated water was able to escape from the space of thermistor through the intervals between the strings.

Nude wet conditions

Before the measurement of the R_et of clothing ensembles, nude wet conditions of a sweating thermal manikin were investigated. To obtain a visual image and rough estimation of the surface temperature on a skin fabric in nude wet conditions, thermography (TVS-700, NEC AVIO, Japan) was used. The thermography measurement had high resolution in temperature (under 0.05°C), but the margin of error for the absolute temperature was ± 2.0°C. To compensate for this shortcoming, we used a water tank of 34.0°C as a reference temperature. The absolute temperature of the skin fabric was corrected using the temperature difference between the mean temperature of four thermistors attached to a black body sheet on the water tank and the mean temperature of the thermogram of the water tank:

(1)

where CST is the mean skin temperature of a wet skin fabric within a zone, STG is the mean skin temperature of a wet skin fabric within a zone measured by thermography, WTT is the mean skin temperature of the water tank measured by four thermistors, and WTG is the mean skin temperature of the water tank measured by thermography.

A thermogram was taken for each zone after the skin temperature reached a steady state in the nude and wet isothermal condition. As the left upper Arm, right upper Arm, left lower Arm, and right lower Arm zones were not divided into front and back sides, these zones were measured from both front and back and then averaged. The mean, maximum, and minimum skin temperature within each zone of the thermogram were measured with the analysis software (PE professional, NEC Avio, Japan). The evaporative resistance of the air layer was calculated using the temperature from the EWS and the mean, maximum, and minimum skin temperature in each zone, as measured by the thermistor or the thermogram.

Evaporative resistance of clothing ensembles

The mean skin temperature for four different types of work clothing ensembles (Table 1, Figure 2) were measured using either MT or an EWS under four conditions: manikin walking at 0.00 (standing), 0.27 (15 steps/min), 0.53 (30 steps/min), and 0.80 (45 steps/min) m/s. The heat flux, the skin temperature of the EWS, the ambient temperature and the relative humidity were recorded on a laptop computer every minute until each parameter reached a completely stable state and then kept constant for more than 30 min. Before and after each measurement, MT were checked to ensure that they were attached firmly to the wet skin fabric. The temperature data measured by MT were stored to data loggers (LT-8, Gram Corp, Japan). For each type of clothing, measurements were made more than three times. The hands of the manikin were covered with gloves; the feet, with socks and safety shoes. The mean skin temperature and the evaporative resistance of the clothing ensembles were calculated for each zone except for the face, hands, and feet by using the temperature from either the EWS or the mean temperature of the thermistors on the wet skin fabric.

Figure 2.

The four types of clothing ensembles that were tested. From the left: SBSE, Summer Business Suits + Underwear; WCE, Work Clothing + Underwear; JCWCE, Japanese Construction Worker Clothing + Underwear; TME, Tyvek@ SoftWear Model III + Underwear.

Table 1.

Material and weight of measured clothing ensembles

Clothing	Material	Weight(g)
Summer Business Suits	Jacket, trousers: 50% wool, 50% polyester	701
	Dress shirts: 50% polyester, 40% cotton, 10% twine	199
	Tie: 100% Silk	50
Work Clothing	80% polyester, 20% cotton	605
Japanese Construction	90% polyester, 10% cotton	1071
Worker Clothing
Tyvek @ SoftWear Model III (DuPont)	100% polyethylene	153
Underwear (Gunze)	Shirts: 60% cotton, 25% polyester, 15% acetate	90
Underwear (Gunze)	Pants: 80% polyester, 15% cupra, 5% polyurethane	44

The R_et of the clothing ensembles were calculated by the following parallel method given in Equation 4.

(2)

(3)

(4)

where i is the manikin zone number (i = 1–20), H_i is the heat generated at zone i (i = 1–20), A_i is the area of zone i (i = 1–20), T_{s_i} is the mean skin temperature of zone i (i = 1–20), A is the summation of the area of zone i (i = 1–20), T_s is the mean skin temperature of all zones, T_a is the ambient temperature, R_{t_i} is the thermal resistance of clothing of zone i (i = 1–20), R_et is the total evaporative resistance of clothing ensembles, P(T_s) is the saturated vapor pressure at T_s, and RH is the relative humidity. The value of H, T_s and R_et were calculated for walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s.

We divided the manikin shell into five parts: Arm, Trunk, Hip, Thigh, and Calf. Each part had four zones. The following equations were used to calculate the mean skin temperature and the evaporative resistance of each part.

(5)

(6)

(7)

where j is the manikin part number (j = 1–5), i is the manikin zone number in part j (i = 1–4), H_j is the heat generated at part j (j = 1–5), H_{j_i} is the heat generated at zone i in part j, A_j_{_i} is the area of zone i (i = 1–4) in part j (j = 1–5),

T_{s_{j_{i}}}

is the mean skin temperature of zone i (i = 1–4) in part j,

R_{t_{j_{i}}}

is the thermal resistance of clothing of zone i in part j (i = 1–4), T_{s_i} is the mean skin temperature of part j (j = 1–5), A_j is the area of part j (j = 1–5), T_a is the ambient temperature, and R_{et_j} is the evaporative resistance of clothing ensembles of part j (j = 1–5). The mean skin temperature and evaporative resistance of each part were calculated for walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s. The thermal resistance of four types of clothing in walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s are shown in Table 2.

Table 2.

Total thermal insulation (IT) of four clothing ensembles

Clothing Ensembles	Total Thermal Insulation (°C · m²/W)
Clothing Ensembles	Standing	Walking at 0.27 m/s	Walking at 0.53 m/s	Walking at 0.80 m/s
SBSE	0.302	0.261	0.245	0.226
WCE	0.277	0.244	0.227	0.210
JCWCE	0.271	0.236	0.214	0.198
TME	0.280	0.244	0.223	0.208

SBSE, Summer Business Suits + underwear; WCE, Work Clothing + underwear; JCWCE, Japanese Construction Worker Clothing + underwear; TME, Tyvek @ SoftWear Model III + underwear. Face, head, hands and feet were excluded in calculation of the thermal insulation of the clothing ensembles.

Results

The thermogram of the chest zone of the wet nude manikin in addition to the water tank showed that skin temperature differed widely, even in the same zone (Figure 3). The range of temperatures for the water tank was much smaller than that for the wet manikin skin. In Figure 4, we show the mean skin temperature as measured by the EWS, the MT, and by thermography for a wet nude manikin in an isothermal condition. The mean, maximum, and minimum temperatures of the four thermistors in each zone were used to calculate the area-weighted average skin temperatures for the manikin and are labeled (B), (C), and (D), respectively (Figure 4). The skin temperature and R_et were measured four times. The mean skin temperature as measured by the EWS (A) was higher than that measured by the MT (B) by 1.2°C (Figure 4). As a result, the calculated R_et of (B) was 24.1% lower than that calculated for (A). The mean skin temperature and R_et measured by MT(B) were close to those measured by thermography (E). When we used the maximum or minimum temperatures of the four thermistors in each zone to calculate whole mean skin temperature of the manikin, the mean temperatures were significantly higher or lower than when using the mean temperatures of the four thermistors (Figure 4).

Figure 3.

A thermogram of the chest zone of the wet nude manikin and the water tank. Manikin temperature (EWS) and room temperature were controlled at 34°C. Upper right figure shows the histogram of skin temperature distribution of the manikin chest zone. Lower right figure shows the histogram of skin temperature distribution of the water tank.

Figure 4.

Mean skin temperatures and total evaporative resistances (R_et) of the wet nude manikin using the EWS, thermistor, and thermography (n = 4). Manikin temperature (EWS) and room temperature were controlled at 34°C. For the thermistor, the mean, maximum and minimum temperatures of the four thermistors in each zone were used to calculate (B), (C), and (D), respectively. Similarly, (E), (F), and (G) were calculated using thermography data. The total evaporative resistance (R_et) of the air layer was calculated using the above temperatures.

For the work clothing ensemble, the mean skin temperatures measured by the MT were lower than those measured by the EWS for both the standing and walking conditions (Figure 5). As the skin temperature of the EWS was controlled at 34°C, the mean temperature of the shell near EWS was thought to remain at 34°C for all conditions. The decreases in mean skin temperature were 0.52, 0.63, 0.72, and 0.87°C for walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s. R_et calculated using the temperature of MT to EWS resulted in a lower value than that by EWS. R_et measured by MT were lower than those by EWS by 10.6, 12.0, 12.9 and 15.3% for walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s (Figure 5). The ratios of the evaporative resistance of each part using the MT to that using the EWS for the Tyvex Softwear Model III ensemble (TME) are shown in Figure 6. In the Arm, the ratios of evaporative resistance decreased more than in the other parts (Figure 6). In Trunk or Hip, the evaporative resistance did not decrease as much as the other parts.

Figure 5.

The mean skin temperatures and total evaporative resistances (R_et) of work clothing ensemble measured by multiple thermistor (MT) and embedded wire sensor (EWS) on wet skin fabric. (Error bars indicate the standard deviation of four measurements.).

Figure 6.

The ratio of evaporative resistance of Tyvex @ SoftWear Model III ensemble for five body parts by multiple thermistor (MT) compared with embedded wire sensor (EWS) while walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s.

For the four types of clothing ensembles, skin temperature was measured and compared for the MT and the EWS (Figure 7). For all types of clothing ensembles, the skin temperature measured by the MT decreased as the walking speed increased. The mean decreases in the ratios of R_et using MT compared with EWS were 9.1, 10.9, 12.5, and 14.4% for walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s, respectively (Figure 8). The decreasing ratios of R_et for the four types of clothing ensembles were very similar.

Figure 7.

The mean skin temperatures measured by the multiple thermistor (MT) or the embedded wire sensor (EWS) for four types of clothing ensembles while walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/sec. (Error bars indicate the standard deviation of four measurements). SBSE, Summer Business Suits + Underwear; WCE, Work Clothing + Underwear; JCWCE, Japanese Construction Worker's Clothing + Underwear; TME, Tyvex@ SoftWear Model III + Underwear; EWS, Embedded Wire Sensor.

Figure 8.

The ratio of total evaporative resistance (R_et) calculated using the temperature of the multiple thermistor (MT) to that calculated by the embedded wire sensor (EWS). SBSE, Summer Business Suits + Underwear; WCE, Work Clothing + Underwear; JCWCE, Japanese Construction Worker's Clothing + Underwear; TME, Tyvek@ SoftWear Model III + Underwear.

Discussion

Every thermal manikin except for Walter, a sweating fabric manikin,¹¹ uses EWS to measure the skin temperature of each zone. However, the skin temperature should be measured on the surface of the wet skin fabric covering the manikin where water evaporates, according to the definition of the equation for calculating the evaporative resistance given in Equation 4. Therefore, we measured the surface temperature of a wet skin fabric by fixing four thermistors to each of the 20 manikin zones and compared them with the temperature of EWS in this study. The average temperature measured by MT was lower than that by EWS. For four types of work clothing ensemble, the temperature using MT in comparison with EWS decreased by 0.49, 0.62, 0.74, and 0.89°C, for walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s, respectively.

Concomitantly, the evaporative resistances of MT decreased by 9.1, 10.9, 12.5, and 14.4% in comparison with those of EWS. To measure the mean skin temperature of the sweating thermal manikin, 80 small thermistors (four thermistors for each of 20 zones), were fixed on the wet skin fabric of the manikin. In nude wet conditions, the surface temperature in the chest zone varied from less than 29°C to more than 34°C from the thermogram (Figure 3). When work clothing ensembles covered the manikin, the average difference of maximum and minimum temperature in each zone was 1.2°C, which was about the same temperature difference in nude condition. These results indicated that the multiple point temperature measurement was required to measure the accurate mean skin temperature of each manikin zone. Multiple sensors helped to suppress the variation of the point skin temperature by averaging the temperature of many thermistors. During walking in particular, different rates of heat dissipation can be assumed between the trunk and the arms or the legs. However, no study has been found in the literature in which the wet skin surface temperature was measured using multiple temperature sensors in each zone.

Another factor causing the skin temperature variation came from skin fabric areas that did not contact with the manikin shell, especially at joint areas. The reason was that the skin fabric was made from one piece of cloth, although the manikin had a complex shape. Even in those places where the wet skin fabric did not touch the manikin, for example the joint area of each zone, water did evaporate from the wet skin fabric. Thus, Figure 3 shows that skin fabric temperatures near the joints were lower than in other places. In contrast to the lower temperatures of skin fabric at the joints, the parts of the manikin shell not covered by wet skin fabric were considered to have higher temperatures than elsewhere on the shell. As the mean temperature of the manikin shell was controlled at 34°C in this experiment, the heat flux from the heater would decrease due to the high temperature of the manikin shell not covered by wet skin fabric; the decreased heat flux led to a lower temperature for the wet skin fabric. From the thermogram of the whole wet nude manikin, we estimated that approximately 11% of the wet skin fabric did not touch the manikin shell. The average temperature difference between the skin fabric touching the manikin and that not touching was about 1.6°C. This decrease in temperature contributed to the decrease in the mean temperature of the wet skin fabric.

In this study, heat dissipation to the environment from the manikin was increased by faster walking. The main avenue of heat transfer was water evaporation from the wet skin fabric in this experiment, because convection and radiation were small due to the small temperature difference between the wet fabric skin and the ambient temperature. To compensate this heat transfer, convective heat transfer was calculated by using the temperature difference and the thermal insulation of clothing ensembles, and was subtracted from the heat transfer released from the manikin heater. The evaporated water was able to escape to the environment by ventilation, such as permeating through the clothing, passing through apertures of the clothing, and the pumping effects of walking.¹² Faster walking increases both ventilation and mass transfer between the ambient environment and the inner clothing.^13,14 This ventilation led to a decrease in the skin temperature on the wet skin fabric as the walking speed increased. In high-ventilation areas, such as Arm, Thigh, and Calf, the ratio of evaporative resistances calculated by MT compared with that by EWS decreased more during faster walking (Figure 6). In Figure 8, the ratio of R_et calculated using the MT to that using EWS (Equation 4) decreased linearly at the walking velocity of 0.8 m/s. If the ratio of R_et decreased linearly by the average walking speed for young men of 1.5 m/s, the decrease in ratio at 1.5 m/s would be about 20% (Figure 8). In actual walking, R_et would decrease further, since there will be an additional wind effect while walking due to the air movement around a person. These results indicated that in higher ventilation it is important to correct the measurement error of skin temperature from EWS.

The relationships between the heat flux and the mean skin temperature of five parts measured by MT are shown in Figure 9. All measurement data for the four clothing ensembles during walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s were included in the figures. Figure 9 shows that each measurement datum of a part was able to be regressed on a linear line. Namely, the decrease in temperature was almost proportional to the heat flux in every part (Figure 9).

Figure 9.

The relationship between the mean heat flux and the mean temperature difference between the multiple thermistor (MT) and the embedded wire sensor (EWS) for four types of clothing ensembles while walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/sec.

According to Fourier’s law, conduction of heat flux (Q) can be calculated¹⁵ as

(8)

where Q (W/m²) is the electrically generated heat in the manikin, T_shell (°C) is the temperature measured by the EWS, T_ws (°C) is the temperature measured by the MT on the wet skin fabric, δ₁ (m) is the thickness of the manikin shell on the EWS, δ₂ (m) is the thickness of the wet skin fabric, k_shell (W/mK) is the effective thermal conductivity of the manikin shell on the EWS, and k_ws (W/mK) is the effective thermal conductivity of the wet skin fabric.

Although Fourier’s law shown above is a simplified one-dimensional equation, it roughly describes the linear relationship between the temperature difference of the EWS and the MT on the wet skin fabric and the heat flux in our results (Figure 9). The slight deflection from the linear relationship could arise from k_ws, which was not constant.¹⁶ Since δ₁ and k_shell differ for other kinds of sweating thermal manikins and δ₂ and k_ws also differ for the other kinds of wet skin fabric, the coefficient of the slope for the linear regression should differ.

In Arm, the mean skin temperature decreased the most with increasing mean heat flux in five parts. This indicates that δ₁ was larger in Arm part than other parts from Equation 8. In Trunk and Hip, heat flux was not large. During walking, the air ventilation in Trunk and Hip inside the clothing ensemble would not be so large as in Arm, Thigh or Calf. The humidity of the air near the wet fabric was kept higher, which led to the lower evaporation rate and heat dissipation in Trunk and Hip. In Thigh or Calf, the mean heat flux increased more than the other parts during walking. The slopes of the regression line were very similar between Thigh and Calf. The relationships between the area-weighted mean skin temperature and the area-weighted mean heat flux of the five parts together are also displayed at the right bottom of Figure 9. Area-weighted average data also showed a linear regression relationship. In all figures, y-intersects of the regression line were at about 34.0°C, except for Hip. Since the temperature of EWS was controlled at 34.0°C, the temperature drop between the EWS and MT, which was 34.0°C minus the temperature of MT, was in proportion to the mean heat flux. However, the y-intersect of the regression line of Hip was significantly lower than 34.0°C. The lower temperature in Hip would be caused by the fact that a large area of the wet fabric in Hip did not come in contact with manikin shell due to the hip joint of the manikin during walking.

If the proportional relationship between the temperature drop and the heat flux can be assumed, the prediction equation of the temperature drop for each part can be written as follows using the data from our four kinds of clothing ensembles while walking at 0.00, 0.27, 0.53, 0.80 m/s.

(9)

(10)

(11)

(12)

(13)

(14)

These equations were statistically evaluated by comparing the root mean squared deviation (RMSD)¹⁷ and the mean standard deviation of the measured wet fabric temperature. The RMSD is expressed as the following equation:

(15)

where d_i is the difference between observed and predicted temperature for each measurement and n is the number of measurement. Table 3 shows the data of the mean standard deviation and RMSD of five parts for four clothing ensembles. In total parts and Calf, the RMSD was within double the value of the mean standard deviation for the four types of clothing ensembles. In Hip, all the RMSD were over two times larger than the mean SD, which showed that the predicted mean skin temperature of wet fabric in Hip was outside the 95% confidence interval. This means that the prediction equation of the skin temperature for Hip should not be applicable. This would be caused by the fact that the large area of Hip did not come into contact with the manikin due to the manikin’s form, leading to the lower skin temperature. Thus, for Hip, the original regression equation in Figure 9 should be used instead.

Table 3.

RMSD values of each part for four tested clothing ensembles

Parts	SBSE		WCE		JCWCE		TME
Parts	Mean SD	RMSD	Mean SD	RMSD	Mean SD	RMSD	Mean SD	RMSD
Arm	0.124	0.249	0.109	0.159	0.053	0.196	0.122	0.205
Trunk	0.077	0.163	0.103	0.193	0.127	0.156	0.071	0.077
Hip	0.077	0.219	0.054	0.442	0.078	0.328	0.111	0.261
Thigh	0.070	0.082	0.048	0.104	0.089	0.191	0.122	0.108
Calf	0.114	0.109	0.044	0.066	0.064	0.078	0.057	0.078
Total	0.039	0.070	0.049	0.090	0.048	0.092	0.062	0.077

RMSD, Root Mean Squared Deviation; SBSE, Summer Business Suits + underwear; WCE, Work Clothing + underwear; JCWCE, Japanese Construction Worker Clothing + underwear; TME, Tyvek @ SoftWear Model III + underwear. Face, head, hands and feet were excluded in calculation of the thermal insulation of the clothing ensembles.

For two clothing ensembles in Arm and Thigh, and for one clothing ensemble in Trunk, RMSD were over two times larger than the mean SD. One of the reasons would be that the temperature of the fabric skin differed within the parts. The air layer or the air permeability of the different clothing ensembles could bring differing ventilation inside the clothing during walking.

We compared these equations with previous research, which includes Havenith’s equation (16)¹⁰ and Wang’s equation (17).¹⁰

(16)

(17)

Figure 10 shows R_et of TME estimated by six kinds of mean skin temperature: EWS, MT, predictive temperature by each individual parts equation, by Equation (14), (16) and (17).

Figure 10.

Total evaporative resistance of Tyvex@ SoftWear Model III ensemble. Six kinds of total evaporative resistance of Tyvex@ SoftWear Model III ensemble were shown while walking at 0.00 (standing), 0.27 (15 steps/min), 0.53 (30 steps/min), 0.80 (45 steps/min) m/sec. From the left, R _et calculated by using the temperature of the embedded wire sensor, the temperature of multiple thermistors on wet fabric, predicted temperature by using our predictive equation of each part, predicted temperature by using our predictive equation of all parts, predicted temperature by using Havenith's equation and Wang's equation. EWS, Embedded Wire Sensor; MT, Multiple thermistors.

The widely used R_et measured by EWS was significantly larger than the others. R_et by MT was correctly predicted by our two equations. R_et by Equation 16 was very close to ours at standing and walking at low speed. One reason for this is that Havenith used the same type of manikin as used here. At higher walking speed, predicted R_et was a little lower than ours. The y-intersect of Equation 16 was a little larger than 34.0°C, rendering the slope of the equation a little steeper. This would cause a slightly lower R_et at higher walking speed. The predictions from Equation 16 and our equations were very close, but Equation (17) predicted a lower R_et than others.

There are several reasons why the results from this equation were different. First, the y-intersect was fixed at 34.00°C in calculating the regression line. When we recalculated the regression line using the data from Figure 3 in the paper by Wang et al.¹⁰ by removing the restriction of the y-intersect at 34.00°C, the y-intersect was 35.4°C and the slope was −0.023. 34.00°C was outside of the 95% confidence interval of the y-intersect. Statistically, it was impossible to write the regression line crossing the y-axis at 34.00°C from their data. Second, the skin temperature of the wet skin fabric was measured using a small number of temperature sensors. It would be difficult to measure the mean skin temperature of the zone. Third, their experimental conditions contained a large temperature difference between manikin shell and the ambient temperature. Since this large temperature difference makes it difficult to measure the skin temperature correctly, the temperature data would have a measurement error. Finally, they used different kinds of manikin from ours. The different types of manikin material, skin fabric, or the depth of temperature sensor would have caused a different heat transfer coefficient.

In this experiment, the reliability of the measured skin temperature with MT was provided from three aspects. The first is that the mean temperature of MT ((B) in Figure 4) was very close to the mean temperature measured by thermography ((E) in Figure 4) in a nude wet condition. The second is that the mean standard deviations of the skin temperature in each part were less than about 0.12°C, which was about the temperature error of the thermistor. The third is that the y-intersects of the linear regression line of MT were almost equal to EWS. Since the temperature of EWS is considered to be correct, the measured temperature by MT would be correct.

Conclusion

For the first time, we have directly measured the surface temperature of the skin fabric by fixing MT to the fabric on each manikin zone and compared it with that of EWS during walking. The decrease in the mean skin temperatures of MT compared with EWS were 0.49, 0.62, 0.75, and 0.89°C; the evaporation resistances of MT compared with EWS were lower by 9.1, 10.9, 12.5, and 14.4% for walking at 0.00 (standing), 0.27, 0.53, and 0.80 m/s, respectively. The equations to predict the mean wet fabric skin temperature were provided for each of five parts: Arm, Trunk, Hip, Thigh, and Calf. The area-weighted mean skin temperature difference (TD) between EWS and MT for all parts was expressed as the following equation: TD = Heat Flux (W/m²) × 0.0092. This equation predicted the mean skin temperature of wet fabric during walking fairly well. These equations would provide an estimate of the mean skin temperature without measuring the temperature directly.

Footnotes

Funding

This study was financially supported by the special research of the National Institute of Occupational Safety and Health.

Acknowledgements

The authors would like to thank Richard Burke and Steve Rodriguez at Measurement Technology Northwest for their valuable suggestions to this study.

References

Havenith

. Heat balance when wearing protective clothing. Ann Occup Hyg 1999; 43: 289–296.

Holmér

. Protective clothing in hot environments. Ind Health 2006; 44: 404–413.

Gonzalez

Bernard

Carroll

Bryner

Zeigler

. Maximum sustainable work rate for five protective clothing ensembles with respect to moisture vapor transmission rate and air permeability. J Occup Environ Hyg 2006; 3: 80–86.

Bishop

Clapp

. Climate under impermeable protective clothing. Int J Ind Ergo 2000; 25: 233–238.

ISO 7933. Ergonomics of the thermal environment – Analytical determination and interpretation of heat stress using calculation of the predicted heat strain (Second edition, 2004-08-15). Geneva, Switzerland: International Organization for Standardization, 2004.

Havenith

Holmér

Den Hartog

Parsons

. Clothing evaporative heat resistance- proposal for improved representation in standards and models. Ann Occup Hyg 1999; 5: 339–346.

McCullough

Jones

Huck

. A comprehensive data base for estimating clothing insulation. ASHRAE Transact. 1985; 91: 29–47.

McCullough

Jones

Tamura

. A data base for determining the evaporative resistance of clothing. ASHRAE Transact 1989; 95: 316–328.

ASTM. ASTM Standard F 2370-05, ‘Standard Test Method for Measuring the Evaporative Resistance of Clothing Using a Sweating Manikin’. West Conshohocken, PA: American Society for Testing and Materials, 2005.

10.

Wang

Kuklane

Chunsai

Holmér

. Development and validity of a universal empirical equation to predict skin surface temperature on thermal manikins. J Ther Biol 2010; 35: 197–203.

11.

Fan

Qian

. New functions and applications of Walter, the sweating fabric manikin. Eur J Appl Physiol 2004; 92: 641–644.

12.

Bouskill

Havenith

Kuklane

Parsons

Withey

. Relationship between clothing ventilation and thermal insulation. AIHA J 2002; 63: 262–268.

13.

Havenith

Heus

Lotens

. Clothing ventilation, vapour resistance and permeability index: changes due to posture, movement and wind. Ergonomics 1990; 33: 989–1005.

14.

Lotens

Wammes

LJA

. Vapour transfer in two-layer clothing due to diffusion and ventilation. Ergonomics 1993; 36: 1223–1240.

15.

Gebremedhin

. A model of evaporative cooling of wet skin surface and fur layer. J Ther Biol 2001; 26: 537–545.

16.

Fan

Luo

. Heat and moisture transfer with sorption and condensation in porous clothing assemblies and numerical simulation. Int J Heat Mass Transfer 2000; 43: 2989–3000.

17.

Haslam

Parsons

. An evaluation of computer-based models that predict human responses to thermal environment. ASHRAE Transact 1988; 94: 1342–1360.