Using a thermal manikin to determine evaporative resistance and thermal insulation

Abstract

Heat transfer from the human body, especially through the evaporation of sweat from the skin, is often restricted when protective clothing is used, which may result in overheating. For this reason, it is important to consider the parameters of protective clothing as input data in physiological models, such as predicted heat strain. The two most important parameters are thermal insulation and evaporative resistance with clothing area factor strongly influencing both. These parameters were determined for two clothing ensembles using a (dry) non-sweating thermal manikin. First, the clothing area factor was determined using the photographic method. Second, thermal insulation was measured in both static and dynamic conditions, and multiple equations for predicting dynamic thermal insulation from static ones were evaluated. Third, methodology for measuring evaporative resistance based on pre-wetted skin was adopted and multiple corrections were assessed. Finally, sensitivity analyses were completed using PHS to determine the impact of different equations on the duration limited exposure. For the thermal insulation measurements, we found that predictive equation (32) from ISO 9920 was the most accurate, but choosing the correct equation for protective clothing proved challenging. Although a manikin’s surface temperature is widely used for calculating evaporative resistance, the skin temperature should be used instead, since it is correct from a physical point of view and there is a difference of up to 15% in the results. Because these measures are used in thermal risk analyses conditions, a high degree of accuracy and a knowledge of the inputs must be guaranteed.

Keywords

Evaporative resistance thermal manikin resultant thermal insulation protective clothing

Introduction

Heat stress is the net heat load to which a person is exposed and is caused by various factors, such as high ambient temperature, the duration of exposure, worn protective clothing and equipment and heavy labor. The physiological response to a heat stress is called a heat strain and it has an impact on a human body via heat transfer processes [1 –3]. In hot conditions, evaporation is the main pathway for heat dissipation from the human body into the environment [4,5]. Multiple physically demanding jobs, e.g. firefighting or agricultural work, are often carried out in hot working environments. In many cases, less permeable, well insulating protective clothing is used. However, these types of clothing often result in reduced sweat evaporation, leading to an elevation in skin temperature, core temperature and/or sweat rate [4]. Combined with insufficient water replacement and resting periods, this can lead to work-related disorders and various serious diseases (cardiac issues or chronic kidney disease (CKD) [6 –8], for example) in the long term.

To prevent occupational disorders, various heat stress prediction models (e.g. PHS) [9] and thermo-physiological prediction models, such as the Fiala model [10] or the FMTK model (an abbreviation in Czech for the Fiala-based thermal comfort model) [11], also contain clothing properties as some of their most important input parameters. This is why clothing parameters should be measured with the highest possible degree of precision to mitigate prediction errors [12 –14]. As it is important to protect workers’ health, the main purpose of these models is to calculate the maximum exposure to a given environment with a given activity level without endangering the subject. Today, more emphasis is being placed on this area of research as weather conditions continue to change and extreme climatic events (e.g. hurricanes, heat waves) become more commonplace [15,16]. Multiple methods for measuring or estimating of clothing properties are currently in use, but the most realistic option involves using a thermal manikin with detailed body segmentation [12,17–19], since a thermal manikin accurately represents the shape of the human body and its movements in real-life situations. Non-sweating thermal manikins (i.e. manikins that lack a built-in sweating system) are often used as they are simpler, more affordable and more commonplace around the world. In both thermal insulation and evaporative resistance measurements, the manikin’s surface temperature is heated to a constant temperature, and the heat flux in each zone is recorded continuously. For the evaporative resistance measurements, a pre-wetted skin method is incorporated to simulate the evaporation of sweat from the body’s surface [20].

The measurement of resultant total thermal insulation (I_tr) is well established as it is part of the ISO 9920 standard [21] and can be repeated with an accuracy of 4% between repetitions. Although it can be used as input data for the physiological modelling to simulate reality as closely as possible, it is an expensive method with regard to time and the extra equipment necessary to enable the manikin to walk, equipment that is not accessible for many laboratories. Obviously, a reliable alternative is needed. One option is to calculate the resultant total thermal insulation values (I_tr) from the static total thermal insulation values (I_t).

There are multiple equations (32) to (36) available in the standard ISO 9920 [21] that can be used to predict the resultant total thermal insulation (I_tr) from non-moving manikin values. Equation (33) from the standard can only be applied to a nude manikin or very low insulated clothing. The rest of the equations differ depending on the properties, such as thermal insulation and air permeability, of the clothing ensembles being tested, or on properties of the environments in which the clothing sets are being measured or used, e.g. low wind speed (Table 1). Standard EN 342 from 2004 [22] includes a similar equation with a similar purpose: predicting resultant thermal insulation (I_tr) for cold protective clothing.

Table 1.

Overview of multiple investigated equations from standards for predicting resultant thermal insulation from total thermal insulation values.

Equation label		Area of application
Standard ISO 9920	(32)	Light or normal clothing	0.6 < I_cl < 1.4 clo
	(33)	No clothing	I_cl = 0 clo
	(34)	LOW insulated clothing	0 < I_cl < 0.6 clo
	(35)	Specialized or high insulated clothing	I_cl > 1.4 clo
	(36)	Very low wind activity
Standard EN 342	(EN342)	Cold protective clothing

Another important clothing parameter that has an impact on heat stress and physiological modelling is evaporative resistance [5, 20]. At present, there is no dedicated EU standard for measuring this quantity, highlighting the fact that more research is needed in this field. Two calculation methods for the evaporative resistance of clothing are provided in the ASTM standard from 2010 [23] – the mass loss method and the heat loss method. The mass loss method was removed from the newest version of the ASTM standard [24]. Its omission was probably related to the challenges associated with using the mass loss method to calculate the localized values for separate body parts, which were added to this new ASTM standard [24], as well as to reduce the number of discrepancies between laboratories caused by different measuring and calculation methodology [25]. According to Wang et al. [26], the exclusion of the mass loss method was not an ideal solution as this method is inherently correct from a physical point of view.

It was previously reported that measuring evaporative resistance in non-isothermal conditions (T_manikin ≠ T_ambient ≠ T_sk) may cause significant error as clothing insulation changes dramatically with absorbed moisture; therefore, isothermal conditions should be used [27 –29]. However, we were unable to set up isothermal conditions (T_manikin = T_ambient = T_sk) for our thermal manikin since most of the available manikins are only able to control manikin’s surface temperature and not the wetted skin temperature. For this reason, a so-called isothermal condition (T_manikin = T_ambient ≠ T_sk) had to be used instead. Issues connected to the manikin’s surface temperature reach beyond the measurement setup where this temperature is also used incorrectly in the calculations of evaporative resistance via both of the methods that were mentioned. Because the water (simulating sweat) evaporates from the pre-wetted skin of the manikin, the skin temperature should be used in the calculations. Multiple corrections from different researchers and for different sweating simulation systems were presented to mitigate this and other possible errors [26,29–35].

Material characteristics of individual garments also play significant role on a heat stress of the workers. For example, it was previously reported that increase in an air permeability (porosity) of an over-garment causes an increase in a ventilation, and thus increases in the heat dissipation from a human body [36, 37]. Porosity and capillary drift have strong effects if we look on a material in detail, however when we consider clothing system, then due to many layers and airgaps present in the ensemble, the influence of a specific material and its porosity is partially reduced. That is probably the reason why porosity is included in the moisture performance evaluation of textile materials, but if we look in any standards and methods for measuring parameters of clothing ensembles, then porosity is not considered specifically in any of these. When measuring evaporative resistance of an ensemble on a thermal manikin, in as stable state as possible, we obtain total evaporative resistance values of the whole ensemble, including the effects of layers, pores etc. As the aim of this paper was related to a standardized and repeatable manikin testing and related calculations to define a technical parameter of an ensemble, we will not focus on the material parameters and layering of the garments in the ensembles in this paper.

The aim of this study was to identify, and possibly enhance, reliable and applicable methods to obtain protective clothing parameters by means of a non-sweating thermal manikin using pre-wetted skin. Both the mass loss and the heat loss methods were used to calculate whole-body total evaporative resistance in this study. Verification of multiple equations used for predicting resultant total thermal insulation (I_tr) from total thermal insulation (I_t) on protective clothing from ISO 9920 [21] was also conducted. Our findings can be used to achieve more precise and reliable physiological predictions for people working in hot environments and possibly enhance the prevention of health problems in both the short and the long term. Published data concerning clothing parameters could be used to estimate values for similar protective clothing ensembles for use in physiological modelling where clothing properties are not available.

Methods

Study design

Measurement of the total (I_t) and resultant (I_tr) thermal insulation and evaporative resistance (R_et) by means of a non-sweating thermal manikin.

Measurement of the clothing area factor (f_cl) by the photographic method and calculation of intrinsic values.

Measured values are considered as reference values.

Investigation of multiple equations from different standards to predict resultant thermal insulation (I_tr) from total thermal insulation (I_t).

Assessment of multiple published corrections for the evaporative resistance calculation in both the mass loss and the heat loss methods.

Sensitivity analyses to demonstrate the impact of clothing parameters on the duration limited exposure using the PHS model.

Equipment

This study provides relevant information concerning methodology and measurement procedures for three important clothing parameters. The thermal manikin TORE [38] at Lund University, Sweden (Figure 1), which was used in this study, is made of plastic covering a metal frame inside that supports the body parts and joints. It is the size of an average Swedish male from the first half of the 1980s. The manikin is 1.71 m tall and weighs around 32 kg. The body surface area of 1.772 m² is divided into 17 zones. The climatic chamber at Lund University, with dimensions height × width × length: 2400 × 2360 × 3200 mm, was used for the measurements. The chamber can be adjusted from +5 to +60°C and the temperature standard deviation (SD) from the set value is less than ±0.2°C. The relative humidity in this chamber can be adjusted from 10 to 95%, depending on the temperature and the humidity SD from the set value is less than ±5%. Air velocity can be adjusted between 0.1 and 0.7 m/s.

Figure 1.

Tested clothing: (a) underwear for both systems (provided by the laboratory); (b) sugarcane harvester’s outfit (glove on only one hand and leg protector on only one leg – sugarcane is held by the hand with the glove and cut by machete in the other hand, the leg protector covers the leg in the direction of the machetés swing); (c) pesticide sprayer’s protective cover all on top of underwear; (d) pesticide sprayer’s complete outfit, complete with outer protective layers.

First, the manikin was placed inside the climate chamber in an upright posture with the arms hanging freely (Figure 1(a)). This posture is typically reported in the literature and standards for investigating the thermal properties of clothing in static conditions. The manikin´s arms and legs were connected to an articulated stand to enable measurement of the resultant total thermal insulation under walking conditions (I_tr). Next, for the evaporative resistance investigation, the articulated stand was not used in order to make room for a scale, which was installed under the manikin to provide mass loss data acquisition. For this reason, the resultant evaporative resistance values were neither measured nor estimated by the ISO 9920 standard [21], as it was not possible to verify the accuracy of the estimations. For these measurements, the manikin’s pre-wetted skin was used since the thermal manikin TORE has no sweating system. Finally, the manikin was taken out of the chamber to measure the clothing area factor (f_cl) as described in the following chapter.

Two different clothing ensembles used by agricultural workers – sugarcane cutters (SC) and pesticide sprayers (PS) were measured, those being the most common types of work in the cultivation of sugarcane fields (Figure 1(a) to (d)). Tested ensembles were obtained from a company that provides protective clothing for workers in the sugarcane fields of Latin America, where very warm (more than 34°C) conditions prevail. The size “large” was chosen as the best fit for the manikin.

Determining the clothing area factor (f_cl)

The clothing area factor (f_cl) is defined as the ratio of the area of a dressed manikin to a nude manikin. This parameter is one of the factors influencing the heat and mass transfer between the skin and the ambient environment. Obtained clothing area factor values allowed us to calculate the intrinsic thermal insulation (I_cl) and the resulting intrinsic thermal insulation (I_clr) according to equation (7) in ISO 9920 standard [21] for each set

I_{t} = I_{c l} + \frac{I_{a}}{f_{c l}}

Equation (7) from ISO 9920

where I_t is the total thermal insulation of the clothing [m²K/W], I_cl is the intrinsic thermal insulation [m²K/W], I_a is the thermal insulation of the air layer [m²K/W] obtained by measurements made of the nude manikin and f_cl is the clothing area factor.

Several methods for determining the clothing area factor (f_cl) exist, such as the 3D scanning method, the photographic method, or its estimation by calculation [12,39]. With respect to the unusual shape of the sugarcane workers´ clothing, we decided to use the photographic method. This method provides high degree of accuracy without the need for specialized equipment. While in the early use of this method up to six photographs were taken from different sides and angles, an acceptable accuracy could also be achieved using photographs from only two positions: 0° – front side of the standing manikin and 90° – right/left side of the standing manikin (Figure 2) [40]. All photographs were taken with the same camera (Nikon D-3700, Nikon, Japan) from the same position. The distance of the camera stand from the manikin was 4 m and the camera itself was placed at the height of the center of the manikin. The pictures were then processed in a photo editing program (Photoshop CC 2018, Adobe Systems, USA) and the clothing area factor (f_cl) was calculated by comparing the number of black pixels on the nude versus the dressed manikins.

Figure 2.

Edited photos of the nude manikin, SC – the sugarcane cutter´s outfit, and PS – the pesticide sprayer’s outfit, which were used to count black pixels for the clothing area factor (f_cl) calculation.

Thermal insulation measurement and evaluation methodology

The heat loss method (using the global calculation method) was used to determine both the total thermal insulation (I_t) and the resultant total thermal insulation (I_tr). The procedure is described in ISO 9920 [21] and all parameters were set within the required ranges. The manikin's surface temperature was set and maintained at 34 ± 0.1°C. The ambient temperature was set and maintained at 20.0 ± 0.1°C (checked against the average of three temperature measurements taken at 0.1, 1.1 and 1.7 m above the level of the sole of the manikin's foot) with 0.21 ± 0.08 m/s air velocity aimed at the manikin's back (measured at 1.2 m above floor level). The relative humidity inside the chamber was maintained at 40 ± 5%. A walking speed of approximately 3.5 km/h (step rate set at 90 steps/min) was used for the resultant total thermal insulation (I_tr) measurement.

Two applicable options (Table 1) to predict the resultant total thermal insulation (I_tr) for clothing ensembles SC and PS were examined in this study. First, with regard to the total thermal insulation range given in the standard (from 1.2 to 2.0 clo), equation (32) was selected for both ensembles. The second option examined was equation (35), which is recommended for protective clothing and clothing with low air permeability. This applies to our clothing ensembles, especially PS. For the nude manikin, equation (33) from the standard was always used. Equations (34) and (36) were not suitable for our clothing sets and environmental parameters. A comparison of the resultant total thermal insulation (I_tr) was drawn between the values from the two predictive equations (32) and (35), from the standard ISO 9920 [21], and the values obtained from our measurements. A similar comparison was done for the resultant intrinsic thermal insulation (I_clr) using measured clothing area factor (f_cl) values. Despite the fact that the clothing ensembles did not fit the description of standard EN 342 [22] for cold protective clothing, we decided to apply this standard as well to evaluate the error compared to ISO 9920 [21]. The equations used in this article are presented below and are labeled in the same way as in the relevant standards

I_{t r_p (32)} = e^{[- 0.281 * (v_{a r} - 0.15) + 0.044 * {(v_{a r} - 0.15)}^{2} - 0.492 * v_{w} + 0.176 * v_{w}^{2}]} * I_{t}

(32)

I_{t r_p (33)} = e^{[- 0.533 * (v_{a r} - 0.15) + 0.069 * {(v_{a r} - 0.15)}^{2} - 0.462 * v_{w} + 0.201 * v_{w}^{2}]} * I_{t}

(33)

I_{t r_p (35)} = e^{{[- 0.0512 * (v_{a r} - 0.4) + 0.794 * 10^{- 3} * {(v_{a r} - 0.4)}^{2} - 0.0639 * v_{w}] * p^{0.144}}} * I_{t}

(35)

I_{t r_p (E N 342)} = (0.54 * e^{(- 0.15 * v_{a} - 0.22 * v_{w})} * p^{0.075} - 0.06 *ln (p) + 0.5) * I_{t}

(EN342)

where I_tr are the resultant thermal insulation values from each of the equations [m²K/W], v_ar is the air velocity relative to the person [m/s], v_a is the air velocity [m/s], v_w is the walking speed [m/s], I_t is the total thermal insulation measured on the manikin in static conditions [m²K/W] and p is the air permeability of the outer layer of the clothing [l/m²s] [low (e.g. coating) = 1; medium = 50; high (open weave) = 1000]. Air permeability was estimated for our clothing sets (100 for SC and 25 for PS) based on previous experience.

Evaporative resistance measurement methodology

Non-sweating manikins (manikins that lack a dedicated sweating system) can be effectively used to determine evaporative resistance using pre-wetted skin. Despite this method not being standardized, it has been presented in a variety of publications and was used for sweat simulation [41] in our study. The tight-fitting skin (thickness d= 0.9 mm, 95% cotton, 5% elastane) covered the manikin's entire body except for the hands and feet. To simulate sweat on these body parts, gloves (100% cotton) and socks (67% cotton, 30% polyester, 3% elastane) were used. Before each test, the skin was wetted with 925 ± 15 g of water with no drippage.

The thermal manikin was placed inside the chamber in the same upright position and in the same place as for the previous measurement (thermal insulation). However, this time the whole manikin system was placed on a scale to monitor mass loss throughout the measurement (the mass loss method was used only for one of the three measured repetitions as the extra equipment needed was not available for the rest). Both the thermal manikin's surface temperature and the ambient temperature were set to 34 ± 0.2°C to ensure so-called isothermal conditions [26,33]. Relative humidity was maintained at below 50% to provide good evaporation from the manikin's wetted skin. The air velocity was raised compared to the thermal insulation tests, to 0.54 ± 0.16 m/s to ensure even humidity distribution inside the climatic chamber. Evaporative resistance was only measured in static conditions as the measurement setup (with the scale) did not allow for the use of a walking stand.

Evaporative resistance calculation

There are two calculation methods for clothing evaporative resistance provided in the ASTM standard from 2010 [23] – the mass loss method and the heat loss method [42].

Heat loss method

Two corrections, for the skin temperature of the manikin [29,31] and for the heat gains from the environment [33], (both of which are present when using pre-wetted skin), are examined in this study and explained in detail below.

Evaporative resistance was calculated by the heat loss method from the area-weighted heat loss observed from the thermal manikin software. According to ASTM standard [24], the total evaporative resistance should be calculated from equation (1) from the manikin's surface temperature as it assumes that isothermal conditions are used (T_manikin = T_sk).

R_{et, h} = \frac{(p_{s k} - p_{a})}{H_{e, heat}}

(1a)

R_{et, h} = \frac{[\exp (18.956 - \frac{4030.18}{T_{manikin} + 235}) * R H_{s k} - \exp (18.956 - \frac{4030.18}{T_{a} + 235}) * R H_{a}]}{H_{e, heat}}

(1b)

where R_et,h is the total evaporative resistance calculated from the heat loss method [m²Pa/W], p_sk and p_a are the vapor pressures of saturated skin and air, respectively [Pa], T_manikin and T_a are the temperatures of the manikin's surface and ambient air, respectively [°C], RH_sk (96% based on previous measurements) and RH_a are relative humidity near the manikin's surface and in the air, respectively [%], and H_e,heat is the total heating power supplied to the wetted parts of the manikin [W/m²]. R_{et,h_manikin} values were calculated from this equation. To correct for lower skin temperature caused by water evaporation, equation (2) was used to mitigate the error in calculation [31]

T_{sk} = T_{manikin} - 0.0132 * H_{e, heat} * A

(2)

where T_sk is the predicted textile skin temperature [°C] and A is the manikin's sweating surface area [m²]. R_{et,h_skin} values were obtained by accommodating equation (2) for skin temperature into the previous equation (1b).

As there is a temperature gradient between the wetted skin and the environment, part of the heat needed to evaporate the water from the skin is also taken from the environment through conduction, convection, and radiation mechanisms as presented in Figure 3. In some cases (e.g. impermeable clothing, high insulation clothing), evaporation is negligible. This may cause the manikin's power regulation to be episodic, which may, in turn, cause the manikin's surface to overheat and exceed the set temperature. Thus, the effect presented above (Figure 3) can be reversed, and the manikin may exhibit extra dry heat loss instead. That is why another correction for the evaporative resistance calculation was proposed and is presented as equation (3). We get the final equation (4), for calculating the corrected evaporative resistance by incorporating both equations (2) and (3) into equation (1)

Q_{evap} = H_{e, heat} + \frac{(T_{a} - T_{s k})}{I_{t}}

(3)

R_{et, h_skin + envi} = \frac{[\exp (18.956 - \frac{4030.18}{T_{s k} + 235}) * R H_{s k} - \exp (18.956 - \frac{4030.18}{T_{a} + 235}) * R H_{a}]}{H_{e, heat} + \frac{T_{a} - T_{manikin} + 0.0132 * H_{e, heat} * A}{I_{t}}}

(4)

where Q_evap is the part of the heat needed for the evaporation taken from the environment [W/m²], I_t is the total thermal insulation measured on a non-sweating manikin according to ‘Determining the clothing area factor (f_cl)’ section [m²·K/W]. R_{et,h_skin+envi} were determined according to equation (4).

Figure 3.

The heat transfer mechanism between the manikin surface, the wetted skin, and the environment in a so-called isothermal conditions (T_a = T_r = T_manikin ≠ T_skin) without clothing, adapted from Wang et al.[33].

A comprehensive overview of the corrections and the comparison of heat loss vs. mass loss are presented in the review article by Wang [43]. Other proposed corrections (e.g. for moisture in the clothing, for skin fabric) by Wang et al. [26,3233,43] are more complicated to incorporate into practical measurements (importance of measuring skin fabric parameters, wet thermal insulation of clothing etc.) and were not considered in this study. With the use of the clothing area factor (f_cl), the whole body intrinsic evaporative resistance could be calculated for both clothing sets.

Mass loss method

This method is based on the measurement of mass loss rate and its subsequent conversion to evaporative heat loss by multiplying the latent heat of the vaporization of water. The ASTM standard [23] uses the manikin's surface temperature for the calculation of the saturated vapor pressure p_sk. The evaporative resistance calculated from the mass loss method is calculated based on equation (5)

R_{et, m} = \frac{[\exp (18.956 - \frac{4030}{T_{manikin} + 235}) * R H_{s k} - \exp (18.956 - \frac{4030}{T_{a} + 235}) * R H_{a}] * A}{λ * \frac{d m}{d t}}

(5)

where λ is the vaporization heat of water at the measured skin temperature [W·h/g] and dm/dt is the evaporation rate of moisture from the wetted skin [g/h].

To be correct from a physical point of view, this calculation should be corrected in the same way as in the heat loss method, using the predicted skin temperature based on equation (2). R_{et,m_skin} is the evaporative resistance, which is calculated based on equation (5) with the manikin's surface temperature (T_manikin) being substituted for the predicted temperature of the wetted skin (T_sk).

Data analysis

All thermal insulation values presented in this study are the averaged values of two independent measurements with a difference lower than 4% between them as required by the ISO 9920 standard [21]. For the evaporative resistance measurements, the values presented, including standard deviation from the heat loss method, were calculated as an average of three independent measurements. However, the mass loss method was measured only once for each clothing ensemble as a control measurement; therefore, no standard deviation could be presented.

Sensitivity study

PHS simulations were conducted as sensitivity analyses to observe the impact of the clothing properties, obtained by the different equations and corrections, on the workers’ maximum exposure time. The maximum exposure time was evaluated using two different criteria: (a) D_Tre is the time it took an average worker to reach a core temperature limit of 38°C (occupational exposure limit), (b) Dwl_50 is the time it took an average worker to reach the limit for water (sweat) loss. All parameters for PHS simulations except for measured clothing parameters [resultant intrinsic thermal insulation (I_clr) and moisture permeability index (i_m), calculated from the measured thermal insulation and evaporative resistance] were constant and corresponded to the environmental parameters during lunch time in the sugarcane fields of Latin America (Table 2). Parameters defining the human body were chosen based on an average male with a posture, walking speed and metabolic production corresponding to work in a sugarcane field. To analyze the difference between multiple predictive equations for the resultant intrinsic thermal insulation (I_clr), the intrinsic evaporative resistance (R_ecl) values were set to a constant for each clothing ensemble. To investigate the impact of the corrections on the evaporative resistance, the opposite approach was used: the resultant intrinsic thermal insulation (I_clr) was set to a constant value and the intrinsic evaporative resistance (R_ecl) was changed according to the corrections.

Table 2.

Constant input parameters for the PHS model for sensitivity analyses of the impact of measured clothing properties on predicted exposure time.

Height	1.85 m
Weight	80 kg
Posture	Standing
Air temperature	34.6°C
Air velocity	0.2 m/s
Relative humidity	23%
Radiant temperature	63.5°C
Metabolic energy production	200 W/m²
Walking speed	1 m/s

Results

Thermal insulation

In Table 3, the total (I_t) and intrinsic (I_cl) thermal insulation and the total (I_tr) and intrinsic (I_clr) resultant thermal insulation results, both predicted (I_{tr_p}, I_{clr_p}) and measured (I_{tr_m}, I_{clr_m}) are presented. The table gives the complete database of all obtained values. A comparison of the measured and calculated values, both by the standard equation (32) of ISO 9920 [21] (applied based on the clothing insulation range) and by equation (35) of the same standard (used based on the low permeability of the clothing sets being measured) was conducted. The percentage difference (Figure 4) was evaluated based on the resultant intrinsic thermal insulation values, calculated using a clothing area factor of 1.26 for SC and 1.41 for PS, the values having been obtained by the photographic method because these values are used as input data for the PHS model. The difference between the predicted and measured values was −3.6 and −0.6% for SC and PS, respectively, when equation (32) was used and 27.9% and 27.3% for SC and PS, respectively, when using equation (35). Similarly, equation (33) was used for the nude manikin to predict total resultant thermal resistance (I_t), and this value was again compared to the measured value. In this case, the difference was 3.5%. When using the equation from the EN 342 standard [22] (used for cold protective clothing), the difference between predicted and measured values for resultant intrinsic thermal insulation was 16.3% and 18.8% for SC and PS, respectively.

Table 3.

Database of total thermal insulation (I_t), resultant total thermal insulation (I_tr), and resultant intrinsic thermal insulation (I_clr) values obtained by measurements and by multiple predictive equations from different standards.

	I_t	I_{tr_m}	I_{tr_p(32)}	I_{tr_p(35)}	I_{tr_p(EN342)}	I_{clr_m}	I_{clr_p(32)}	I_{clr_p(35)}	I_{clr_p(EN342)}
	[m²K/W]
SC	0.191	0.143	0.138	0.173	0.157	0.083	0.080	0.115	0.099
PS	0.257	0.188	0.185	0.236	0.217	0.134	0.133	0.184	0.165
SC (SD)	0.002	0.001
PS (SD)	0.001	0.001

Figure 4.

Percentage differences between measured resultant intrinsic thermal insulation (I_{clr_m}) and predicted values from equation (32) from ISO 9920 (I_{clr_p(32)}), from equation (35) from ISO 9920 (I_{clr_p(35)}) and from the equation in the EN342 standard (I_{clr_p(EN342)}).

The total thermal insulation of the manikin's skin, which was used for the evaporative resistance calculation was measured only in static conditions and reached 0.131 m²K/W.

The sensitivity analysis was done based on the PHS simulations (Table 4), with all parameters set as constants, except for resultant intrinsic thermal resistance, which was changed according to the equations. The core temperature limit (D_Tre) was not reached during any of the simulations in the selected environmental parameters when clothing set SC was used. However, when clothing set PS was used, the core temperature limit was reached in all cases within 25 to 30 min.

Table 4.

Results from PHS simulations using constant intrinsic evaporative resistance (R_{ecl_m}) as measured by the mass loss method with multiple values of resultant intrinsic thermal insulation (I_clr) based on different predictive equations.

	I_clr	R_ecl,m	i_m	D_Tre (min)	Dwl50 (min)
SC
I_{clr_m}	0.083	20.0	0.248	480	340
I_{clr_p(32)}	0.080	20.0	0.239	480	335
I_{clr_p(35)}	0.115	20.0	0.344	480	382
I_{clr_p(EN342))}	0.099	20.0	0.296	480	362
PS
I_{clr_m}	0.134	81.4	0.098	25	280
I_{clr_p(32)}	0.133	81.4	0.098	25	280
I_{clr_p(35)}	0.184	81.4	0.135	30	280
I_{clr_p(EN342))}	0.165	81.4	0.121	28	280

Evaporative resistance

All data for both ensembles are presented in Table 5. Three different values for total evaporative resistance, calculated by the heat loss method and using different corrections, are shown. First, R_{et,h_manikin} values were calculated from equation (1b). Second, R_{et,h_skin} values were obtained by incorporating the correction for skin temperature [equation (2)] into the previous equation (1b). Third, R_{et,h_skin+envi} values were determined according to equation (4). Finally, two values were obtained from the mass loss method. R_{et_m} and R_{et_m_skin} were calculated according to equation (5), using the manikin's surface temperature and the predicted skin temperature, respectively. The total evaporative resistance of the manikin, wearing only pre-wetted skin, was also measured and calculated from equation (1b) using the heat loss method – 8.2 m²Pa/W. Because this value was only measured once (when testing the entire measurement setup), no statistical analyses could be applied to it. All intrinsic evaporative values (R_ecl) were calculated using the clothing area factor (f_cl) and are also presented in Table 5.

Table 5.

Database of total (R_et) and intrinsic (R_ecl) evaporative resistance, obtained by both the mass loss and the heat loss methods, using various possible calculations and corrections.

	R_et,m	R_{et,m_skin}	R_{et,h_manikin}	R_{et,h_skin}	R_{et,h_skin + envi}	R_ecl,m	R_{ecl,m_skin}	R_{ecl,h_manikin}	R_{ecl,h_skin}	R_{ecl,h_skin + envi}
	[m²Pa/W]
SC	26.7	23.2	26.7	23.0	21.0	20.0	16.5	20.0	16.3	14.3
PS	87.4	83.5	83.7	79.9	79.8	81.4	77.5	77.7	73.9	73.8
SC (SD)			0.9	0.9	0.7
PS (SD)			2.8	2.8	2.8

Note: R_et,m and R_{et,h_manikin} are the evaporative resistances calculated by the mass loss and the heat loss method using the manikin´s surface temperature, R_{et,m_skin} and R_{et,h_skin} are the evaporative resistances calculated by the mass loss and the heat loss methods using the predicted skin temperature of the manikin and R_{et,h_skin+envi} is the evaporative resistance calculated by the heat loss method based on the correction for gains from the environment. Additionally, all of these values were recalculated using the clothing area factor (f_cl), and their intrinsic values are given as well.

Percentage differences between evaporative resistance calculated by the mass loss method using the manikin's surface temperature, a common method at present, and other types of calculations using different corrections are shown in Figure 5. The average deviation and percent deviation were also calculated for five presented calculation methods for both sets, reaching 2.05 and 8.49 for the SC ensemble and 2.42 and 2.92 for the PS ensemble, respectively.

Figure 5.

Percentage differences between evaporative resistance calculated by the mass loss method using the manikin´s surface temperature (a common method that is currently favoured), evaporative resistance calculated by the mass loss method using skin temperature (R_{et,m_skin}), evaporative resistance calculated by the heat loss method using the manikin´s surface temperature (R_{et,h_manikin}), skin temperature (R_{et,h_skin}), and skin temperature with a correction for environmental gains (R_{et,h_skin+envi}).

The sensitivity analysis was done based on the PHS simulations (Table 6) with all parameters set as constants except for the intrinsic evaporative resistance, which was changed according to the corrections mentioned in the ‘Evaporative resistance calculation’ section. Both the core temperature limit (D_Tre) and the water loss limit (Dwl50) could be used for the evaluation as both were reached for both sets in most cases.

Table 6.

Results from PHS simulations using constant measured resultant intrinsic thermal insulation (I_clr) and multiple values of intrinsic evaporative resistance (R_ecl) based on different corrections used for its calculation.

	R_et,m	I_clr	i_m	D_Tre (min)	Dwl50 (min)
SC
R_ecl,m	20.0	0.083	0.248	480	340
R_{ecl,m_skin}	16.5	0.083	0.301	56	368
R_{ecl,h_manikin}	20.0	0.083	0.248	480	340
R_{ecl,h_skin}	16.3	0.083	0.303	55	369
R_{ecl,h_skin + envi}	14.3	0.083	0.346	45	382
PS
R_ecl,m	81.4	0.134	0.098	25	280
R_{ecl,m_skin}	77.5	0.134	0.103	26	280
R_{ecl,h_manikin}	77.7	0.134	0.103	26	280
R_{ecl,h_skin}	73.9	0.134	0.109	27	280
R_{ecl,h_skin + envi}	73.8	0.134	0.109	27	280

The PHS model is simple compared to other physiological models [10, 44, 45] and it would seem that its simplicity demands further consideration. Simplicity might be the reason why a small change in the input data, e.g. pre-set metabolic energy production, causes a considerable change in the final maximal exposure time.

Discussion

Differences between calculated and measured resultant thermal insulation values

We found that the difference between measured values of the resultant intrinsic thermal insulation and those calculated according to equation (32) in ISO 9920 ranged from −0.6 to −3.6%. The difference decreased with the rising total insulation of the clothing being evaluated. The accuracy of the prediction from equation (32) is sufficient; however, it is not clear which of the equations from ISO 9920 should have been used for our ensembles. From the perspective of thermal insulation, equation (32) from ISO 9920 is the best fit. On the other hand, equation (35) is meant to be applied to specialized clothing with impermeable layers. When we used equation (35), however, the difference between the predicted and the measured values was significantly higher: −27.9% for SC and −27.3% for PS. Moreover, a similar predictive equation also exists in the standard for cold protective clothing, EN 342 [22]. The difference between the predicted and the measured values of the resultant total thermal insulation for both sets was lower when using this equation (Figure 4) than equation (35) of the ISO 9920 standard. Obviously, the equation from the standard for cold protective clothing yields results that are closer to the measured ones than equation (35), which is meant to be used for impermeable clothing. Our findings underscore the difficulties associated with choosing the correct equation for a given application. To avoid confusion in the future, a versatile and robust predictive equation for protective clothing needs to be developed. Such an equation could then be applied to the predictive models to enhance their accuracy.

In the next step, a sensitivity analysis was carried out to measure the impact of the inputs on the resulting PHS. Based on water loss criteria, the sensitivity analyses showed that there was very little difference in the exposure time in the SC ensemble. The most significant time difference between predictions based on the measured value and any of the predictive equations was 42 min [measured vs. the equation (35)], which is relatively small but not negligible, especially in relation to an 8-h workday. The core temperature limit was not reached for the whole working day when the SC clothing set was used. For the PS set, even though there were relatively high differences in absolute values between the results from the measurements and the calculations, no significant differences were found in PHS predictions while using these values. The exposure time based on water loss did not change at all and the core temperature maximum exposure time varied by only 5 min (Table 4). However, we cannot conclude that these error margins have a negligible impact on the PHS predictions. As seen in Figure 5, a small change in metabolic production can cause, in some cases, a significant change in the resulting exposure time. This stems from the simplicity of the PHS model, and it would be of a great interest to carry out a similar analysis on a more sophisticated physiological model. However, the more complicated models are not affordable for many researchers.

Differences between evaporative resistance values obtained using multiple corrections

First, two methods for measuring evaporative resistance were compared (Table 5). For the SC ensemble, the same results were obtained (with a difference of 0.002%) by both methods when the corrections were not used (R_{et,h_manikin} = R_et,m ₌ 26.7 m²Pa/W). For the PS set, the difference between the heat loss method (R_{et,h_manikin} = 83.7 m²Pa/W) and the mass loss method (R_et,m = 87.4 m²Pa/W) was slightly higher, amounting to 4.4% (Figure 5).

Second, we investigated the discrepancies caused by the use of multiple corrections (Figure 5). In the mass loss method, the differences between values calculated from the manikin's surface temperature and from the manikin's skin temperature were 13.2% for the SC ensemble and 4.4% for the PS ensemble. Similarly, the heat loss method involved differences of 13.7% and 8.6% for SC and PS respectively. Moreover, when the correction for gains from the environment was used in the heat loss method, the differences compared to the raw values (calculated from the manikin's surface temperature) were even higher 21.2% for SC and 8.7% for PS. We could see that the percentage differences between both the mass loss and the heat loss method are not significant when the same temperature (either the surface temperature or the skin temperature of the manikin) is used for their calculation. However, calculations based on the manikin's surface temperature should not be used as this is not correct from a physical point of view. Water evaporates from the manikin’s skin and not its surface, thus the vapor pressure of saturated skin needs to be used in the calculations. As we were neither able to control nor measure the manikin’s skin temperature with the manikin available to us, it was necessary to use the predicted skin temperature instead.

The results from the sensitivity analyses (Table 6) support our conclusions about using the manikin's surface temperature. For the SC ensemble, the core temperature criteria were not reached when the manikin’s surface temperature was used in either the mass loss or the heat loss method, whereas when the predicted skin temperature was used in both methods, the maximum exposure time was only around 55 min. This represents a huge difference and could result in very inaccurate PHS predictions, which would have the potential to adversely affect the health of the sugarcane workers. As seen in Figure 5, a small change in the input metabolic production in the PHS model can, in some cases, cause a significant jump in the exposure time. Moreover, in one case for the SC ensemble, a lower evaporative resistance value resulted in a lower exposure time, which should not have been the case. These uncertainties are probably caused by the simplicity of the PHS model and should be taken into account when using this model in real-life situations. The PS ensembles insulation was relatively high and also contained multiple impermeable layers, meaning heat transfer between the skin and the environment was negligible and the corrections mentioned had almost no effect on the calculated evaporative resistance values, something that was supported by the results from the sensitivity analysis. As the impermeability of the clothing set has a significant impact on the PHS predictions, there were no differences in exposure time based on water loss criteria and the core temperature limitation was reached fairly quickly (at around 30 min) in all cases.

Results from this study show the need to correct for the pre-wetted skin temperature in the calculations of both methods in order for them to be physically correct. Other corrections should not be needed as they do not improve the results further although creating a larger clothing database would be advisable. However, the correction for the pre-wetted skin temperature can be omitted when impermeable and high insulated clothing is used. These results are in conformity with the conclusions by Wang et al. [33], who stated that the corrections should not be needed for impermeable or high insulated clothing (>2.5 clo). Although our measurement on the PS set supports this statement, it was previously concluded on the basis of only one measurement [33]. Therefore, we recommend further tests with high insulated and impermeable clothing being carried out for verification. Actually, although it might be interesting to pinpoint the exact cut-off point in the thermal insulation range from which corrections are no longer necessary. The corrections have minimal impact when high insulated and impermeable clothing is used, and therefore they can be used in all cases.

The same clothing ensembles were also used in the study presented on the 12i3m – 12th International Meeting on Thermal Manikins and Modeling, Empa, St. Gallen, Switzerland [46].

Conclusion

Three clothing properties that are important for heat stress modelling – clothing area factor, thermal insulation, and evaporative resistance were measured for two clothing ensembles currently used by workers in sugarcane fields. First, for the thermal insulation measurements, we found that equation (32), a predictive equation from ISO 9920 standard is the most accurate for our ensembles; however, there is considerable confusion when determining the most suitable equation from this and other standards when protective clothing is used. Although, there was no significant difference in the resulting PHS predictions when using different equations for calculating the input thermal insulation, an advanced physiological model should be used to verify these results as differences in absolute values were significant in some cases. A new versatile equation for predicting resultant thermal insulation for various kinds of protective clothing should be developed in the future to avoid confusion when choosing from multiple equations. Second, both the mass loss and the heat loss methods for calculating evaporative resistance were used and some proposed corrections for the heat loss method were included. Even when there is no significant difference between the values calculated by the heat loss method and the mass loss method using the manikin’s surface temperature, these calculations should not be used because they cannot accurately represent realistic physical conditions. The correction for the pre-wetted skin temperature should be used instead, especially in cases where there are no impermeable layers and low insulated clothing is being considered. Results of the PHS sensitivity analyses showed that using the manikin’s surface temperature with such clothing can cause inaccurate predictions and is potentially dangerous when used in real-life situations. Usage of other corrections depends largely on the type of clothing being measured, and more investigation involving a more extensive set of data is needed.

Footnotes

Authors' note

Kalev Kuklane is now affiliated with Institute for Safety (IKV), Zoetermeer, Netherlands.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by two projects (RV9080000301, FSI-S-17_4444) at the Brno University of Technology, Czech Republic.

ORCID iD

Róbert Toma

References

Parsons

Human thermal environments: the effects of hot, moderate, and cold environments on human health, comfort, and performance. Boca Raton, FL, USA: CRC Press, Inc., 2014.

ISO 15265. Ergonomics of the thermal environment – risk assessment strategy for the prevention of stress or discomfort in thermal working conditions. Geneva, Switzerland: Author, 2014.

Malchaire

Kampmann

Mehnert

, et al. Assessment of the risk of heat disorders encountered during work in hot conditions. Int Arch Occup Environ Health 2002; 75: 153–162.

Holmer

Protective clothing in hot environments.

Ind Health 2006; 44: 404–413.

Holmer

Protective clothing and heat-stress.

Ergonomics 1995; 38: 166–182.

Wesseling

Aragón

González

, et al. Kidney function in sugarcane cutters in Nicaragua – a longitudinal study of workers at risk of mesoamerican nephropathy. Environ Res 2016; 147: 125–132.

Roncal-Jimenez

García-Trabanino

Barregard

, et al. Heat stress nephropathy from exercise-induced uric acid crystalluria: a perspective on Mesoamerican nephropathy. Am J Kidney Dis 2016; 67: 20–30.

Flouris

Dinas

Ioannou

, et al. Workers’ health and productivity under occupational heat strain: a systematic review and meta-analysis. Lancet Planet Heal 2018; 2: e521–e531.

Malchaire

Piette

Kampmann

, et al. Development and validation of the predicted heat strain model. Ann Occup Hyg 2001; 45: 123–135.

10.

Fiala

Lomas

Stohrer

A computer model of human thermoregulation for a wide range of environmental conditions: the passive system. J Appl Physiol 1999; 87: 1957–1972.

11.

Pokorný

Fišer

Fojtlín

, et al. Verification of Fiala-Based human thermophysiological model and its application to protective clothing under high metabolic rates. Build. Environ 2017; 126: 13–26.

12.

Fojtlín

Psikuta

Fišer

, et al. Local clothing properties for thermo-physiological modelling: comparison of methods and body positions. Build Environ 2019; 155: 376–388.

13.

Veselá

Psikuta

Frijns

AJH.

Local clothing thermal properties of typical office ensembles under realistic static and dynamic conditions. Int J Biometeorol 2018; 62: 2215–2229.

14.

Veselá

Kingma

BRM

Frijns

AJH.

Local thermal sensation modeling – a review on the necessity and availability of local clothing properties and local metabolic heat production. Indoor Air 2017; 27: 261–272.

15.

Vitousek

PM.

Beyond global warming – ecology and global change. Ecology 1994; 75: 1861–1876.

16.

McMichael

Woodruff

Hales

Climate change and human health: present and future risks. Lancet 2006; 367: 859–869.

17.

Psikuta

Allegrini

Koelblen

, et al. Thermal manikins controlled by human thermoregulation models for energy efficiency and thermal comfort research – a review. Renew Sustain Energy Rev 2017; 78: 1315–1330.

18.

Holmer

Thermal manikin history and applications.

Eur J Appl Physiol 2004; 92: 614–618.

19.

[19]

McCullough

EA.

The use of thermal manikins to evaluate clothing and environmental factors. Elsevier Ergonom Book Ser 2005; 3: 403–407.

20.

McCullough

Kenney

WL.

Thermal insulation and evaporative resistance of football uniforms. Med Sci Sports Exerc 2003; 35: 832–837.

21.

ISO 9920. Ergonomics of the thermal environment – estimation of thermal insulation and water vapour resistance of a clothing ensemble. Geneva, Switzerland: Author, 2007.

22.

EN 342. Protective clothing – ensembles and garments for protection against cold. Brussels, Belgium: Author, 2004.

23.

ASTM F2370 – 10. Standard test method for measuring the evaporative resistance of clothing using a sweating manikin. West Conshohocken, PA: Author, 2010.

24.

ASTM F2370-16. Standard test method for measuring the evaporative resistance of clothing using a sweating manikin. West Conshohocken, PA: Author, 2016.

25.

Richards

MGM

McCullough

EA.

Revised interlaboratory study of sweating thermal manikins including results from the sweating agile thermal manikin. Journal of ASTM International 2005; 2(4): 27–39.

26.

Wang

Lai

Shi

, et al. Effects of fabric thickness and material on apparent ‘wet’ conductive thermal resistance of knitted fabric ‘skin’ on sweating manikins. J Therm Biol 2017; 70: 69–76.

27.

Xiaohong

Chunqin

Yingming

, et al. The thermal insulation difference of clothing ensembles on the dry and perspiration manikins. Meas Sci Technol 2010; 21: 85203.

28.

Chen

Fan

Zhang

Clothing thermal insulation during sweating. Text Res J 2003; 73: 152–157.

29.

Wang

Kuklane

Gao

, et al. Effect of temperature difference between manikin and wet fabric skin surfaces on clothing evaporative resistance: how much error is there? Int J Biometeorol 2012; 56: 177–182.

30.

Wang

Gao

Kuklane

, et al. A study on evaporative resistances of two skins designed for thermal manikin tore under different environmental conditions. Text Bioeng Inform Symp Proc 2009; 1: 211–215.

31.

Wang

Kuklane

Gao

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

32.

Wang

Shi

, et al. Effects of moisture content and clothing fit on clothing apparent ‘wet’ thermal insulation: a thermal manikin study. Text Res J 2016; 86: 57–63.

33.

Wang

Zhang

YH.

Correction of the heat loss method for calculating clothing real evaporative resistance. J Therm Biol 2015; 52: 45–51.

34.

Ueno

Sawada

Correction of the evaporative resistance of clothing by the temperature of skin fabric on a sweating and walking thermal manikin. Text Res J 2012; 82: 1143–1156.

35.

Koelblen

Psikuta

Bogdan

, et al. Comparison of fabric skins for the simulation of sweating on thermal manikins. Int J Biometeorol 2017; 61: 1519–1529.

36.

Havenith

den Hartog

Martini

Heat stress in chemical protective clothing: porosity and vapour resistance. Ergonomics 2011; 54: 497–507.

37.

Bernard

Ashley

Trentacosta

, et al. Critical heat stress evaluation of clothing ensembles with different levels of porosity. Ergonomics 2010; 53: 1048–1058.

38.

Kuklane

Heidmets

Johansson

Improving thermal comfort in an orthopaedic aid: better Boston brace for scoliosis patients. Hong Kong SAR: The Hong Kong Polytechnic University, 2006, pp. 343–351.

39.

Psikuta

Frackiewicz-Kaczmarek

Mert

, et al. Validation of a novel 3D scanning method for determination of the air gap in clothing. Meas J Int Meas Confed 2015; 67: 61–70.

40.

Havenith

Kuklane

Fan

, et al. A database of static clothing thermal insulation and vapor permeability values of Non-Western ensembles for use in ASHRAE standard 55, ISO 7730 and ISO 9920. ASHRAE Trans 2015; 121: 197–215.

41.

Wang

Zhou

, et al. Empirical equations for intrinsic and effective evaporative resistance of multi-layer clothing ensembles. Ind Text 2010; 61: 176–180.

42.

Havenith

Richards

Wang

, et al. Apparent latent heat of evaporation from clothing: attenuation and ‘heat pipe’ effects. J Appl Physiol 2008; 104: 142–149.

43.

Wang

FM.

Measurements of clothing evaporative resistance using a sweating thermal manikin: an overview. Ind Health 2017; 55: 473–484.

44.

Pokorny

Fiser

Fojtlin

, et al. Verification of Fiala-Based human thermophysiological model and its application to protective clothing under high metabolic rates. Build Environ 2017; 126: 13–26.

45.

Havenith

Fiala

Blazejczyk

, et al. The UTCI-Clothing model. Int J Biometeorol 2012; 56: 461–470.

46.

Kuklane

Regional and total insulation and evaporative resistance values of clothing for sugarcane harvesters and chemical sprayers in Latin America. In: 12th international meeting on thermal manikins and modeling, Empa, St. Gallen, Switzerland, 29–31 August 2018.

Using a thermal manikin to determine evaporative resistance and thermal insulation – A comparison of methods

Abstract

Keywords

Introduction

Methods

Study design

Equipment

Determining the clothing area factor (fcl)

Thermal insulation measurement and evaluation methodology

Evaporative resistance measurement methodology

Evaporative resistance calculation

Heat loss method

Mass loss method

Data analysis

Sensitivity study

Results

Thermal insulation

Evaporative resistance

Discussion

Differences between calculated and measured resultant thermal insulation values

Differences between evaporative resistance values obtained using multiple corrections

Conclusion

Footnotes

Authors' note

Declaration of conflicting interests

Funding

ORCID iD

References

Determining the clothing area factor (f_cl)