The Study on Effects of Walking on the Thermal Properties of Clothing and Subjective Comfort

2.1 Objective measuring methods

2.2 Wear trials: subjective judgments and physiological data

The thermal manikin imitates an adult male with the height of 185 cm and chest girth of 100 cm, and correspondingly three male subjects of resembling body proportions were scanned and laboratory trials were performed. The subjects of the same age (23 years), with the height of 184 cm, chest girth ranging from 98 to 102 cm, and body mass varying from 77 to 79 kg, were scanned in compliance to ISO 13402 – 3 [50] and ISO/TR 10652 [51]. Since thermal manikins are designed for steady-state measurements, they do not work well under transient environmental conditions [52] and the human subjects were employed as evaluators of the clothing thermal quality during short-term test trials [53]. Simultaneously, the skin temperature and RH were measured for Li's study during environmental transients, pointed to the relation between the comfort perception to the skin temperature and microclimatic RH [54].

After scanning, the same subjects conducted a series of the laboratory trials according to ISO 10551 [55]. The subjects did one trial at the time in specific set of conditions throughout all three phases and were exposed to three sets of the climatic conditions to better correspond to the real conditions of wearing selected garments in the different cool and the moderate workplace environments.

The wear trials are important in order to define a range of ambient temperatures in which clothing can be worn. The basic intention of the produced jackets is to give the appropriate thermal protection under cool environments. The clothing's effective thermal insulation (I_cle) is required to have a minimum value of 0.190 m²/KW and resultant effective thermal insulation (I_cler) of 0.170 m²K/W in order to protect against cool environments [56], with ambient temperatures −5°C < ta < 15°C [57]. The three sets of the climatic conditions varied in the value of the air temperature (t_a), while the RH = 50% and the air speed (v_a=1.2 m/s) remained the same. In the climatic condition A, air temperature was set to 20°C (in order to test the thermal comfort felt by humans in warmer environments and to compare the results to the standardized values of the climatic conditions during the thermal manikin testing), in B it was set to 15°C, and in C to 10°C (in order to simulate the cool environments).

The wind speed of 1.2 m/s was chosen since Havenith et al. [12] proved that the windup to 4 m/s has a small effect on the intrinsic clothing insulation, because it does not penetrate materials and thus does not disturb the entrapped air layers.

In this study, different activities (walking and resting) and ensembles (CE 2 to CE 5) were combined to examine the insulation values and the subjective judgments. During each laboratory trial, which lasted for altogether for 70 minutes and was divided into three phases, the questionnaires were filled thrice, at the end of each trial phase, under artificially designed ambient conditions (A, B, or C) in the computer-controlled climatic chamber (Figure 1).

Figure 1

Phases of wear trials

The cold exposure starts with a cooling period of 20–40 minutes, since the heat content of the body tissues is reduced, particularly the skin and the extremities. For the calculation of the heat balance and for some comparisons between conditions and the heart rate were averaged over 10 and 20 minutes during the subjective ratings. The purpose of the selected time interval for averaging was to describe the most constant phases of each experiment and to reduce the effects of small variations during the period [58]. The thermal equilibrium is then restored for the values of the mean skin temperature.

Special questionnaire was developed, and the subjects were asked to fill it during trials. The objective of the questionnaire was to determine the influence of the thermal environment and the clothing on the overall thermal state of the subjects in the transient conditions. The answers to the questions were given in the form of the evaluative judgment scale which is one pole scale consisting of 5°, where 0 stands for point of origin and 1, 2, 3, and 4 are degrees describing intensity. Grade 0 was taken as the complete comfort and corresponds to the satisfaction of 100%, while grade 4 corresponds to 0% of satisfaction or a complete lack of comfort felt by the subject. The grades have only positive values.

The physiological parameters have been measured simultaneously. The physiological reactions during wear trials can be objectively measured as a change of the physiological parameters or perceived subjectively, based on the assessment of the thermal comfort using the subjective judgment scales [55].

In order to validate the evaporation of the sweat from the surface of the skin, a body mass loss (Δm_sw) due to sweating was calculated according to ISO 9886 [59]:

where Δm_g is the gross body mass loss (kg), Δm_sw is the body mass loss due to sweat loss during the time interval (kg), Δm_res is the body mass loss due to evaporation in the respiratory tract (kg), Δm_o is the mass loss due to the mass difference between carbon dioxide and oxygen (kg), Δm_wat is the mass variation of the body due to intake and excretions of water (kg), Δm_so is the mass variation of the body due to intake and excretions of solids (kg), Δm_clo is the mass variation due to variation of clothing or sweat accumulation in the clothing (kg).

The evaluation of the thermal strain due to sweat loss was calculated as an amount of sweat evaporated from the body and condensed into the clothing. The overall amount of sweat excreted was calculated from the difference in the gross body mass of the test subjects (Δm_g), the mass variation due to variation of clothing or to sweat accumulation in the clothing (Δm_clo), and the amount of sweat evaporated to the environment. The electronic scale TPT3 (Libela, Celje) with the accuracy of ±2g was used to measure the garment’s weight, and the electromechanical scale E 2100/5 (Libela, Celje) with accuracy of ±5 g was used to measure the subject’s weight. Since there was no intake of fluids or food and no excretions of urine or solids, the mass variation of the body due to intake and excretions of water (Δm_wat) and the mass variation of the body due to intake and excretions of solids (Δm_sol) were equal to 0. The body mass loss due to evaporation in the respiratory tract (Δm_res) and the mass loss due to the mass difference between carbon dioxide and oxygen (Δm_o) were calculated on the basis of the average metabolic rate (M = 115 W/m² under walking phase and 60 W/m² under resting phase) [60], the respiratory quotient (R = 0.80), and the DuBois and DuBois body area (2.029 m²) [61].

The measurement of body temperature is another important parameter to determine the thermal state of the human body [62] but implies the maintenance of body temperature within relatively narrow limits [63]. When the mean skin temperature falls below 31°C, discomfort occurs as the effect of cooling in the cold [64]. Another study by Liz et al. concluded even higher values of the mean skin temperature of resting person who feels thermal comfort. The mean skin temperature of 32.6°C is the limit between cool discomfort and comfort, and 33.7°C is the limit between comfort and warm discomfort [65].

The mean skin and the skin’s RH were deduced by a modular signal recorder MSR 12 (MSR Electronics GmbH, Switzerland). The average or the mean skin temperature (t_sk) was measured at the skin’s surface with 8 external digital waterproof temperature sensors encased in an epoxy (the working range −40 to 125°C, the resolution 0.0625°C, the accuracy ±0.5°C at −10 to 85°C). Although accurate measurement of the body temperature with temperature sensors encased in an epoxy is less precise [66], noninvasive measuring methods were used. The accumulation of moisture to clothing was measured during trials by weighting clothing before and after trials, since during activities human body starts sweating. The evaporated sweat penetrates the fabric [67] in close contact with the skin and can cause the subjects to feel uncomfortable [68].

The mean skin temperature is the mean value of the skin over the whole body according to ISO 9886 [59], and the weighting scheme according to 8 measuring points was selected. It was calculated by weighting each of the local temperatures with the coefficients corresponding to the relative surface of the body area that each measuring point represented [59]:

2.3 Statistical analysis

The statistical analysis was performed using MS Excel and the statistical software STATISTICA. The average mean skin’s temperature (t_sk [°C]) and the average skin’s RH (RH_sk [%]) between different climatic conditions and for the selected clothing ensembles were calculated.

The statistical analysis of each scale was assigned by the standard ISO 10551 [55]. Since this is the preliminary research of the volume impact on the insulation of the three-layered clothing, only three subjects were chosen to perform wear trials.

The percentage of judgments expressing discomfort was calculated for the answered questions. The values have been converted to the percentage and analyzed. Since Li’s study [24] pointed to the relation between the comfort perception to skin a temperature and microclimatic RH during environmental transients, the multiple regression analysis was performed.

The correlation was performed between the percentage of judgments expressing discomfort and the thermal state.

The multiple regression analysis was performed to analyze the connection between the percentage of judgments expressing discomfort (D_ex [%]), the mean skin temperature (tsk [°C]) and skin’s (RH_sk [%]). The similar analysis was done to compare the mutual influence of the ensemble’s microclimatic volume (V_ml [dm³]) on the effective thermal insulation of clothing ensemble in static conditions (I_cle [m² K/W]) or in dynamic conditions (I_cler [m² K/W]) and the percentage of judgments expressing discomfort (D_ex [%]).

3.1 Volume calculations and their impact on the effective thermal insulation values

The Geomagic Verify Viewer software following the scan reconstruction calculated the volumes and the surface areas of the naked body and the clothed body. As summarized in Table 3, the volume (V_CE) and the area (A_CE) of the selected clothing ensembles were calculated initially, and later the increase in the ease allowance value in comparison to jacket CG 1 was calculated, while Table 4 summarizes the volume of the microclimatic air layer under the selected clothing ensembles (V_ML,CE) and jackets (V_ML,CG). The average ( ) and standard deviation (σ) values are summarized in Tables 3 and 4 as calculated by previous study [31].

Table 3

Calculation of the overall ensemble volume (V_CE), the overall ensemble area (A_CE) using the Reverse Engineering software Geomagic Verify Viewer together with calculated average ( ) and standard deviation (s) values

Scanned object	Statistics	VCE (dm3)	ACE (m2)
Naked body		78.58	1.95
	s	0.0205	0.0094
Body dressed in CE 0 (no jacket is added as outer garment)		97.83	2.26
	s	0.0544	0.0125
Body dressed in CE 1 (jacket variant CG 1 added as outer garment)		109.34	2.32
	s	0.0321	0.0102
Body dressed in CE 2 (jacket variant CG 2 added as outer garment)		111.39	2.36
	s	0.0267	0.0055
Body dressed in CE 3 (jacket variant CG 3 added as outer garment)		112.15	2.40
	s	0.0163	0.0125
Body dressed in CE 4 (jacket variant CG 4 added as outer garment)		112.69	2.43
	s	0.0090	0.0060

As summarized in Table 3, the average overall ensemble volume is much bigger than previously reported [4, 69, 70], but do roughly correspond to study done by McQuerry et al. [71], although this study was done on three-layer combinations for fire protection with much bigger ease allowances added. This was expected since most of the studies do not report the measurements for three-layered ensembles.

The results listed in Table 4 also point to much bigger values of the microclimatic air volume than previously reported by any other studies [24, 27].

There was a small difference in the calculated volume (V_CE) and the area (A_CE) between the clothing ensembles CE 2, CE 3, and CE 4. This is due to fabric folding and overlapping, regardless of the actual ease allowance value added to the jacket construction [20]. Fabric drape of the textile materials is directly related to the sense of fullness and graceful appearance of a garment during wear [72]. Loose-fitting clothing provides a larger surface area as well as a greater clothing microenvironment volume [73], but provides more draping. When compared to the area of the pattern parts as summarized in Table 2, there is a distinguished difference between the area of the scanned jackets and the area of the two-dimensional (2D) patterns. This difference is the consequence of overlapping and wrinkling of the fabric when the garments are worn over the body.

Since previous studies verified the strong influence of wind and human moving on the thermal insulation value of clothing, the insulation values were compared between static (thermal manikin standing) and dynamic conditions (thermal manikin moving).

The average total and effective insulation values, together with the calculated variance (s²), for the selected clothing garments, are summarized in Table 5. The maximum thermal insulation with the thermal manikin stationary was measured for jacket’s microclimatic volume of 13.56 dm³ [31].

Table 5

Calculation of the average total and effective insulation value for the selected clothing garments together with the calculated variance (s²)

Jacket variant			s2
CG 1	0.1472	0.0728	5.05×10−05
CG 2	0.1569	0.0826	7.66×10−05
CG 3	0.1561	0.0817	4.80×10−05
CG 4	0.1510	0.0766	1.93×10−05

The average total and effective insulation values, together with the calculated variance (s²), for the selected clothing ensembles are summarized in Table 6. For ensembles’ microclimatic volumes up to 33.57 dm³ in static [31] and up to 30.76 dm3 in dynamic conditions, the insulation enlargement was measured as summarized in Table 6. There was an obvious change in the insulation value due to the effects of walking.

Table 6

Calculation of the average total and effective insulation value for the selected clothing ensembles together with the calculated variance (s²)

Ensemble variant			s2			s2
CE 0	0.4042	0.0627	1.25×10−04	0.3814	0.0591	3.69×10−05
CE 1	0.6052	0.0938	2.01×10−04	0.5002	0.0775	7.97×10−05
CE 2	0.6094	0.0945	2.79×10−04	0.4862	0.0754	1.27×10−04
CE 3	0.6165	0.0956	2.50×10−04	0.4953	0.0768	5.66×10−05
CE 4	0.6100	0.0945	3.67×10−04	0.4943	0.0766	5.21×10−05

For microclimatic volume of 33.57 dm³ formed under ensembles with corresponding 14.32 dm³ of air volume formed under jacket, the maximum thermal insulation in static conditions was measured [31]. For dynamic conditions, the maximum insulation was measured for ensemble's microclimatic volume of 30.76 dm³ (with 11.51 dm³ of air volume under jacket). The jacket with smallest microclimatic volume provided 5.2–13.5% less insulation than wider jackets with larger microclimatic volume.

The ensembles’ insulation ranged 0.0938 < I_cle < 0.0956 and 0.0754 < I_cler < 0.0775, thus pointing to conclusion that subjects will start to feel discomfort during rest when exposed to lower ambient temperatures (10°C and 15°C).

The ensembles with tighter jackets added as the outerwear garments showed 0.74–1.9% less insulation in static conditions [31] and 0.9–2.7% more insulation in dynamic conditions. The ensembles’ insulation increased by 49.6–52.4% with jacket added on top of underlying garments in static conditions and by 27.6–31.1% in dynamic conditions. There was a significant fall in the effective thermal insulation values of the ensembles in static and dynamic conditions.

The increase in the ease allowance value caused simultaneous air volume amplification in the microclimatic area. The percentage of the volume and insulation increase for ensembles was calculated in comparison to nude body as shown in Figure 2.

Figure 2

The percentage of the increase for the microclimatic volume and the effective thermal insulation of the ensembles

Between CG 1 and CG 2, there was an increase in the ease allowance for 18.2%, which caused ensemble's microclimatic volume increase by 17.3% (CE 2) and this volume enlargement accounted for 51.8% increase in ensemble's insulation value during rest and for 26.45% insulation increase in dynamic conditions. Additional ease allowance increase in jackets in amount of 36.4% (between CG 2 and CG 3) caused the overall ensemble's volume enlargement for another 0.97% (between CE 2 and CE 3) and the corresponding insulation increase by 1.81% in static conditions and by 2.28% in dynamic conditions. Further 18.1% ease allowance increase between jackets CG 3 and CG 4 caused the overall ensemble's volume increased by 0.69% (CE 4), thus providing ensemble's insulation decrease by 1.7% in static conditions and 0.24% insulation decrease in dynamic conditions. In other words, the enlargement of the microclimatic volume causes the insulation increase in both jackets and ensembles up to a certain critical point. It also causes corresponding ensembles’ insulation decrease with the thermal manikin moving.

The correlation analysis showed moderate positive linear relationship between the volume of ensembles to effective thermal insulation measured at static conditions (r = 0.7) but weak negative linear relationship to effective thermal insulation measured at dynamic conditions (r = −0.44).

The effective thermal insulation value was reduced by 20.98% between the standing and moving manikins for CE 1. For CE 2, the reduction in the effective thermal insulation in dynamic conditions in comparison to the static conditions was 25.34%. The overall reduction in insulation value decreased from 25.34% to 24.48% for CE 3, and even further insulation reduction to 23.41% was shown for CE 4.

If the main intention is to provide the smaller reduction in the effective thermal insulation value during activities, increase in the ease allowance of the outerwear clothing should be either smaller than 47% or bigger than 153% if we presume the value of ease allowance in amount of 15 cm to be the reference point. The highest effective thermal insulation value was observed for the jacket's ease allowance value of 34 cm with thermal manikin stationary. However, in dynamic conditions, with the thermal manikin moving, the highest effective thermal insulation value was recorded for jacket with 22 cm of the ease allowance added. The smaller the increase in the ease allowance value, the smaller the reduction in the effective thermal insulation in dynamic conditions.

3.2 Wear trials: the subjects’ physiological data and grade analysis

The physiological data, recorded simultaneously during the wear trials in controlled simulated laboratory environments, are important mean by which the ergonomist can control whether or not the subjects feel comfortable during testing. They also test the influence the different human sensations on the overall thermal comfort perception under different activity levels and different environment conditions. In this section, the mean temperature and the RH of skin are analyzed. The secreted sweat was also calculated in order to see the average body mass loss and sweat accumulation into clothing during actual clothing usage. The average mean skin temperature and RH were calculated for each climatic condition, worn ensemble, and trial phase separately.

The average mean skin's temperature values (t_sk [°C]) for each subject's dresses in selected clothing ensembles are shown in Figure 3 during all trial phases in different climatic conditions. In the climatic condition A with the highest ambient temperature of 20°C, the average value of the mean skin temperature at rest was 32.10°C, at first walking phase was 32.25°C and 31.84°C at second walking phase. With the drop of the ambient temperature for 5° in the climatic condition B, the average mean skin values decreased for about 0.91° in the first phase, for about 1.23° in the second and 1.67 in the third phase. With further drop of the ambient temperature for another 5°, the mean skin temperature in the first phase decreased for about 0.31°, for about 0.57° in the second and 0.85 in the third phase. The mean skin temperature generally ranged from minimum 28.52°C to maximum 33.65°C in the climatic condition C, from 29.03°C to 34.09°C in the climatic condition B, and from 29.61 to 33.30°C in the climatic condition A.

Figure 3

The average mean skin's temperature (t_sk [°C]) and variances (s²) for each ensemble during all trial phases in different climatic conditions

Since the mean skin value of 30–32°C is considered comfortable, pointing to conclusion that subjects are exposed to neutral comfort zone. This also points to conclusion that the perceptual adaptation was fully achieved at ambient temperature of 20°C. When the mean skin temperatures falls below 31°C, discomfort occurs [64] as seen with temperature decrease in phases 2 and 3 at ambient temperature of 15°C and through all phases at ambient temperature of 10°C.

Generally, the highest mean skin temperature values were recorded for CE 4 at the climatic condition A during all three phases. In the climatic conditions B and C, the highest recorded values of the mean skin temperature were while wearing CE 1. Generally, conditions, the highest mean skin temperatures, were recorded at the first walking phase during all climatic conditions.

The recorded values of the skin's RH are presented in Figure 4. The highest values of RH were recorded for the climatic condition A, especially at the resting phase while the second highest values of the humidity were measured in the coldest environment during the first walking phase. Overall, the highest recorded values were shown for CE 2 in the climatic condition A, for CE 3 in the climatic condition B and CE 4 in the climatic condition C.

Figure 4

The average skin's relative humidity (RH_sk [%]) and variances (s²) for each ensemble during all trial phases in different climatic conditions

In the climatic condition A with the highest ambient temperature of 20°C, the average value of the skin's RH at rest was 77.8%, 75.6% at first walking phase and 74.0% at second walking phase. With the drop of the ambient temperature for 5° in the climatic condition B, the average skin's RH values decreased for about 22.1% in the first phase, for about 26.4% in the second and 31.6% in the third phase. With further drop of the ambient temperature for another 5°, the skin's RH started increasing for about 11.7% in the first walking phase, for about 13.8% in the second and 16.3% in the third phase.

The values of the body mass loss due to sweating were also calculated (Figure 5), but the measurements were taken after each trial, so the values were not given for each phase separately. The highest values of the body mass loss due to sweating were calculated for CE 4 in the climatic conditions A (203.6 g) and C (188.9 g) as expected since the measured values of the mean skin temperature and RH were also high.

Figure 5

Calculation of the average body mass loss due to sweating (Δm_sw) with the average difference in ensemble’s weight (Δm_cl) shown as numeric value for the selected clothing ensembles

The garment's weight was measured before subjects entered climatic chamber and after they completed trials. The average difference in total ensemble's weight points to sweat accumulation into clothing which is particularly pronounced at lower ambient temperatures as shown in Figure 5. Under coolest climatic conditions, the skin's RH increased, causing subjects to accumulate sweat into their clothing, especially while wearing tighter jackets.

The subjects filling-in questionnaires expressed no or little dissatisfaction with the environmental conditions or the clothing. The grades leaning toward discomfort were shown for the lowest ambient temperature at resting phase although the subjects stated that the thermal environment is generally acceptable considering the grades given to question 3. The grades are in the range of the thermal neutrality. The slight discomfort was determined only for the clothing ensembles CE 1 and CE 3 (Figure 6).

Figure 6

The percentage of judgments expressing discomfort with the thermal state (question 1), the thermal protection provided by the clothing (question 2), and the overall acceptability (question 3)

Generally, the highest grades were given while wearing CE 4 (the outerwear was jacket with the highest value of the ease allowance and consequently the greatest volume of the microclimatic air layer inside the ensembles). The lowest grades were displayed for CE 1 and 3. When sweat starts to condense into clothing, it will cause the subjects to feel uncomfortable.

The subjects exposed to lower ambient temperatures (the ambient conditions B and C) gave lower grades to the ensembles with the tighter jackets. Although CE 4 did not have the highest value of the effective thermal insulation in the dynamic conditions, it was judged as the best clothing ensemble by the subjects in the climatic conditions C (t_a = 10°C).

The percentage of judgments expressing discomfort was calculated for the questions 1, 2, and 3. The highest percentage of judgments expressing discomfort was given under the climatic conditions B and C in the resting phase (phase 2) for all of the questions, especially for CE 1 and CE 2. There was almost no or negligible percentage of dissatisfied subjects in the first walking phase under all environmental conditions except while wearing CE 3 under the condition C. In the third walking phase, the highest percentage of judgments expressing discomfort was found for CE 1 and CE 3 under the condition C (Figure 6). The highest percentage of the comfort was given for CE 4 throughout all the climatic conditions. As previously reported, discomfort occurs when the mean skin temperature falls below 31°C. Decrease in the mean skin temperature below that value was accompanied with a higher percentage of judgments expressing discomfort during exposures to cool environments.

The percentage of judgments expressing discomfort with the thermal state negatively correlates with ensemble's volume through all phases of trials and all climatic conditions. Strong negative linear relationship was proven during resting phase at t_a = 10°C (r = −0.9). Moderate linear relationship during resting phase at t_a = 10°C was also proven to effective thermal insulation measured at static (−0.7) and dynamic conditions (0.8).

Since Li's study [24] pointed to the relation between the comfort perception to skin a temperature and microclimate RH during environmental transients, the multiple regression analysis was performed. It showed that 42.43% of the variation to the percentage of judgments expressing discomfort (D_ex [%]) with the clothing thermal protection is explained by the variation of the skin's (RH_sk [%]) and the mean skin temperature (t_sk [°C]):

In order to compare the results and to examine the relationship between several independent and criterion variables, the multiple regression analysis has been performed. The criterion variable was the percentage of judgments expressing discomfort (D_ex [%]). Since the multiple regression can be used to find relationship between variables, the comparison was performed to the measured values of the microclimatic volume (V_ml [dm³]), the effective thermal insulation of clothing ensemble in static conditions (I_cle [m² K/W]), and the effective thermal insulation of clothing ensemble in dynamic conditions (I_cler [m² K/W]).

From Table 8, the observed F-test statistics for all comparisons has the p-value (significant F) greater than 0.05, except in the case of climatic condition A when comparing the percentage of judgments expressing discomfort to the measured values of the microclimatic volume and the effective thermal insulation of clothing ensemble in static conditions. Since the p-value is not less than 0.05, we do not reject the null hypothesis that the regression parameters are zero at significance level 0.05. We reject the null hypothesis only in the abovementioned example and can conclude that the regression parameters are not zero at significance level 0.05.

When observing the computed p-values of the regression parameters as summarized in Table 8, one can conclude that the parameters are jointly statistically insignificant at the significance level of 0.05 for most of the comparisons, although the percentage of the mutual influences are high. Only in the case of the climatic condition A, when comparing the values of the percentage of judgments expressing discomfort to the measured values of the microclimatic volume and the effective thermal insulation of clothing ensemble in static conditions, the observed coefficient of V_ml has estimated the standard error of 0.38, t-statistic of −21.0435, and p-value of 0.03. It is therefore statistically significant at significance level a = 0.05 as p < 0.05.

The multiple regression of the ensemble's microclimatic volume (V_ml [dm³]) to the effective thermal insulation of clothing ensemble in static conditions (I_cle [m² K/W])) and the percentage of judgments expressing discomfort (D_ex [%]) shows that 99.8% of the variation of D_ex is explained by the regression of V_ml and I_cle only for climatic condition A.

The observed results point to conclusion that this preliminary study should be further expanded involving more subjects to analyze the impact of the microclimatic air volume formed under the ensembles to the subjective judgments expressed by the subjects on the thermal comfort felt during wear trials and thermal insulation decrease due to movements.