Performance Study of Protective Clothing against Hot Water Splashes: from Bench Scale Test to Instrumented Manikin Test

Ann. Occup. Hyg., 2015, Vol. 59, No. 2, 232 242 doi:10.1093/annhyg/meu087 Advance Access publication 27 October 2014 Performance Study of Protective Clothing against Hot Water Splashes: from Bench Scale Test to Instrumented Manikin Test Yehu Lu 1,2, Guowen Song 3,4, * and Faming Wang 1,2 1.Laboratory for Clothing Physiology and Ergonomics, National Engineering Laboratory for Modern Silk, Soochow University, Suzhou 215123, China 2.Department of Human Ecology, University of Alberta, Edmonton T6G 2N1, Canada 3.Department of Apparel, Events, and Hospitality Management, Iowa State University, Ames, IA 50011, USA 4.College of Textile, Tianjin Polytechnic University, Tianjin 300387, China *Author to whom correspondence should be addressed. Tel: +1-515-294-3012; fax: +1-515-294-6364; e-mail: gwsongsgs@gmail.com Submitted 6 June 2014; revised 8 September 2014; revised version accepted 15 September 2014. Abstract Hot liquid hazards existing in work environments are shown to be a considerable risk for industrial workers. In this study, the predicted protection from fabric was assessed by a modified hot liquid splash tester. In these tests, conditions with and without an air spacer were applied. The protective performance of a garment exposed to hot water spray was investigated by a spray manikin evaluation system. Three-dimensional body scanning technique was used to characterize the air gap size between the protective clothing and the manikin skin. The relationship between bench scale test and manikin test was discussed and the regression model was established to predict the overall percentage of skin burn while wearing protective clothing. The results demonstrated strong correlations between bench scale test and manikin test. Based on these studies, the overall performance of protective clothing against hot water spray can be estimated on the basis of the results of the bench scale hot water splashes test and the information of air gap size entrapped in clothing. The findings provide effective guides for the design and material selection while developing high performance protective clothing. Keywords hot liquid splash; protective clothing; protective performance; skin burn; spray manikin Introduction People encounter various environmental hazards during working scenarios, such as flame, radiation, chemical substances, and thermal stress (Hodder and Parsons, 2007; Ceballos et al., 2011; Lu et al. 2013a). Protective clothing is widely used to protect the wearers from these hazards. The industrial workers in the oil, gas, and petrochemical industries often wear a single layer coverall made of flame resistant material to protect from accidental exposure to flash fires, which might happen when oil or gas leaks are ignited (Crown and Dale, 2005). Workers of oil and gas industry are also at risk of steam, hot water, and drilling mud scald burn injuries (Sati et al., 2008). These could occur if the high temperature drilling fluid splashes during the oil drilling process and the oil with high pressure spouts when the transportation pipeline breaks. Traditional flame-resistant protective clothing usually worn by Published by Oxford University Press on behalf of the British Occupational Hygiene Society 2014. 232

Performance study of protective clothing 233 workers cannot provide adequate protection against hazards of steam and hot liquid splashes. The hazard of hot liquid splashes is quite different from flash fire in terms of temperature, heat and mass transfer modes, and pressure. It has been reported that the highest proportion of burns results from contact and scalds, and 68.7% of the patients are industrial workers (Taylor et al., 2002). During 2001 2010 in USA, there was about 32.7% scald burn injury, approximate to the burn caused by heat and flame (National Burn Repository, 2011). The systematic investigation on the protective performance upon hot liquid splashes is urgent and relatively unexplored in the area of protective clothing for industrial workers and first responders. The current testing standard on protection of materials from hot liquid splashes is ASTM F2701-08 (2008) Standard test method for evaluating heat transfer through materials for protective clothing upon contact with a hot liquid splash. The test method determines whether or not the fabrics or fabric combinations can provide protection to extend the time to second-degree burn injury exposed to hot liquid splashes. Jalbani et al. (2011) have modified the hot water protection device in terms of the liquid spout and delivery system. It was concluded that fabric properties were the key factors in protective performance. The position of the membrane in fabric systems affected the thermal performance ( Jalbani et al., 2011). Air permeability of the fabric was a critical factor in determining the protection against hot water ( Jalbani et al., 2011; Ackerman et al., 2011a; Gholamreza et al., 2013). It was found that minimizing the mass transfer during a hot liquid splash was the most important mechanism to provide a high level protective performance (Lu et al., 2013b; Gholamreza et al., 2013). The protective performance of fabrics exposed to various thermal hazards was investigated and the critical factors influencing thermal protection were identified (Mandal et al., 2013). In addition, the investigation of mass transfer characteristics through protective materials under the same exposure condition has been conducted (Lu et al., 2013b, 2014a). High temperature of the liquid also decreased the surface energy of the fabric, resulting in more mass transfer through the fabric system (Lu et al., 2014a). Furthermore, the relationship between mass transfer and protective performance was discussed in our previous study (Lu et al., 2013b). The effect of liquid flow on the skin burn injury of protective fabrics exposed to hot liquid was investigated (Gholamreza et al., 2012). The inclination of the sensor board affected the heat and mass transfer through protective materials (Gholamreza and Song, 2013; Lu et al., 2014a). The existing standards and studies focused a lot on the thermal performance of the fabric or fabric combinations. These bench scale tests are simpler, quicker, less costly and provide good reproducibility; while the full-scale manikin test is time consuming and expensive but relatively represents actual wear conditions (Lee et al., 2002). Therefore, measuring the heat transferred to the wearers will provide a more useful index of thermal protective performance than bench scale tests (Crown et al., 1998). As known, the air gap plays a very important role in heat transfer. While in the standard or the modified bench scale test, the candidate fabric directly contacts with the cooper or the skin simulant sensor, which cannot capture the effect of air gap on the heat transfer. In our previous study, under exposure to the hot liquid splashes, the air gap between the fabric and sensor significantly decreased the energy transfer to skin, improving the thermal protective performance of the fabric (Lu et al., 2013d). Under the actual wear condition, the air gaps between human body and a garment are not evenly distributed (Kim et al., 2002; Song, 2007; Mah and Song, 2010), namely clothing may be in direct contact with the skin in some areas while hanging loosely in others. In addition, bench scale tests cannot simulate the location of air gaps distributed over the body, nor can they predict the areas of the body that will be burned. How to relate the fabric properties to garment performance has been a challenge up to now. Actually, when the hot liquid is sprayed to protective clothing, it will interact with the fabrics, which has a great influence on heat transfer. Consequently, the angle of the fabric deposition is very important. Although various directions can be simulated on a bench scale test, only one direction can be employed at a time. The high hot liquid splash may apply a compression force to the clothing, resulting in garment deformation and physical properties change. The change may affect heat transfer in protective clothing and its discharge of stored thermal energy. Preliminary full-scale hot water spray tests using a manikin were carried out and the material combinations used in protective coveralls have been compared (Ackerman et al., 2011b). The results demonstrated that impermeable clothing provided better protection.

234 Performance study of protective clothing The effects of fabric properties and design features on thermal protection against hot liquid splashes have been systematically investigated by using the instrumented manikin test (Lu et al., 2013c). In previous studies on heat and flame protection, the correlation between the flame manikin test and the bench scale test was analyzed (Behnke, 1984; Pawar, 1995; Lee et al., 2002). In addition, the protective performance of garment was evaluated by using a copper manikin exposed to steam climate chamber. The tests on fabrics and on garments were in good agreement. Water vapor impermeable fabrics and garments provide greater protection against hot steam. Moreover, the thicker the sample or garment, the higher the protection it provides (Desruelle and Schmid, 2004). However, currently there is a lack of understanding and correlation between full-scale and bench-scale protective clothing tests against hot liquid splashes. Whether the small-scale fabric test can be employed to characterize the thermal protection of protective clothing against hot liquid splash is still unclear, which initiates this study. In this work, the protective performance of full-scale garment exposed to hot liquid spray was investigated using a newly developed hot liquid manikin system (Lu et al., 2013c). Three-dimensional (3D) body scanner was used to characterize the air gap size of the protective clothing. The performance of fabric was evaluated by a modified bench scale test. The air Table 1. Configurations of protective clothing gap of 6 mm was also involved in the bench scale test. The relationship between the bench top test and the manikin test was discussed and a regression model was established to predict the overall percentage of skin burn while wearing protective clothing. MATERIALS AND METHODS Materials and garments A number of protective garments for industrial workers, which were made of permeable, semipermeable, and impermeable fabrics, were selected in this study. Table 1 shows the garment configurations. They were all provided by the thermal protective clothing manufacturers. The garments G1 G8 are size of 42. In order to investigate the effect of garment size on the protective performance, two additional sized garments (size 40 and 44) are manufactured for G2. All the experimental garments consist of a double layer foldover collar and a top fly in the front center as well as a horizontal segment line at the waist. Pockets such as chest patch pockets with flap, rear patch pockets, and in-seam pockets are designed. Reflective tape is also sewed at the shoulder, cuff of sleeve and leg, and back. 3D body scanning The 3D body scanning using a whole body laser scanner (Human Solutions, Germany) was conducted to Garment Fabric Weight (g m 2 ) Air permeability a cm 3 /(cm 2 s) Permeability G1 100% Nomex 169 25.8 Permeable G2 88% cotton/12% nylon 305 5.62 Permeable G3 88% cotton/12% nylon 237 26.9 Permeable G4 88% cotton/12% nylon 305 18.3 Permeable G5 100% cotton 360 2.97 Permeable G6 88% cotton/12% nylon with polymer 412 0 Semipermeable finishing G7 Polyurethane-coated Aramid knit 260 0 Semipermeable G8 Polyvinyl chloride-coated cotton 450 0 Impermeable G9 88% cotton/12% nylon 305 5.62 Permeable G10 88% cotton/12% nylon 305 5.62 Permeable a The fabric air permeability is tested according to ASTM F2701-08 (2004).

Performance study of protective clothing 235 calculate the air gap between protective clothing and the manikin surface (Mah and Song, 2010). A replicate of 40R spray manikin was used for the air gap investigation. Both the nude and clothed scans were required to be aligned as accurate as possible. The nude manikin was scanned first, and then the clothed manikin was scanned with the same condition. Each garment was scanned for three times. In order to reduce error, a specific dress protocol was followed, which involved gently pulling downwards on the waist, sleeves, and leg cuffs of the garments. Pictures of the dressed mannequin were taken and compared across garments of the same style to ensure consistency in dress. A specific procedure was applied to dress the spray manikin to reproduce the air gap distribution as accurately as possible. The 3D scan data were processed in Rapidform XOR (Rapidform, Inc., USA) and the air gap size was determined in Rapidform XOV by a specific protocol as shown in our previous study (Lu et al., 2014b). Spray manikin test The instrumented manikin testing system was used to investigate the protective performance against hot water spray (Lu et al., 2013c), as shown in Fig. 1. The manikin with size of 40R is made from fiberglass and resin. 110 skin simulant sensors are evenly distributed over the manikin surface. The manikin keeps an upright posture and stands on the shutter of square pool that is used to hold the water flow during the test. The manikin torso is sprayed by four sets of cylinder spray jets, which are located at four corners and automatically controlled by the computer. Each set comprises three nozzles. The test clothing was dressed on the manikin followed the dress protocol and compared with photos taken from 3D body scanning to ensure a similar wear condition. During the test, hot water was simultaneously sprayed by the surround twelve nozzles. The hot water was heated to 85 C by superheater and pumped by a motor. The pressure of the hot water spray was set at around 250 kpa to mimic the hot water splashes in industrial work scenarios (Lu et al., 2013c). The exposure time is controlled for 10 s in this study. During the exposure, the heat flux at the surface of the mannequin and its variation with location and time were determined. The temperature of hot liquid spray (<100 C) is very significantly lower than that of a flame engulfment situation (the maximum temperature of flash fire is around 1500 C), and thus the 1 The instrumented hot water spray manikin evaluation system. heat transfer to which the clothing is exposed is much less. While the stored energy in the fabric layers continues to be transferred to the sensor (or human skin) after the exposure, our experimental data showed that this was completed in couple seconds. Therefore, the data acquisition time was set to 60 s to capture the full energy flux, which includes both the exposure time and cooling period. The second-degree burn time and the third-degree burn time of each sensor were predicted using a three-layer skin model and Henriques Burn Integral. The percentage of second, third-degree burn, burn distribution over the manikin, and the total absorber energy were calculated and reported. Three replicates of each garment were tested. Hot water splashes test The bench scale hot liquid splashes tester used in this study was modified based on the device described in ASTM F2701-08 (2008). A spacer was used to create an air gap between the fabric and sensor (Lu et al.,

236 Performance study of protective clothing 2013d). The instrument comprised a reservoir with temperature control, a liquid delivery and circulation system, a sensor board, and a data acquisition system (Lu et al., 2013b). The hot water was controlled to 85 C. The flow control device was used to regulate the flow rate of 40 ± 1 ml s 1 in this study. Two configurations were applied, with and without air spacer. In configuration of without spacer, the testing specimen was directly placed on the sensor board, and with spacer configuration, an air gap of 6 mm between the fabric and sensor board was created. Three skin simulant sensors embedded in the sensor board were used to record the temperature profile and the heat flux during the time course, as shown in Fig. 2. The hot liquid splash exposure time was 20 s and the data acquisition time was 120 s. The second-degree burn time was predicted by the Pennes three-layer skin model and the Henrique s Burn Integral model as described in ISO 13506.3 (2008). The total absorbed energy measured from these three sensors was calculated as well. RESULTS AND DISCUSSION Performance assessment in the bench scale and manikin tests Table 2 shows the second-degree burn time and total absorbed energy during the bench scale tests with and without air gap. The difference in predicted protection provided by the selected fabrics shows significant in both configurations (P < 0.01), moreover, the protective performance provided by permeable fabrics (G1 G5) is worse than those of semipermeable and 2 The modified hot water splash tester. impermeable fabrics (G6 G8). Interestingly, the fabric G7 predicts no burn injury during the test in both cases. The fabrics G5 G8 generate no second-degree burn time in configuration of 6 mm air spacer, whereas the fabrics G1 G4 develop second-degree burn in both cases. In terms of absorbed energy, under both test conditions, it demonstrates more energy transferred through permeable fabrics than those through semipermeable and impermeable fabrics. Under condition of no air gap, however, the selected fabrics can be divided into five categories, whereas with 6 mm air gap, the fabrics show four levels of protective performance. The percentage of second and third-degree burn and total absorbed energy during the hot water spray test for selected clothing were shown in Table 3. Based on these results, the selected protective clothing provides distinct thermal protection upon hot water spray (P < 0.01). The maximum percentage of skin burn developed is G1 (56.23%); whereas no burn predicted for G7 and G8, providing the best performance over the range tested. The predicted percentage of third-degree burn is <10% for all the garments. The total absorbed energy calculated ranges from 16.46 to 171.05 kj m 2. Obviously, the absorbed energy for semipermeable and impermeable garments shows low value, less than 50 kj m 2 ; whereas that for permeable garments presents higher value of 90 kj m 2 during the test. In summary, the semipermeable and impermeable garments show better protective performance than those of permeable garments as the result of mass transfer occurring. Generally, the selected garments can be divided into six categories based on the protective performance. Generally, the bench scale tests with and without an air gap and full-scale spray manikin tests can provide different levels of testing and evaluation. In this study, the second-degree burn time and absorbed energy in the bench scale test shows similar fabric performance in the configuration of no air gap. The total burn injury and absorbed energy also present a similar order on the garment performance. The comparison of bench scale tests with these two configurations indicates that the tests without air gap can differentiate the semipermeable and impermeable fabrics in terms of second-degree burn time and total absorbed energy; the tests with an air gap provide distinct difference for permeable fabrics. Therefore, the bench scale test should consider the

Performance study of protective clothing 237 Table 2. Protective performance of fabrics in bench scale hot water splashes test Fabric No air gap 6 mm air gap Second-degree burn time (SD) (s) Absorbed energy (SD) (kj m 2 ) Second-degree burn time (SD) (s) Absorbed energy (SD) (kj m 2 ) G1 0.41 (0.12) a 408.10 (11.36) a 11.45 (1.28) 282.15 (16.89) a G2 1.61 (0.05) a 396.25 (6.16) a 14.53 (2.05) 205.23 (29.12) b G3 0.65 (0.08) a 410.20 (8.28) a 12.09 (1.42) 269.16 (14.24) a G4 0.82 (0.06) a 399.61 (9.15) a 12.62 (1.19) 250.83 (12.65) a,b G5 4.33 (0.16) b 349.75 (18.75) b 126.43 (19.54) c G6 14.73 (0.51) d 140.17 (14.50) d 69.67 (10.39) d G7 e 53.93 (8.29) e 34.54 (4.59) d G8 12.30 (1.40) c 176.02 (12.44) c 41.15 (3.21) d means no burn predicted. SD is the abbreviation of standard deviation. The subset a, b, c, d, and e in a column means the homogenous value by Tukey s post hoc test. Table 3. Burn injury and total absorbed energy of tested garments Garment Second-degree burn (%) (SD) Third-degree burn (%) (SD) test configurations. In addition, both the bench scale test and manikin test demonstrate that semipermeable and impermeable fabrics and garments exhibit better protective performance than permeable ones, showing in a good agreement. In a previous study, the impermeable fabrics and garments provided higher protection upon hot steam than permeable ones (Desruelle and Schmid, 2004). The correlation between bench scale Total burn injury (%) (SD) Total absorbed energy (kj m 2 ) (SD) G1 48.07 (4.51) 8.17 (4.33) 56.23 (5.60) a 171.05 (7.62) a G2 38.02 (2.12) 1.17 (1.02) 39.18 (1.65) c 132.26 (3.62) c G3 49.63 (1.30) 2.17 (1.93) 51.80 (0.83) a,b 157.51 (2.91) a,b G4 46.87 (0.28) 2.40 (2.26) 49.27 (2.18) b 149.89 (9.16) b G5 16.77 (5.50) 0.00 (0) 16.77 (5.50) d 90.51 (4.54) d G6 6.35 (1.92) 0.00 (0) 6.35 (1.92) e 49.29 (2.71) e G7 0.00 (0) 0.00 (0) 0.00 (0) e 22.00 (2.12) f G8 0.00 (0) 0.00 (0) 0.00 (0) e 16.46 (1.37) f G9 40.63 (2.48) 0.70 (0.62) 41.33 (2.06) c 132.13 (7.04) c G10 34.78 (0.88) 0.65 (0.56) 35.43 (1.28) c 127.51 (6.54) c SD is the abbreviation of standard deviation. The subset a, b, c, d, e, and f in a column means the homogenous value by Tukey s post hoc test. steam test and steam climate test shows a similar trend with that found in this study. Correlation of bench scale test without air gap with manikin test The correlations of different indices obtained in bench scale test without air gap and manikin test are analyzed, as shown in Table 4. The second-degree burn

238 Performance study of protective clothing Table 4. Pearson correlations among different indices T 2 TAE bt TPBI TAE mt T 2 1.0 0.998** 0.914** 0.938** TAE bt 0.998** 1.0 0.899** 0.942** TPBI 0.914** 0.899** 1.0 0.989** TAE mt 0.938** 0.942** 0.989** 1.0 T 2 is second-degree burn time, TAE bt is total absorbed energy in bench scale test, TPBI is total percentage of burn injury, and TAE mt is total absorbed energy in manikin test. *Significant difference at level of 0.05. **Significant difference at level of 0.01. time shows a negative relationship with other indices, whereas the other three indices present positive correlations. It is clear that these four parameters are significantly correlated; therefore further regression analysis is necessary and meaningful. Fig. 3a shows the relationship between the seconddegree burn time and the total absorbed energy during the bench scale hot water splashes test without air gap. The two indices show a strongly negative linear relationship with R 2 = 0.996. The regression is given in equation (1). The fabric that predicts a longer time to reach the second-degree burn decreases the energy transferred and absorbed by human skin during the hot water splashes test. The relationship between total burn injury and total absorbed energy during hot water spray manikin test is analyzed in Fig. 3b. A strongly positive linear relationship with R 2 = 0.976 is observed and the regression is presented in equation (2). If the energy absorbed by human skin is higher, the developed percentage of burn injury will be larger. The results demonstrate that parameters obtained in bench scale and manikin test are well correlated. TAE bt = 19. 381 * T 2 + 422. 02 (1) TAE mt = 2.6131 *TPBI + 26.884 (2) The relationship between the second-degree burn time and the percentage of burn injury is presented in Fig. 4a. These two parameters show a strong logarithmic correlation with R 2 = 0.977. It indicates that the garment with a longer second-degree burn time predicted 3 Correlation analysis of parameters in bench scale test (a) and manikin test (b), respectively. in the bench scale test provides better protective performance during the hot water spray manikin test. Therefore, the increase in protective performance of the fabric can improve the overall performance of the full-scale garment and reduce the percentage of skin burn injury. Fig. 4b analyzes the relationship between the total absorbed energy in the bench scale test and that in the manikin test. The result also demonstrates a strong exponential correlation with R 2 = 0.941. It indicates that the bench scale test and the manikin test are well correlated. It is noted that the origin point is involved by the regression while considering the fact that there should be no absorbed energy predicted during the manikin test when no energy transferred to the skin in the bench scale test. TPBI = 1598. * ln ( T 2 ) + 44. 169 (3) TAE = 17. 2* exp ( 000. 5646 * TAE ) 17. 2 (4) mt bt

Performance study of protective clothing 239 4 Relationship between bench scale test without air gap and manikin test, (a) second-degree burn time versus the percentage of burn injury, (b) the total absorbed energy in the bench scale test versus that in the manikin test, (c) absorbed energy in the bench scale test versus predicted burn injury in manikin test. The second-degree burn time or the total burn injury can intuitively reflect the protective performance of fabrics or garments. It should be noted that the Henriques Burn Integral to predict burn severity levels is not a continuous criterion; even small deviations in data acquisition and wearing condition may cause a different degree of burn injury (Lu et al., 2013b). In some cases, the second-degree burn time or the total burn injury is not predicted, and thus the differentiation of garment performance fails. The prediction of overall burn injury while wearing protective clothing may not be made by equation (3). The total absorbed energy is a continuous index to characterize the heat transfer through the protective fabric systems or garments to human skin. The correlation of percentage of burn injury with total absorbed energy in the bench scale test without air gap is further analyzed, as shown in Fig. 4c. The regression model is established in equation (5). As the absorbed energy is less than 176 kj m 2, the percentage of burn injury is close to zero, and that increases sharply when the absorbed energy is higher than 300 kj m 2. TPBI = 00. 2228 * exp ( 00. 1906* TAE ) 0. 02228 bt (5) Relationship between the bench scale test with air gap and the manikin test Fig. 5a shows the relationship between the total absorbed energy in the bench scale test with 6 mm air gap and that in the manikin test. A strong positive linear relationship with R 2 = 0.988 is predicted and the regression is shown in equation (6). The (0, 0) point is included in the regression as well. In the both cases, as shown in Figs 4b and 5a, the absorbed energy in bench scale test shows a positive correlation with that in the manikin test. Moreover, the two indices present a linear relationship when there is a 6 mm air gap involved in the bench scale test, however, a nonlinear relationship is found as there is no air gap. It confirms that the bench scale test with a 6 mm air gap can linearly represent the full-scale manikin test. TAE = 0. 6116 * TAE (6) mt where TAE bt6 is the total absorbed energy in the bench scale test with a 6 mm air gap. To predict the overall protective performance of protective clothing, the relationship between the total absorbed energy in the bench scale test with a 6-mm bt6

240 Performance study of protective clothing air gap and the total percentage of burn injury in the manikin test is analyzed in Fig. 5b. A significant linear relationship (R 2 = 0.996) is observed and the regression is given in equation (7). Comparing with the case without air gap, this regression equation is linear and shows a good fit. As the absorbed energy in the bench scale test is less than 40 kj m 2, there is no predicted skin burn. TPBI = 0. 2315 * TAE 9. 5613 (7) Prediction of protective performance of garments The garment s thermal protective performance upon hot water spray depends on the fabric properties and garment design features (Lu et al., 2013c). The air gap between clothing and the skin represents the garment bt6 fit and governs the heat transfer to skin. Pearson correlation shows that the average air gap size presents a strong negative relationship with the percentage of burn injury (r = 0.083, P = 0.003 < 0.01). Based on the results found in this study, the second-degree burn time in the bench scale test without air gap and the average air gap size of the garment are employed in multi-linear regression method to estimate the protective performance of protective clothing. The regression equation is presented in equation (8). The regression is significant at level of 0.01 with the R 2 = 0.942, providing a high accuracy. All of the variables in the equation are significant at level of 0.01 by t-test. The comparison of percentage of burn injury predicted by the regression model with that obtained in the experiments is shown in Fig. 6. The absolute difference of skin burn percentage between predicted value and measured value ranges from 1.04 to 7.72. It confirms that the regression model is acceptable and provides believable results. However, the garment G5 gives the largest difference between the predicted value and the measured value (namely 7.72). This might be related to the better performance than other permeable garments due to the large air gap size, whereas the performance of the fabric is similar to other permeable fabrics. TPBI = 97. 671 2. 597 * T 2 1. 84*AAG (8) where AAG is the average air gap size of protective clothing. 5 Relationship between bench scale test with 6 mm air gap and manikin test, (a) absorbed energy in bench scale test versus absorbed energy in manikin test, (b) absorbed energy in bench scale test versus predicted burn injury in manikin test. 6 The measured and predicted percentage of burn injury by equation (8).

Performance study of protective clothing 241 The total absorbed energy in the bench scale test and the average air gap size are used in multilinear regression model to predict the percentage of burn injury of protective clothing. The regression model is significant (P = 0.00 < 0.01) and the R 2 = 0.891. However, only the total absorbed energy significantly contributes to the model by t-test (P = 0.009 < 0.01). The comparison of the percentage of burn injury predicted by the regression model with that obtained in the experiments is shown in Fig. 7. The absolute difference of skin burn percentage between predicted value and measured value ranges from 1.0 to 10.51. The predicted model provides a good prediction of the garment performance. Consequently, this model can be applied to estimate the garment performance if the second-degree burn time is not predicted in the bench scale test. However, the garments G5 G7 that provide better performance give the relatively largest difference between the predicted value and the measured value. This might be related to the nonlinear feature of the skin burn injury prediction method (Lu et al., 2013b). The total absorbed energy in the bench scale test with a 6-mm air gap and the average air gap size are also employed in the multilinear regression model to predict the percentage of burn injury of protective clothing. Although the regression model provides a significant equation (P = 0.011 < 0.05) with R 2 = 0.722, the variables in the model are not significant by t-test. Therefore, this model fails to predict the garment performance. TPBI = 49. 653 + 0. 101* TAE 1. 712 * AAG (9) 7 The measured and predicted percentage of burn injury by equation (9). bt Conclusions The performance of protective clothing upon hot water splashes hazards was evaluated by the bench scale test and the full-scale manikin test. The correlations of the indices obtained in these tests have been analyzed in this study. The correlation analysis shows that the second-degree burn time and the total absorbed energy in the bench scale test without air gap presents a linear relationship. The percentage of burn injury is also linearly related to the total absorbed energy in the manikin test. The results obtained in the bench scale test without air gap correlate well with those obtained in the manikin test, and the relationships show nonlinear, whereas the significantly linear correlations are found when an air gap exists between the test specimen and the sensor. With air gap size of clothing, the results of bench scale test could be applied to accurately predict the percentage of burn injury while wearing protective clothing. The research finding demonstrates that the bench scale test can estimate the overall performance of garment based on the garment air gap size information. To comprehensively evaluate the garment performance and the skin burn distribution, the full-scale hot water spray manikin test is recommended. From the perspective of protection against hot water spray, semipermeable or impermeable protective clothing would be effective. Garments which provide a relatively large air gap size and minimal air gap change will exhibit better performance. FUNDING China Postdoctoral Science Foundation funded project (2014M551657), and Natural Science Research Project for Colleges and Universities in Jiangsu Province (14KJB540001). ACKNOWLEDGEMENTS The authors appreciate the technical support from Stephen Paskaluk in Department of Human Ecology, University of Alberta. We are grateful to the manufacturers who provide the tested garments. The authors declare there are no conflicts of interest in relation to this article. References Ackerman MY, Crown EM, Dale JD et al. (2011a) Project update: protection from steam and hot water hazards. Edmonton, AB, Canada: Protective Clothing Systems for Safety 11.

242 Performance study of protective clothing Ackerman MY, Song GW, Gholamreza F et al. (2011b) Analyzing thermal protective clothing performance against the impact of small splashes of hot liquid. 9th symposium on performance of protective clothing and equipment: emerging issues and technologies. Anaheim, CA: ASTM Committee F23 on Personal Protective Clothing and Equipment. ASTM D737-04. (2004). Standard test method for air permeability of textile fabrics. West Conshohocken, PA: ASTM International. ASTM F2701-08. (2008). Standard test method for evaluating heat transfer through materials for protective clothing upon contact with a hot liquid splash. West Conshohocken, PA: ASTM International. Behnke WP. (1984) Predicting flash fire protection of clothing from laboratory tests using second degree burn to rate performance. Fire Materials; 8: 57 63. Ceballos DM, Yost MG, Whittaker SG et al. (2011) Development of a permeation panel to test dermal protective clothing against sprayed coatings. Ann Occup Hyg; 55: 214 27. Crown EM, Ackerman MY, Dale JD et al. (1998) Design and evaluation of thermal protective flightsuits: part ii: instrumented mannequin evaluation. Cloth Textile Res J; 16: 79 87. Crown EM, Dale JD. (2005) Protection for workers in the oil and gas industries. In: Scott R, editor. Textiles for protection. Cambridge, UK: Woodhead Publishing, pp. 699 713. Desruelle AV, Schmid B. (2004) The steam laboratory of the Institut de Médecine Navale du Service de Santé des Armées: a set of tools in the service of the French Navy. Eur J Appl Physiol; 92: 630 5. Gholamreza F, Ackerman MY, Song GW. (2012) Thermal protective clothing performance: hot liquid splash and its flow effect on skin burn. The 5th ECPC and Nokobetef 10 Future of protective clothing: intelligent or not? Valencia, Spain: Aitex Textile Research Institute. Gholamreza F, Song G. (2013) Laboratory evaluation of thermal protective clothing performance upon hot liquid splash. Ann Occup Hyg; 57: 805 22. Hodder SG, Parsons K. (2007) The effects of solar radiation on thermal comfort. Int J Biometeorol; 51: 233 50. ISO 13506.3. (2008). Protective clothing against heat and flame test method for complete garments prediction of burn injury using an instrumented manikin. Geneva, Switzerland. Jalbani SH, Ackerman MY, Crown BM et al. (2011) Modification of ASTM F 2701 08 apparatus for use in evaluating protection from low pressure hot water jets. 9th symposium on performance of protective clothing and equipment: emerging issues and technologies. Anaheim, CA: ASTM Committee F23 on Personal Protective Clothing and Equipment. Kim IY, Lee C, Li P et al. (2002) Investigation of air gaps entrapped in protective clothing systems. Fire Materials; 26: 121 26. Lee C, Kim IY, Wood A. (2002) Investigation and correlation of manikin and bench-scale fire testing of clothing systems. Fire Materials; 26: 269 78. Lu YH, Li J, Li XH et al. (2013a) The effect of air gaps in moist protective clothing on protection from heat and flame. J Fire Sci; 31: 99 108. Lu YH, Song GW, Ackerman M et al. (2013b) A new protocol to characterize thermal protective performance of fabrics against hot liquid splash. Exp Thermal Fluid Sci; 46: 37 45. Lu YH, Song GW, Li J (2013c) Analyzing performance of protective clothing upon hot liquid exposure using instrumented spray manikin. Ann Occup Hyg; 57: 793 804. Lu YH, Song GW, Li J (2013d) Effect of air gap on the heat transfer through protective materials upon hot liquid splash. Textile Res J; 83: 1156 68. Lu YH, Song GW, Zeng HB et al. (2014a) Characterizing factors influencing impact penetration performance of fabrics used in protective clothing. Textile Res J; 84: 174 86. Lu YH, Song GW, Li J. (2014b) A novel approach for fit analysis of thermal protective clothing using three-dimensional body scanning. Appl Ergonomics; 45: 1439 46. Mah T, Song GW. (2010) Investigation of the contribution of garment design to thermal protection. Part 1: characterizing air gaps using three-dimensional body scanning for women s protective clothing. Textile Res J; 80: 1317 29. Mandal S, Song GW, Akcerman M et al. (2013) Characterization of textile fabrics under various thermal exposures. Textile Res J; 83: 1005 19. National Burn Repository. (2011) 2011 Annual Report Dataset Version 7.0. American Burn Association. Pawar M. (1995) Analyzing the thermal protective performance of single layer garment materials in bench scale and manikin tests. Rayleigh, NC: North Carolina State University. Sati R, Crown EM, Ackerman M et al. (2008) Protection from steam at high pressures: development of a test device and protocol. Int J Occup Saf Ergon; 14: 29 41. Song GW. (2007) Clothing air gap layers and thermal resistance performance in single layer garment. J Ind Textil; 36: 193 205. Taylor AJ, McGwin G Jr, Cross JM et al. (2002) Serious occupational burn injuries treated at a regional burn center. J Burn Care Rehabil; 23: 244 8.