Thermal protective performance evaluation

Many researchers took the initiative to evaluate the thermal protective performance of fabrics and clothing in a laboratory setting [75,76]. For thermal protective performance evaluation, development of a proper type of sensor was critical to estimate heat flux flow through clothed firefighters’ bodies under a particular thermal exposure. Currently, several types of sensor are used within the scientific community for the evaluation of thermal protective performance using bench-scale tests, full-scale stationary, and dynamic manikin tests [331,332].

Development and application of different sensor types

In order to develop a sensor that can measure heat flux flow through human (firefighters’) bodies, numerous specific features need to be considered [333]. The developed sensor must be (1) a close representation of human skin, (2) small and light-weight, (3) inexpensive, (4) accurate (within ±10%) and reliable to measure convective and radiative heat flux in an operating range of 0-105 kW/m2, (5) rapid for proper data acquisition, (6) immune from noise produced under the convective and radiative heat exposures, (7) highly sensitive to detect heat flux in the lowest operating range, with only slight variation due to heat leakage/loss or thermal storage within the sensor, (8) designed to minimize the storage of thermal energy within the sensor under repetitive and highly intensive thermal exposures, (9) rugged/durable enough to withstand repeated thermal exposures and cleaning, (10) minimally impacted by the thermal history (heat sink, temperature gradient) of any overlaying materials used in the sensor, and finally (11) easy and economical to fabricate and/or easily available in the market [331,333-336]. Keeping these features in mind, several sensors have been developed: the thermal protective performance (TPP) sensor (copper slug sensor), the embedded sensor (thin-skin sensor), the skin simulant sensor, the NCSU PyroCal sensor (new copper slug sensor), and the water-cooled sensor [333,337]. In the following section, each of these sensors is thoroughly discussed and they are compared to identify the most suitable sensors for close study of thermal protective clothing. Furthermore, application procedures of these identified sensors in the context of thermal protective clothing are discussed below and a few case studies of their application are also highlighted.

TPP sensor (copper slug sensor): In this sensor, a copper disk of 1.6 mm in thickness and 40 mm in diameter is used (Fig. 5.4). The thickness is determined so that the temperature increase of the disk approximates the temperature increase of human tissue under a particular thermal exposure; and the large diameter of the disk is preferred to conveniently monitor its surface temperature in a wide range [331,332]. On the back surface of the disk, four 30-gauge type-J (iron-constantan) thermocouples are uniformly secured in a parallel manner at 120 degree intervals and at the center to measure the average front surface temperature of the disk. This average temperature is further used to determine the heat flux through the sensor. The four thermocouples are used to average out any variation of the front surface temperature of the disk; the 30-gauge thermocouples are employed because they are large enough to work with, and small enough to minimize, heat loss. The uniform distribution of the thermocouples can cover the entire surface of the disk and the parallel configuration can average the voltage of all four thermocouples. The entire disk with thermocouples is mounted in an insulating block. The front face of the disk is blackened to approximate its emissivity characteristics to that of human skin. In this sensor, the heat flux under a particular thermal exposure is calculated using Eq. (5.1), where, q = heat flux (cal/s/cm2), M=mass of the disk (g), Cp = specific heat of the disk (cal/g °C), A = area of the disk (cm2), ДТ = temperature rise of the disk (°C), and At = exposure time (s). M can be further represented by Eq. (5.2), where, A = area of the disk (cm2), b = thickness of the disk (cm), and p = density of the disk (g/cm3). Eventually, Eq. (5.1) can be rewritten as Eq. (5.3), which can be conveniently used to calculate heat flux through the sensor. According to Eq. (5.3), it is clear that the p and Cp are constant; consequently, q is directly dependent upon b and ДТ/At. It seems that the accurate measurement of ДТ/At is essential to precisely calculate q, and b is the most important affecting parameter to accurately measure ДТ/At or q. The result of the TPP sensor can also be used to predict human skin burns based on the Stoll Curve (Fig. 3.2) [129,131,154,155]. TPP sensor

Fig. 5.4 TPP sensor.

Adapted from S. Mandal, G. Song, Text. Res. J. 85 (1) (2015) 101-112.

This type of sensor is highly reliable, accurate, rugged for repetitive thermal exposures, durable during intensive thermal exposures, and can also withstand long thermal exposures. As a consequence, this sensor is widely accepted within the scientific community. However, this sensor does not factor in potential heat loss under a particular thermal exposure when calculating heat flux. Hence, the calculated heat flux from this sensor can be overestimated [331].

Embedded sensor (thin-skin sensor): The embedded sensor was developed by the DuPont Company with the Thermo-Man flash fire manikin (Fig. 5.5). This type of sensor is constructed using a thermo-set polymer resin that is cast into a small solid cylindrical-shaped plug [35,332,334]. This polymer resin exhibits thermal inertia (a product of density, thermal conductivity, and specific heat) similar to that of undamaged human skin. In this skin-simulant resin-cast plug, a type-T thermocouple (copper-constantan) is embedded just below the front surface of the resin-cast plug (at a depth of 0.127 mm). This thermocouple measures the front surface temperature of the resin-cast plug under a particular fire exposure; this temperature is later used to evaluate the heat flux flow through the sensor. The resin-cast plug is designed with

Embedded sensor

Fig. 5.5 Embedded sensor.

Adapted from S. Mandal, G. Song, Text. Res. J. 85 (1) (2015) 101-112.

a 26 mm diameter and 27 mm thickness. The sensor can be treated as infinite slab geometry for certain durations of fire exposures.

This type of sensor is rapid and accurate for properly measuring the heat flux under an intensive thermal exposure. However, the measurement of heat flux by this sensor is highly dependent on the exact location of the thermocouple. Because determining the exact location of the thermocouple is usually uncertain, this type of sensor has an inherent tendency to produce inaccurate results [35,334]. Additionally, this sensor is not very durable, because the used polymer resin may crack under repetitive thermal exposures [331].

Skin simulant sensor: This sensor, developed by researchers at the University of Alberta, Canada (Fig. 5.6), behaves similarly in heat transfer to human skin [24,331,337]. This sensor is made up of an inorganic material called colorceron, a mixture of various compounds such as calcium, aluminum, silicate, asbestos fiber, and a binder. This inorganic material does not have the same values of density (p), thermal conductivity (k), or specific heat (Cp) when compared with human skin, but the thermal inertia [a product of p (kg/m3), k (W/m °C), and Cp (J/kg °C) or thermal absorptivity (a square root of thermal inertia)] of the material is similar to that of human skin. As thermal inertia is the most important property of a material in determining heat flow in sensor development, colorceron is used for the skin simulant sensor. In this sensor, a type-T thermocouple (copper-constantan) is held on the surface of a colorceron slab by an epoxy-phenolic adhesive that can tolerate the temperature up to 370°C. A hole is drilled along the length of the colorceron slab to allow the thermocouple to permeate inside the sensor. The thermocouple’s attached surface is generally painted black to control the emissivity of the sensor. The thermocouple measures the temperature increase in any intensified thermal environment and this increase is used to calculate the heat flux through the sensor using Duhamel’s theorem [Eq. 5.4, where, Ti = initial uniform surface temperature (°C), Ts(t) = surface temperature (°C) at time t (s), q"(t) = heat flux (W/m2) at time t] [36,335]. The length and diameter of this sensor are 32 mm and 19 mm, respectively. The length of the sensor is chosen for convenience because the colorceron is commercially available in 32 mm thick slabs; the diameter is selected such that sufficient lead length could be given to the thermocouple on the surface for eliminating the conductive heat transfer effect at the junction of the colorceron and thermocouple.

Skin simulant sensor

Fig. 5.6 Skin simulant sensor.

Adapted from S. Mandal, G. Song, Text. Res. J. 85 (1) (2015) 101-112.

This type of sensor is rapid and accurate for properly measuring the heat flux under a short-duration intensive fire exposure. This sensor is widely used to evaluate the performance of protective clothing using the Harry Burns full-scale instrumented flash- fire manikin test. However, this sensor does possess a few drawbacks, one of which is that it cannot withstand long duration (>120 s) exposure at high heat flux [331]. If the measuring surface of this sensor is not properly painted black, the thermocouple bead (placed on the surface of the sensor) will be directly exposed to fire, and it may reflect some amount of radiative energy away from the surface of the sensor, resulting in a lower sensor response rate.

NCSU PyroCal sensor (new copper slug sensor): Recently, the TPP sensor was modified and improved by Grimes [333] from NCSU. This improved sensor is called the PyroCal sensor (Fig. 5.7). The PyroCal sensor is thermally insulated; hence, it causes less heat loss than the noninsulated TPP sensor under fire exposure. As the mass of the insulated PyroCal sensor (1.3 g) is much lower than that of the TPP sensor (17.9 g), heat storage within the PyroCal sensor under repetitive and highly intensive fire exposures is less compared to the TPP sensor. Overall, the PyroCal sensor demonstrates more reliability, repeatability, and versatility than the traditional and extensively used TPP sensor [35,331].

PyroCal sensor

Fig. 5.7 PyroCal sensor.

Adapted from S. Mandal, G. Song, Text. Res. J. 85 (1) (2015) 101-112.

In the PyroCal sensor (as shown in Fig. 5.7), a copper 110 alloy (density = 8910 kg/m3; specific heat=0.385 kJ/kg) disk (about 12.7 mm in diameter and 1.5 mm in thickness) surrounded by a radial thin copper ring (ring acts as a thermal guard) is mounted in an insulating block (diameter 26.3 mm and thickness 26.6 mm) to minimize heat transfer to and from the body of the disk. This design approximates a one-dimensional energy flow through the sensor. The copper disk with ring is secured within the insulating holder by three pins, located 120 degree apart, perpendicular to the central axis and 1.02 mm back from the front surface of the disk. The pins are inserted through 1.57 mm holes in the insulating holder and the copper disk, and seated into conical notches (0.76 mm diameter and

0.38 mm deep) at the front surface of the disk in order to reduce the amount of heat loss from the disk. A 6.35 mm air gap is kept behind and on the side of the disk; this insulates the back and side faces of the disk. The disk temperature is measured by a 30-gauge type-T thermocouple (copper-constantan), affixed atthe backof the disk; this temperature is used to assess the heat flux through the sensor. As described by Grimes [333], a flat base hole of 1.19 mm diameter and 1.02 mm depth is drilled into the rear surface of the disk and into this hole two leads of the thermocouple are inserted and secured in place by an 18 gauge copper plug. The pressure exerted by the plug results in an intimate contact between the disk and thermocouple, which can improve the sensor’s accuracy reliability. After the proper attachment of the disk and thermocouple, a 3.18 mm length of heat-shrunk tubing is applied to the remaining length of the thermocouple to protect it. To secure the thermocouple, it is further fed through a strain relief tube mounted into the insulating disk-holder. The entire assembly (copper disk with ring, insulating holder, and thermocouple) is then inserted and secured with #10-32 nuts within a protective shell made from the thermo-set polymer resin used in the embedded sensor. This shell protects the insulating holder and keeps the copper disk with ring in place. The front surface of the disk is usually painted with a 0.025-0.038 mm layer of low gloss, high-temperature black enamel paint to protect the surface and raise its diffuse emissivity to a value close to 1.0 under a particular fire exposure.

Based on the above sensor design, heat received at the front face of the copper disk is equal to the energy conducted axially within the sensor. However, it is impossible to keep the disk in a perfect insulator; thus, some amount of heat transfers to and from the disk and results in heat loss. It is assumed that the incident (q" incident) heat (W/m2) on the sensor under any intensified thermal environment can lose through convection (q" convection), conduction (q" conduction), and/or radiation (q" radiation). Simultaneously, some amount of heat does store (q" storage) inside the sensor (Fig. 5.8).

Heat loss through PyroCal sensor

Fig. 5.8 Heat loss through PyroCal sensor.

Adapted from S. Mandal, G. Song, Text. Res. J. 85 (1) (2015) 101-112.

As per Fig. 5.8, the thermal energy balance of a PyroCal sensor can be written as

where p = density of the copper disk (kg/m3), Cp = specific heat of the copper disk (J/kg °C), dL = length of the heat exposed area (m), and dT/dt = change of temperature with respect to time (°C/s). Thus,

The convective and radiative losses can be defined as a Newtonian cooling term and a function of temperature to the fourth power, respectively. The convective and radiative losses can be represented by Eqs. (5.8), (5.9), respectively, where, h = convective heat transfer coefficient (W/m2K); e = emissivity of the copper disk; a = Stefan- Boltzmann constant (W/m2 K4); T(t) = temperature of the copper disk (°C); and Tair = temperature of the outside air (°C). Moreover, the conductive loss can be broken down into two distinctive components: radial and axial, as shown in Fig. 5.9. Assuming that there is no contact resistance, the conductive loss can be calculated based on Eq. (5.10), where k = thermal conductivity (W/m °C) of the gasses in Lgap (m). By entering the values of Eqs. (5.8)-(5.10) into Eq. (5.7), the incident heat of heat flux value is calculated; this final equation for heat flux calculation is Eq. (5.11).

NCSU water-cooled Prototype sensor

Fig. 5.9 NCSU water-cooled Prototype sensor.

Adapted from S. Mandal, G. Song, Text. Res. J. 85 (1) (2015) 101-112.

The PyroCal sensor possesses superior features such as simplicity in design, small size, ease of operation, durability under repetitive fire exposure, accuracy under highly intensive exposures, and easily available materials (eg, copper disk and ring, insulating holder, thermo-set polymer resin). Thus, this sensor is extensively used for evaluating the performance of firefighters’/industrial-workers’ clothing using the “PyroMan” full-scale instrumented flash fire manikin test. However, the application of this sensor possesses several drawbacks [27,331]. For example, this sensor does not absorb heat accurately in the same fashion as human skin in long duration exposures [27]. Actual human skin temperature rises faster than the PyroCal sensor; thus, the calculated heat flux by the sensor may inaccurately represent heat transferred into skin.

Water-cooled sensors: It can be identified from the above-mentioned sensors that they are all unsuitable for use in long duration fire exposures. To overcome this problem, the water-cooled sensor is used. The water-cooled sensor is an example of a classic sensor, which uses a novel dynamic cooling technology for its development. Three different types of water-cooled sensors are available: The Schmidt-Boelter sensor, the Gardon sensor, and the Prototype sensor [334,337-339]. The Schmidt-Boelter and Gardon sensors are standardized by ASTM and NIST and the Prototype sensor is developed by NCSU. Normally, the measuring thin constantan foil or copper disk faces of these sensors (placed and secured in a disk holder using sealants) are painted black to properly absorb all the radiant heat, especially when exposed to any intensified fire exposure. The Schmidt-Boelter sensor operates by measuring the axial temperature differential between the front face of the disk and a water-cooled copper or aluminum heat sink with thermopile. The Gardon sensor operates by measuring the radial temperature differential between the front face of the disk and a water-cooled copper heat sink with thermopile [338]. In the Prototype sensor, the temperature differential of water flowing in (coolant inlet) and out (coolant outlet) of the area under the copper disk is measured (Fig. 5.9). The temperature differential is used to determine the heat flux within these sensors. In this context, Eq. (5.12) is developed to calculate the heat flux by the Prototype sensor, where, q" = total heat flux (kW/m2), pcu = density of the copper disk (kg/m3), CPcu = specific heat of the copper disk (kJ/kg K), tcu = thickness of the copper disk (m), Acu = area of the copper disk (m2), Tcu = temperature of the copper disk (K), t = time step of the experiment (s), mh2o = the mass flow rate of the coolant (kg/s), CPh O = specific heat of the coolant (kJ/kg K), TI = incoming coolant temperature (K), TO = exit coolant temperature (K), hH = heat transfer coefficient of the sensor housing (kJ/s m2 K), AH = wetted area of the sensor housing (m2), and TH = temperature of the sensor housing (K).

A water-cooled sensor has a built-in cooling system that removes the energy absorbed by the disk during an exposure; this heat removal process prevents an overload in disk temperature during exposure. This allows the disk to maintain a constant temperature gradient even with continuing exposure of the heat source. This means the water- cooled sensor can function for much longer duration of exposures than the other nonwater-cooled sensors (TPP, embedded, PyroCal, skin simulant) [331,337,340]. However, the water-cooled sensor is unsuitable for measurement of rapid, intensive heat transfer because of the thermal inertia of its cooling system and its relatively slow response time; the capability and accuracy of this sensor is also limited to intensity and duration of exposure. In addition, the accuracy of any water-cooled sensor is directly associated with the characteristics of its water cooling system. Thus, there is a need to precisely evaluate the heat evacuated by the sensor’s cooling system in order to accurately assess the heat flux. Other significant drawbacks associated with this type of sensor are that it is expensive and cumbersome to use [334]. Due to all these drawbacks, water-cooled sensors are not preferable for regular commercial use; however, this sensor is acceptable for use as a calibrator for other sensors [35,334,340].

Types of sensors and their applicability: The TPP, embedded, skin simulant, and PyroCal sensors operate using thermocouple technology; however, the water-cooled sensors work based on dynamic cooling technology. Embedded and skin simulant sensors measure temperature increases of predetermined duration by thermocouples, and these temperatures are used to determine heat flux (Eq. 5.4). Because embedded and skin simulant sensors require an inverse heat transfer calculation to estimate the heat flux from temperature, these sensors may cause an estimation error associated with heat flux calculations. The TPP, PyroCal, and water-cooled sensors directly measure the heat flux (Eqs. 5.1,5.11,5.12), thus avoiding miscalculations of heat flux as is the case with embedded and skin simulant sensors [331,336,340]. Among the sensors, embedded and skin simulant sensors represent a human skin model (Fig. 5.10); eventually, the heat transmission behavior within these sensors is similar to that of human skin [23,26,341]. As embedded and skin simulant sensors work according to the skin model, the temperature readings of these sensors under fire exposures can be used to calculate the time required to generate skin burn injuries based on the Henriques burn integral equation (Eqs. 3.1, 3.2), which can be used to gauge the performance of thermal protective clothing [125,129]. Table 5.1 summarizes the advantages and disadvantages of these sensors in protective clothing studies.

Considering the above applicability of different sensors, several standards have been developed by international and national organizations, such as ASTM, NFPA (National Fire Protection Association), ISO (International Organization for Standardization), and CGSB (Canadian General Standard Board). These test standards (eg, ASTM F 1939:2008, ASTM F 2702:2008, ASTM F 2703:2008, ASTM F 1930:2013, NFPA 1971:2013, ISO 9151:1995, CGSB 155.20:2000) are mainly used to evaluate the thermal protective performance of fabric or clothing under a particular fire exposure using bench-scale tests or full-scale manikin tests [137,138,342-347]. In these standards, the sensors are used in bare and clothed conditions for two main purposes: (1) heat flux calibration in bare conditions, and (2) heat flux measurement in clothed conditions. For these tests, the sensors measure the heat transferred through them under radiant heat and/or flame exposures. Subsequently, the analog output signals from these sensors are fed into an analog amplifier multiplexer to produce a digital voltage signal. Next, these digital voltage signals are fed into a software program to translate into temperature readings every second; these temperature readings are simultaneously used to determine heat flux.

Table 5.1 Advantages and disadvantages of various thermal sensors

Thermal sensors



Skin simulant




  • • Reliable and accurate
  • • Rugged for repetitive fire exposures
  • • Durable in intensive fire exposures
  • • Simulates human skin
  • • Rapid and accurate in intensive fire exposures
  • • Simulates human skin
  • • Rapid and accurate in highly intensive, short-duration exposures
  • • Simple design
  • • Small in size
  • • Ease of operation
  • • Relatively rugged under repetitive fire exposure
  • • Accurate under highly intensive exposures
  • • Suitable for calibration purposes of the other sensors: TPP, embedded, skin simulant, PyroCal
  • • Functions for longer duration than other sensors in radiant heat exposures


• Uncertain heat loss from this sensor renders it

unsuitable for low- intensity and long time exposure

Calculated heat flux from this sensor can be underestimated in long duration exposures

  • • Calculated heat flux is highly dependent on exact location of thermocouple within the sensor
  • • Nondurable under long duration and/ or repetitive fire exposures
  • • Cannot withstand long exposures at high heat flux
  • • Directly exposed thermocouple may result in inaccurate heat flux under intensive fire exposures


absorption dissimilar to human skin in long duration fire exposures

  • • Unsuitable for measurement of rapid, intensive heat transfer
  • • Relatively slow response time
  • • Limited capability and accuracy
  • • Expensive and cumbersome to use
Skin model for a sensor

Fig. 5.10 Skin model for a sensor.

Case studies on the applications of different sensor types: In most of the above- mentioned standards, TPP sensors are used. However, many researchers also conducted their studies using other sensors; these studies are useful to analyze the prediction variability of sensors depending upon the intensity of fire exposure and clothing used for testing [331,332,340]. In this context, Barker et al. [331] compared the heat flux readings obtained from four sensors (TPP, embedded, skin simulant, and PyroCal) in both bare/nude (calibration) and clothed conditions under different intensive fire exposures. In the case of bare sensors, three different types of fire exposures were applied: (1) 100% radiant heat source at 0.14 cal/cm2/s for 20 s, (2) 100% radiant heat source at 0.30 cal/cm2/s for 20 s, and (3) 50/50 convective/radiant heat source at 2 cal/cm2/s for 5 s. Only a single fire exposure (50/50 convective/radiant heat source at 2 cal/cm2/s for 10 s) was selected for the clothed sensors. For the clothed conditions, the front surfaces of the sensors were covered with specimens of three different fabrics (Nomex-III of 5.33 oz/yd2, fire-retardant cotton of 6.93 oz/yd2, and fire-retardant wool of 10.23 oz/yd2). In this comparison, the bare PyroCal sensor corresponded closely with the bare TPP sensor under all three types of fire exposures, as both were copper slug sensors. On the other hand, the heat flux reading of the bare skin simulant sensor could not be compared to the bare TPP sensor in a fire exposure of 2 cal/cm2/s, because the skin simulant sensor reflected some of the radiative energy away from the surface of the sensor. The bare embedded sensor did provide a reasonably close estimate of average heat flux with the bare TPP sensor for all three fire exposures; however, there was considerable variation in heat flux readings between individual bare embedded sensors because the position of thermocouples can be different in these sensors. Additionally, both the bare PyroCal and embedded sensors provided a consistent average reading of heat flux in repeated measurements for the exposure conditions tested. Nevertheless, the PyroCal sensor data was more repeatable than the embedded sensor data. Furthermore, this study demonstrated that the clothed PyroCal sensor responded quickly and accurately to calculate the heat flux under a particular fire exposure in comparison with the clothed embedded or skin simulant sensor. This is because the PyroCal sensor provided a larger integrating measurement surface area to promptly monitor the bulk heat transfer through the fabric specimen. This measuring surface can also even out the constructional variation in the specimen and give an accurate average for the heat transmitted through the interstices (made by the intersection of warp and weft yarn) in the tested fabric specimen. On the other hand, the buried or surface mounted thermocouple in the clothed embedded or skin simulant sensor can reflect the heat away from the surface of the sensor and this reflection can slow the response to calculate heat flux. It is also notable that most of the bare or clothed sensors lose a significant amount of energy under fire exposures; hence, there is a need to accurately measure a calibration factor to correct the heat loss from the sensors. Barker, Hamouda, and Grimes [332] determined the heat flux through each of four sensors in clothed conditions under a single fire exposure for 5 s (50/50 convec- tive/radiant heat source at 1.3 cal/cm2/s). Here, the surface of the sensors were covered with a poly(p-phenylene terephthalamide)/polybenzimidazole fabric (Kombat 750 of 7.5 oz/yd2). After the 5 s exposure, the heat flux values of these sensors were continually determined for another 5 s with or without any fabric. In this case study, it was found that the clothed TPP, skin simulant, and PyroCal sensors measured similar levels of heat flux both during and after exposure when tested with a fabric after the exposure. However, these sensors measured similar levels of heat flux only during exposure when tested without a fabric after the exposure. It was also found that the skin simulant sensor measured the highest heat flux (^0.6 cal/cm2/s), whereas, the embedded sensor measured the lowest heat flux (^0.4 cal/cm2/s). In this case study, it was suggested to carefully study the repeatability and durability of the sensors under different fire exposures along with their heat flux estimation.

< Prev   CONTENTS   Source   Next >