Printer Friendly

Measurements of pipe insulation thermal conductivity at below ambient temperatures Part I: experimental methodology and dry tests.


Heat exchangers inside buildings are often adopted to provide air conditioning, cooling and dehumidification to the zones of buildings. A cold fluid, usually water or water-glycol solutions, is used to transport and distribute the refrigerating capacity from central chillers to the liquid-to-air heat exchangers and a network of chilled pipes is installed inside the building structure. The surface temperature of these pipelines is commonly at 40 [degrees] F (4.5 [degrees] C) and below the dew point temperature of the surrounding ambient air. Mechanical insulation systems are installed around these cold pipes to limit the heat gain in the pipelines and to prevent moisture condensation on their surface. Insulation jackets, vapor retarders, and vapor sealing of the joints and fittings are normally adopted to create a barrier to the moisture ingress into permeable insulation. However, experience shows that mechanical pipe insulation systems are not completely vapor tight and inevitably moisture accumulates in permeable insulation (ASHRAE 2009). The fact that the pipe surface temperature is below the ambient room temperature has three main implications:

1. there is a temperature gradient across the insulation system that drives a radial inward heat flow, that is, sensible heat is leaking in from the ambient to the cold pipe through the cylindrical shaped insulation material

2. there is a gradient of the water vapor partial pressure across the pipe insulation and moisture diffuses in the permeable insulation material toward the cold pipe surface. If the surface temperature of the pipe is below the dew point temperature of the ambient air, moisture might condense at the cold pipe surface and form water droplets that accumulate and drip onto the building surfaces

3. the moist flow also represents an additional heat flux leaking into the insulation system and it decreases the insulation thermal conductivity of the system (Guldbrandsen et al. 2011)

To limit moisture accumulation few papers proposed to continuously remove the condensate from the pipe surface by utilizing a wicking action of hydrophilic fabrics (Crall 2002) and (Korsgaard 1993). While this approach has some value it is generally recognized that the mechanical insulation systems for cold pipe suffer from moisture accumulation issue and the ASHRAE Handbook mentioned to such issue since 1980s. Currently the standard ASTM C335 (2010a) is used to determine the thermal conductivity of cylindrical shaped pipe insulation systems. However, this standard is based on a heated pipe, with the heat flow outward, and it is generally for use in above room temperature applications. When testing at ambient temperatures below the normal room temperature, theoretical analysis shows that the experimental heat flow direction is unimportant for a perfect homogenous material but if the properties of the insulation vary in the radial direction, the experimental heat flow direction will affect the measured thermal conductivity (Wilkes et al. 2002). In addition, if the pipe is above the room temperature moisture accumulation and water vapor condensation phenomena are virtually absent since the moisture might be driven outward, that is, from the pipe surface to the ambient. Another approach in the literature is to approximate the effective thermal conductivity of the materials used in pipe insulation systems with the thermal conductivities of materials for insulation panels (ASTM 2010b, 2010c). Unfortunately the radial configuration and longitudinal joints typical of pipe insulation systems can produce different thermal properties than slabs.

To address the above mentioned challenges, a novel experimental apparatus that measures the actual thermal conductivity of mechanical pipe insulation systems in cold pipe applications was developed. This paper introduces the novel test set up, describes the laboratory methodology, and discusses the experimental results of the actual pipe insulation thermal conductivity at below ambient temperatures and in dry non-condensing ambient conditions.


Thermal conductivity is an important property for the proper selection of the pipe insulation systems. The thermal conductivity of pipe insulation depends on several factors including material density, porosity, type of structure (fibrous and cellular foam type), temperature of the insulation, moisture content, method of fabrication and cutting, thickness, and time if the material is subjected to aging effects, that is, if CFCs gases initially present in the insulation material start to leak out and are replaced by air as time passes. Preventing conduction and convection is the basic physics RI insulation materials. To decrease the conduction certain types of insulation provide severe barriers to the thermal path between the hot and cold sides. Other types of insulation material inhibit convection by trapping air molecules in the micro voids inside the material itself.

Budaiwi et al. (2002) and later Abdou and Budaiwi (2005) studied several insulation materials with the aim to identify the temperature effects on the actual thermal conductivity of the insulation. Their findings seem to suggest that the k-values, with k being the measured thermal conductivity of the material, have a strong dependence on the temperature and their observations indicated that measured values at room conditions of 75 [degrees] F (25 [degrees] C) should not be used for below ambient temperatures. Wilkes et al. (2002) measured the thermal conductivity of two types of foam insulation and the results showed that the thermal conductivity of pipe insulation lead to a 1.5% to 2.5% lower value than the ones from insulation boards. They pointed out that for materials subjected to aging effects, the difference in the thermal conductivity might have been caused by different ways of diffusion between board-type and pipe-type insulation specimens. Whitaker and Yarbrough (2002) tested five types of pipe insulation systems commonly used from 1982 through 1997, with specimen mean temperatures between 100 [degrees] F (37 [degrees] C) to 1000 [degrees] F (537 [degrees] C). The effect on thermal conductivities caused by various material physical properties, like thickness, density and methods of installation were evaluated on the test apparatus, which was based on a 3 inches (76.2 mm) heating pipe, with three heating circuits controlled for the metering area and end guards. Wang et al. (2000) investigated half C-shell insulation specimen and the results indicated that a sizeable amount of heat flow occurred through the insulation small gaps. This effect was estimated to be difficult to avoid due to the installation method typical of pipe insulation systems. The effect of gap voids were numerically studied and discussed in their work.

Two methodologies are commonly used to determine the thermal conductivity of insulation systems: a guarded plates method and a thermal conductivity probe. These approaches are summarized next.

Hot and Cold Guarded Plates Approaches

As early as 1970's, Howard et al. (1973) presented technical details of an apparatus to measure the thermal conductivity by using two constant temperature plates: one at cold and one at hot temperature, respectively. McFadden (1986) explained this method in more detail a few years later. The guarded plate technique employed in their work consisted of a hot plate which operates at 212 [degrees] F (100 [degrees] C) and a cold plate which was set at 122 [degrees] F (50 [degrees] C). The idea was to create a controlled thermal gradient across the insulation specimen. Wijeysundera et al. (1989) used the guarded plate method, placing a piece of test insulation sandwiched between the hot and cold plates. Two heat flow meters were located within a square cut of a rubber sheet. The heat flux measuring pads were guarded with rubber sheet to reduce lateral heat flow into the heat flow meter. During the wet test, certain amount of water would be sprayed onto the surface of the test insulation with an atomizer. From the experimental data on flat fibrous insulation, the authors concluded that the equivalent thermal conductivity of the slab in presence of condensation had a linear trend and increased with increasing of mean temperature and temperature difference across the insulation specimen.

Log (1991) used the hot-strip method for determining thermal conductivity and thermal diffusivity, similar to Gustafsson et al. (1978), Gustafsson et al. (1985) and Saxena et al. (1989). A 1 inch (25.4 mm) thick, 5/16 inch (8 mm) wide and 2 3/4 inches (70 mm) long iron strip heater was sandwiched between two halves of insulation samples. The thin metal foil was used as a constant heat source and a sensor of temperature. The method was reported to have an accuracy within 5% for the thermal conductivity.

Hay (1984) pointed that one difficulty in making the thermal conductivity measurements by using hot plate guard approach was the length of time required for samples to reach equilibrium. As mentioned before, he suggested the orientation of the sample also should be considered as variable. With additional tests, he concluded that equilibrium would be reached at different times based on the locations of high moisture and low moisture regions.

Thermal Conductivity Probe Approaches

Compared to such instruments which can only measure the overall thermal property, a thermal conductivity probe can be inserted into specific locations in the test insulation. This sensor has a self-heated thermistor located at the tip of the probe. By measuring the energy dissipation and temperature difference, thermal conductivity at certain point within the insulation can be determined. When moisture gradually enters the insulation, it may vary from one side to the other. At this point, the thermal conductivity sensor might fail since its sensitivity might not be able to estimate the overall change of the insulation. Batty et al. (1984) presented a detailed description of the probe commonly used by the manufacturer, and the thermal-probe technique was assessed on wet clay and moist masonry materials.

Thermal conductivity probes might introduce some systematic errors when applied in extreme conditions. Yu et al. (2009) used the thermal probe to find the moisture content effects on the uncertainty of sand thermal conductivity. They reported that with moisture content higher than 25% in volume, thermal conductivity is easier to be measured. Due to their operating principles, thermal probes generate a heat flux during the measurements. With high water content, the water evaporation rate in the region next to the sensor might be low, and because of the capillary force, water might flow toward the sensor in the adjacent region. In such steady state conditions, the thermal conductivity could be measured accurately with a series of thermal probes. However, with low moisture content, the region next to the sensor might easily dry

out and a systematic error in the measurements is introduced by the local heat flux generated from the probes themselves.

Heat Flux Meter Methods for Measuring the Thermal Conductivity of Pipe Insulation

In recent years, several laboratory methodologies were developed for measuring flat plate and pipe insulation thermal conductivity. The ASTM (2010a) standard provides a standard test method for steady-state measurements of pipe insulation. The technique is based on hot pipe with an outward heat meter principle. The core heater creates a measurable heat flow that is dissipated through the cylindrical test section. In order to obtain good accuracy, the axial temperature imbalance is no greater than 0.5% of the radial temperature drop across the test specimen, thus, the radial temperature differences must be at least 18 [degrees] F (10 [degrees] C) (ASTM 2010a). The end guard sections should be long and well insulated to minimize the axial heat conduction effects at the end sides. Different types of heaters were suggested in the standard to guarantee uniform surface temperature so that the axial heat flow can be limit within 1% of the total heat transfer on the test section. Chyu et al. (1997a, 1997b) and Wilkes et al. (2002) developed new pipe insulation apparatus based on a similar heat meter principle. Although the core section of the pipe insulation apparatus was colder than ambient room temperature, the test methodology of Wilkes et al. (2002) did not consider the thermal conductivity of pipe insulation systems with the moisture ingress and the inner tube surface temperature was higher than the room air dew point temperature. Chyu et al. (1997a, 1997b) proposed to simulate field moist absorption conditions by immersing the pipe insulation specimens in a water tank. Their proposed method could not be used to study the thermal conductivity of pipe insulation systems under actual operating conditions of chiller pipes but could provide some estimates of the maxi-mum amount of water that a particular insulation material is able to hold.

Other Techniques Used to Measure Thermal Conductivity at Below Ambient Temperatures

Fleischmann et al. (2003) and Hao et al. (2004) presented a contactless technique for thermal conductivity measurements at low temperatures. Instead of using electric wires/heaters, the temperature gradient in the samples was generated via optical heating. Two magnetization thermometers were used for temperature readouts. Since the heat sink was able to apply a power dissipation lower than 3.412x[10.sup.-5]Btu/h ([10.sup.-5] W), this contact-free method was well-suited for very small thermal conductivity measurements.

Rywotycki (2003) investigated the moisture content in solid food products. It is proved that the resistivity of food has a linear function with conductor temperature, and a non-linear trend with moisture content. The food moisture content was determined by measuring its electric resistance or capacitance. A unique apparatus was designed for this purpose. The food was placed under an adjustable plate with a screw that could be adjusted to maintain same thickness. The advantage of his design is that food moisture content could be measured with good repeatability because density of samples could be maintained constant during the experiments. Litovsky et al. (2008) focused to measure the thermal conductivity of thin layers of strongly compressed fiber insulation materials. The apparent thermal conductivity was determined using the method of monotonic heating based on the Fourier equation. Freitas et al. (1991) proposed a method to determine the moisture content in the porous insulation. Their method was based on the attenuation of gamma-ray. The intensity of transmitted radiation was expressed as a function of moisture content in the volume.

All these techniques provide good approaches to measure the thermal conductivity and the moisture content in materials. Although each approach has some benefits it does not turn out to be quite feasible when considering mechanical pipe insulation systems.


The experimental apparatus consists of three parts: the pipe insulation tester (PIT), the refrigeration system, and the psychrometric chamber to control the ambient conditions. Each part is described next.

Pipe Insulation Tester (PIT)

The pipe insulation tester (PIT) consists of an aluminum pipe of 3 inches (76.2 mm) Nominal Pipe Size (NPS). The Aluminum pipe is installed in between two end sections that act as thermal guards as shown in Figure 1. A center copper pipe carries the refrigerating fluid in two-phase vapor and liquid mixture. The fluid enters the PIT and undergoes a slow evaporation process. It exits the PIT as still vapor-liquid mixture. Because of the low evaporation process of the refrigerant, heat is absorbed from the pipe insulation system creating two simultaneous effects: (i) an inward heat flow through the pipe insulation system and (ii) a uniform temperature distribution along the axial direction of the Aluminum pipe, since the saturation temperature of the refrigerant is practically unchanged throughout the test section. For the PIT, the instrumentation includes 20 surface temperature sensors distributed on Aluminum pipe surface, 20 surface temperature sensors distributed on pipe insulation system exterior surface, 6 surface temperature sensors along the surface of the refrigerating cold copper pipe, 2 pressure sensors and a Coriolis type mass flow meter, not shown in Figure 1 for clarity. The refrigerant was R134a or R404A, depending on the temperature range that is required during the measurements of the thermal conductivity of the pipe insulation specimen.


The Aluminum pipe was filled with pre-dried sand that acts as an intermediate insulation material between the refrigerant cold pipe and the Aluminum pipe. Sand was selected based on its compact structure and appropriate thermal property for this application (Tarnawski et al. 2009). The compact structure of sand filling the Aluminum pipe reduced the possibility of formation of small air pockets. If water vapor enters the Aluminum pipe it might frost and accumulate in these air pockets. The sand formed a fairly homogenous filling along the entire Aluminum pipe and provided a suitable thermal conductivity so that large temperature gradients between the inner copper tube and Aluminum pipe surface were guaranteed during the experiments. The sand was compatible with the thermocouples wires that run along the exterior surface of the copper tube and sand particles filled the gaps in between the thermocouple wires. Sand had good thermal stability, which was an important characteristic of the filling material since the assembly process occurred at room temperature while the working conditions of the sand during the experiments were well below freezing temperature. Finally sand was a practical choice for conducting regular inspections and calibrations of the thermocouples attached on the copper surfaces without damaging the instrumentation. It was an inexpensive and readily available material that could be used for filling the Aluminum pipe without requiring any custom-made tool. The sand was sealed in vapor barrier plastic bags and additional plastic plugs were glued and sealed at each end of the Aluminum pipe. This was done to prevent moisture ingress inside the Aluminum pipe and contaminating the sand. Test insulations were precisely cut to fit the center test section of the Aluminum pipe. The pipe insulation dimensions were measured according to the standard ASTM C585 (ASTM 2009) before installation. Certain types of joint sealants were applied to the longitudinal joints, depending on the characteristics of the test insulations and on the manufacturer recommendations. Additional 20 thermocouples, which were calibrated in-situ, were placed around the exterior surface of the pipe insulation approximately following a spiral path. The tips of these thermocouples were covered by a low thermal conductivity silicone gel to prevent interference with the surrounding air and provide a more accurate measurement of the local pipe insulation surface temperature.

The end sections, referred in this paper as thermal end guards, were designed as two 48 inches (1.2 m) long, with an outer diameter of 10 1/2 inches (26.7 mm). The wall thickness of the thermal end guards was 3.5 inches (88.9 mm) for the purpose of limiting heat transfer at the end sections of the apparatus. The end guards added a large thermal resistance and moisture ingress was limited since the material of the end guard was inherently vapor resistant. Two sections of nominal 3 inches NPS (7.6 mm) ID, 12 inches (0.3 m) long plastic pipes were installed at the end edges of the Aluminum pipe. Rubber couplings were used to join the two lateral plastic pipes to the central Aluminum pipe and about 1/2 inch (12.7 mm) of air gap was maintained in between the pipes to minimize axial thermal bridges in the apparatus. Polyethylene foam rubber, which is also vapor resistant, was inserted inside the plastic pipe and rubber caps were installed at their ends. Cellular glass cylindrical inserts were placed in between the rubber caps and refrigerating cold copper pipe in order to eliminate any direct contact between the highly conductive rubber material and the copper pipe. Then all other surfaces of the end guard were covered by cellular glass pipe insulation.

System to Control Ambient Conditions

Measuring of the actual thermal conductivity of pipe insulation systems under controlled ambient temperature and humidity could be performed in psychrometric chambers. In this work a psychrometric chamber was designed so that a slow motion of the air was produced with air ascending in the room from a perforated floor. This displacement ventilation system, shown in Figure 2, consisted of a conditioning loop, an under floor air plenum supply, and a set of adjustable ceiling filters. Air was circulated through the conditioning loop first, shown in Figure 2, by using a variable speed fan. The air flow rate was adjusted with the fans and a set of electro-mechanical dampers were used to control the supply air to the room. With reference to Figure 2, during the first process, the air goes inside the conditioning loop and is cooled and dehumidified through water-to-air cooling coils. The coils surface temperature and capacity was controlled by a variable speed pump, electronic mixing valves and bypass valves. These parameters were adjusted so that the air is cooled and dehumidified at about 43 [degrees] F (6 [degrees] C) of dew point temperature. To achieve dry non-condensing conditions, the dew point temperature of the ambient must be below the Aluminum pipe surface temperature of 40 [degrees] F (4.5 [degrees] C). In order to reach such conditions solid desiccant silica gel material, not shown in Figure 2, was installed along the air stream inside the conditioning loop. The solid silica gel material was replaced periodically to insure that the room humidity was maintained low enough throughout the experiments such that non-condensing conditions were achieved. After being cooled and dehumidified, air stream was guided to a series of electric resistance heating coils, which raised the air temperature up to the required ambient room temperature for the tests. The electric resistance heaters allowed for precise temperature control and an immediate response time. The power to these heaters was controlled by a PID control algorithm referencing the room dry bulb temperature. The control strategy was developed so that the cooling coils were set at constant load during one test while the heaters were adjusted very quickly to maintain the desired conditions. Since the cool-ing coils had a large thermal inertia this control strategy resulted in stable and steady ambient conditions around the pipe insulation specimen throughout the entire period of the tests.


Four humidity sensors and additional thermocouples were installed around the PIT to detect the local air psychrometric conditions. A panel of flexible elastomeric rubber insulation was positioned directly underneath the pipe insulation specimen, as shown in Figure 3. This configuration avoided direct impingement of the air on the bottom exterior surface of the insulation specimen, which stood about 33 inches (0.8 m) high from the perforated floor. The measurements indicated that this configuration produced an extremely well uniformed distribution of the temperature, humidity, and velocity of the ambient air around the pipe insulation system.


Refrigeration System Used to Achieve Below Ambient Temperature of the Pipe Insulation

A refrigeration system was used to cool the copper pipe in the range from 20 to 25 [degrees] F (-29 to -4 [degrees] C) via two-phase flow evaporation. The selection of the refrigerant was R134a for pipe insulation systems with moderate thermal conductivity and R404A for pipe insulation systems that have a very-low thermal conductivity. For the latter case, a very low temperature of the refrigerant was required in order to create a measurable radial heat flux across the pipe insulation test specimen. A two-phase flow boiling approach was chosen because it provides a good distribution of surface temperature along the axial direction of the pipe. It was also stable with time, which is a desirable feature for the time-extensive experiments in wet ambient conditions. Refrigerants entered and exited the PIT as the two-phase vapor and liquid mixture, with thermodynamic vapor quality within the saturated region. Two PITs were connected in series with respect to the refrigerant flow, that is, the outlet of the first PIT was connected to the inlet of the second PIT as shown in Figure 4. The sensors on the first PIT aimed to measure the thermal conductivity of the pipe insulation specimen, while the second PIT served a dual purpose: (i) it allowed monitoring the sand thermal conductivity during the dry tests, and (ii) it measured the moisture content in the pipe insulation system during the wet tests. A flexible strip electric heater was installed along the surface of the Aluminum pipe of the second PIT during the dry tests. A large amount of insulation was wrapped around the heater to limit radial heat gain from the surrounding ambient. The power to the heater was controlled to obtain similar Aluminum surface temperature in the second PIT as the one measured in the first PIT during the dry tests. Since the two PITs were of identical geometry, constructed at the same time, and maintained at the same temperature and humidity boundary conditions, any variation in the sand conductivity measured with the second PIT was postulated to be reflected with an equivalent variation of the sand conductivity inside the first PIT. In other words, while the absolute value of the sand thermal conductivity for the two PITs was slightly different, the increase of the sand thermal conductivity due to moisture ingress into the Aluminum pipe of the first PIT was assumed to be proportional to the increase of sand thermal conductivity measured in the second PIT, if present. This approach allowed monitoring the thermal conductivity of the sand and detecting any change of the sand thermal conductivity due to potential failure of the sealing and moisture ingress inside the Aluminum pipe.


A tube-in-tube suction line heat exchanger (HX) was installed after the second PIT device, referred in Figure 4 as Tube/Tube Suction Line HX. The refrigerant exits from the second PIT device flows into the refrigerant suction line HX, which increased the cooling capacity available for the PITs in the refrigeration unit. In the non-contact tube-in-tube suction line HX, the room temperature refrigerant coming from the air-cooled condenser was cooled well below the condenser saturation temperature by the low temperature refrigerant from the second PIT. Then the refrigerant was throttled in the metering valve and the entire throttling process shifted toward the liquid side of refrigerant thermodynamic two-phase dome by adopting the suction line HX in the refrigeration loop. Lower thermodynamic vapor quality at the inlet of the first PIT allowed accommodating larger enthalpy variations before the refrigerant reached a superheated state at the outlet of the second PIT. This translated in larger heat transfer available during the two-phase flow evaporation of the refrigerant inside the first and second PITs. The suction line HX was made by welding a 48 inches (1.2 m) long, 1/2 inch (12.7 mm) OD copper pipe around the system 3/8 inch (9.5 mm) copper tube. Before returning to the compressor, proper degree of superheat at the compressor suction were achieved by an auxiliary water-to-refrigerant tube and tube heat exchanger. Hot water was circulated in the auxiliary loop with an in-line heater, whose power is adjusted to guarantee proper suction conditions at the compressor. It should be noted that electric heaters apply directly on the surface of the cold copper pipe were intentionally avoided in the test apparatus because they created large axial temperature gradients in the refrigeration copper pipelines used inside the PIT devices. These large axial temperature gradients were the results of conduction across the walls of the pipes and they affected the uniformity of the surface temperature along the axial direction of the pipe Insulation as well as introduced a systematic error during the calculation of pipe insulation thermal conductivity. Thus, we purposely avoided to have any electric heater directly in contact with any metal surface of the test apparatus. This strategy minimized the undesired effect of axial conductions.

Test Conditions

The test apparatus evaluated pipe insulation that was applied to a 3 inches (76.2 mm) NPS pipe with the surface temperature controlled to 40 [degrees] F (4.5 [degrees] C). The insulation thick-ness ranged from 1 to 2 inches (25.4 to 50.8 mm) in wall thick-ness. Dry non-condensing conditions were achieved by operating the chamber room with very low ambient humidity. In these conditions, water vapor did not condensate on the Aluminum pipe surface and the tests are referred throughout this paper as dry non-condensing tests. The thermal conductivity of the pipe insulation was measured for average insulation temperature ranging from 56 to 73 [degrees] F (13 to 23 [degrees] C), that is, cold surface side of 40 [degrees] F (4.5 [degrees] C) and hot surface side from 70 to 104 [degrees] F (22 to 40 [degrees] C). The hot surface temperature was controlled by varying the ambient temperature of the surrounding air from 77 to 110 [degrees] F (25 to 43.4 [degrees] C). Any jacketing system and vapor barriers were removed from the exterior of the pipe insulation specimen. More than 50 temperature sensors were used to monitor the interior and exterior local surface temperature of the pipe insulation specimen. An example of the axial and angular temperature measurements is shown in Figure 5. Pipe insulation surface temperatures were measured by 20 thermocouples positioned around the exterior surface of insulation specimen and following a spiral configuration. By blocking the air stream right below the test apparatus with a panel of flexible elastomeric insulation, the temperature distribution of the pipe insulation exterior surface was within [+ or -]0.95 [degrees] F (0.5 [degrees] C). Twenty additional thermocouples were positioned around the Aluminum pipe exterior surface and they were installed inside longitudinal grooves of about 1/8 inch (3.2 mm) depth. The grooves were cut out in the Aluminum pipe so that the thermocouple wires did not interfere with the installation of the pipe insulation around the Aluminum pipe. The local surface temperatures of the Aluminum pipe were within [+ or -]0.5 [degrees] F (0.3 [degrees] C) when considering both axial and angular directions, as shown in the central plot of Figure 5. Finally, 6 thermocouples were attached to the copper tube and the temperature variation in both axial and angular directions was within 2 to 3.5 [degrees] F (1 to 2 [degrees] C). This is shown in the bottom plot of the figure. The average copper surface temperature was slightly higher than the refrigerant saturation temperature inside the copper pipe, which was estimated from the average refrigerant pressure inside the first PIT device.


Data Reduction and Uncertainty Analysis

Calibration tests were conducted at the beginning of each experiment to determine the actual thermal conductivity of the sand filling the 3 inches (76.2 mm) NPS Aluminum pipe. In the calibration tests, a tape resistor-type heater was applied uniformly around the exterior surface of the Aluminum pipe and several layers of rubber elastomeric insulation were installed around the heater to create a thermal insulation barrier from the ambient. When the electric heater was energized, the electric power was carefully measured using a precision watt meter. The total heat transfer into the Aluminum pipe, [Q.sub.Al,pipe], was estimated by considering both the electric power converted into heat transfer rate in the tape resistor heater, [Q.sub.heater], and a small amount of heat transfer leaking in from the surrounding ambient through the thick rubber insulation layer and from the axial heat conduction through the end sections of the test apparatus, [Q.sub.leak,in], as given in (1):

[Q.sub.Al, pipe] = [Q.sub.heater] + [Q.sub.leak, in] (1)

During the calibration of the PIT device the ambient temperature was adjusted such that the temperature difference across the outer thermal insulation barrier resulted less than 5.4 [degrees] F (3 [degrees] C). The axial temperature gradient along the Aluminum pipe was less than 0.43% of the radial temperature gradient. Thus, it was reasonable to assume that the axial heat losses were small compared to the magnitude of the radial heat flow. However, even though the resulting [Q.sub.leak,in] was fairly small it was accounted for during the calibration procedures to eliminate a systematic error on the actual thermal conductivity of the sand inside the Aluminum pipe. It should be emphasized that the actual thermal conductivity of the sand inside the Aluminum pipe was slightly different than the thermal conductivity of pure dry sand because the measured values during the calibration procedure account for the end effects of the thermal guards, for the contribution due to small axial heat flow, and for the effects due to the non-uniformity and non-homogenous properties of the sand inside the Aluminum pipe. The thermal conductivity of the sand also depended on the percent of quartz and residual moisture in the sand batch used for filling the pipe. Testing the Aluminum pipe at the same thermal conditions as the ones occurring during the actual measurements of the insulation thermal conductivity allowed accounting for all of these non-ideal effects. An equivalent sand thermal conductivity was calculated to approximate the 3D geometry and non-homogeneity of the PIT device as to a 1D heat conduction model that follows the Fourier law.

Once the effective conductivity of the sand filling the Aluminum pipe, [k.sub.sand], was known from the calibration procedure, a 1-D heat transfer balance equation along the radial direction was applied to the PIT device, as shown in the schematic of Figure 6. As a result, the pipe insulation thermal conductivity, [k.sub.ins], was determined directly by the following expression:


[k.sub.ins] = [k.sub.sand] x [([T.sub.Al, pipe] - [T.sub.cold, copper, pipe])/([T.sub. exterior, ins, specimen] - [T.sub.Al, pipe])] x [G.sub.f]) (2)

Where [G.sub.f] is a geometry factor that depends on the ratios of pipe insulation exterior diameter, cold Aluminum pipe outer diameter, and refrigerating copper pipe outer diameters, as shown in Figure 6. The diameters of the pipe insulation were measured according to the standard ASTM C585 (ASTM 2009).

The average temperature of the test insulation specimen was calculated from the measurements of the surface temperature sensors as follows:

[T.sub. avg, ins, specimen] = ([T.sub.Al, pipe] + [T.sub. exterior, ins, specimen])/2 (3)

Uncertainty Analysis

A complete uncertainty analysis was conducted and calculations were carried out by using a numeric model of the test apparatus that was developed in EES (Engineering Equation Solver (Klein 2006)). The experimental data for the axial and radial temperature measurements were input to the numerical model of the PIT device. The uncertainty on the insulation thermal conductivity was estimated according to the Taylor series expansion method as follows (Taylor 1997):


where [U.sub.Y] represents the uncertainty of the variable Y and [U.sub.X] represents the precision accuracy of the measured variable X. The uncertainty on the pipe insulation thermal conductivity was calculated following a similar approach and it resulted as given in equation (5):


Where each term in parenthesis represents the coefficient of sensitivity for the variable [X.sub.i] and the [U.sub.Xi] on the right hand side of equation are the uncertainty associated with the variable [X.sub.i.] The uncertainty on the thermal conductivity depended on the sensor accuracy and on the spatial uniformity of the temperatures along the axial and angular directions of the PIT device. All temperature sensors were calibrated in-situ and the bias uncertainty was computed according to the approach provided by Johnson et al. (1998) and summarized as below:

[B.sub.Tdistribution] = ([T.sub.max] - [T.sub.min])/2 (6)

[B.sub.T] = [([B.sub.Tdistribution.sup.2] + [B.sub.Tdistribution.sup.2].sup.0.5] (7)

Where [B.sub.Tdistribution] is the bias uncertainty resulting from the actual non-uniform temperature distribution, [B.sub.Tuniform] is the bias uncertainty present even with a uniform temperature distribution and it was estimated from equation (6) with [T.sub.max] and [T.sub.min] set as the acceptable limits for temperature uniformity condition, BT is the total bias uncertainty, and n is the total number of measuring points in the grid.

The overall uncertainty was calculated taking into consideration the error from the accuracy of the temperature sensors, repeatability of the measurements, and spatial uniformity of the temperature measurements and the results are summarized in Figure 7. Accuracy and precision of the measurements are also given in Table 1. The uncertainty on the effective thermal conductivity of the sand filling the Aluminum pipe is the main factor impacting the uncertainty of the pipe insulation thermal conductivity. The two uncertainties follow similar trends with the latter being amplified by the accuracy of the temperature measurements. It was estimated that the uncertainty on the pipe insulation thermal conductivity was about [+ or -]6% when the uncertainty of the sand thermal conductivity was within [+ or -]5%. In the work presented in this paper, the novel experimental apparatus was used in the dry non-condensing ambient conditions but it offers significant advantages with respect to the existing test apparatuses for wet condensing conditions, in which water vapor ingress into the insulation is promoted.

Table 1. Accuracy and Max Spatial Variation of the Temperature
Measurements of the PIT

Description       Range       Accuracy      Max spatial
                                            ([T.sub.max] -

Thermocouples  -4 - 32      [+ or -]0.36  1.5 [degrees] F
(Copper Tube)  [degrees] F  [degrees] F   ([+ or -]0.8
               (-20 - 0     ([+ or -]0.2  [degrees] C)
               [degrees]    [degrees]
               C)           C)

Thermocouples  36 - 43      [+ or -]0.18  0.8 [degrees] F
(Al Pipe)      [degrees] F  [degrees] F   ([+ or -]0.4
               (2.2 - 6.1   ([+ or -]0.1  [degrees] C)
               [degrees]    [degrees]
               C)           C)

Thermocouples  59 - 104     [+ or -]0.18  2.1 [degrees] F
(exterior      [degrees] F  [degrees]F    ([+ or -]1.2
insulation)    (15 - 40     ([+ or -]0.1  [degrees] C)
               [degrees]    [degrees]
               C)           C)

Description     Uncertainty

Thermocouples  [+ or -]0.44
(Copper Tube)  [degrees] F
               ([+ or -]0.25
               [degrees] C)

Thermocouples  [+ or -]0.18
(Al Pipe)      [degrees] F
               ([+ or -]0.1
               [degrees] C)

Thermocouples  [+ or -]0.21
(exterior      [degrees] F
insulation)    ([+ or -]0.12
               [degrees] C)


Two pipe insulation systems were studied initially with the new experimental apparatus described in this paper. The new apparatus was validated with these two pipe insulation systems, cellular glass and Polyisocyanurate (PIR), used to benchmark our measurements against data available in the public domain The measurements were taken at several average temperatures of the pipe insulation specimen to allow some extrapolation of the thermal conductivity for temperatures slightly above room ambient. Again this was done in order to compare our measurements with the ones available in the public domain for these two materials. The ambient temperature was varied from 68 to 117 [degrees] F (20 to 47 [degrees] C) and the refrigerant temperatures were controlled from 12 up to 39 [degrees] F (-11 to 4 [degrees] C) depending on radial thermal gradient required to maintain the Aluminum pipe surface temperature at 40 [degrees] F (4.5 [degrees] C). The estimates from our measurements at above ambient temperatures were then compared with the measured thermal conductivity of flat blocks of the same material and, in the case of the Polyisocyanurate, of the same batch of insulation material. During this comparison the aim was to verify that the thermal conductivity measured with the new apparatus presented here were of the same order of magnitude as the data obtained from flat blocks, even though we expected some percent of deviations due to the radial configuration and split joints.

Cellular Glass Pipe Insulation

Three thicknesses of cellular glass pipe insulation were selected for validation of the test apparatus: pipe insulation systems with 1 inch (25.4 mm), 1 1/2 inches (38.1 mm) and 2 inches (50.8 mm) wall thicknesses. Due to manufacturing constraints, the insulation samples were only provided in length of 2 ft (0.6 m) section (ASTM 2007). Test specimens were machined to fit and staggered when installed on the Aluminum pipe, preventing air circulation through the gaps. All the joints were sealed with joint butyl rubber sealant (BOSS 368), which provides excellent adhesion in between the C-shells of the pipe insulation. The sealant was initially applied following an S pattern on the edge of the pipe shell, with a thickness of about 1/8 inch (3.2 mm) before the two halves were tightly pressed and the effect of the joint sealant will be discussed later in the paper.

Since cellular glass material is practically not subjected to aging effect, the three samples were tested at different time and with two techniques for measuring the surface temperature of the insulation as listed in Table 2. Figure 8 compares the measured thermal conductivity versus insulation average temperature and at different thickness of cellular glass pipe insulation systems. The error bar in the data represents the maximum deviation of experimental data from a linear fit correlating all the data. Compared to the manufacturer's data, whose value of thermal conductivity was obtained for flat block configuration according to the standard ASTM C518 (ASTM 2010c), the experimental data from the new test apparatus were in agreement within [+ or -] 3%, which is consisted with the findings from the literature. The plots in Figure 8 were obtained by attaching the thermocouples on the exterior insulation surface using small neoprene adhesive square strips and by using droplets of silicone gel covering the tips of the thermocouples. These two techniques produced quite different results as shown in Figure 8. Due to the open-cell structure of neoprene, the local thermocouples on the exterior of the pipe insulation surface measured an average temperature between local ambient air temperature and local surface temperature of the material. From comparisons with measurements without any neoprene strips on the top of the thermocouple we concluded that the local temperature was weighted toward the ambient temperature rather than the pipe surface temperature. On the contrary, the measurements obtained using silicone gel droplets as surface adhesive covering the tips of the thermocouples seemed to be more accurate because the local surface temperature of the exterior of the insulation was lower than the ambient temperature of the surrounding air. Silicon gel used in these tests had a low thermal conductivity and thus act as an adhesive and as a large thermal barrier between the tip of the thermocouple and its immediate air film around the insulation. The measurements of the thermal conductivity of the 1 inch (25.4 mm) wall thickness cellular glass insulation specimen were repeated four months later using the silicone gel method and the average thermal conductivity was observed to be about 4.5% higher than the previous tests in which neoprene was used. This was explained because of the effect of the thermocouple attachment method and of the joint sealant thickness, which slightly increased the inner diameter of the insulation creating air pockets for the second tests. We concluded that the silicon method produced more accurate and more repeatable results. Thus, for all other insulation materials only the measurements that adopt silicone gel for attaching the thermocouples to the exterior surface of the pipe insulation are reported in this paper.

Table 2. Validation Experiment Results of Cellular Glass

Cellular glass  Days after   Thermocouple  [k.sub.pipe,insulation] =
thickness inch  calibration   attachment     a * T + b Btu-in/hr-
(mm)              of PIT                     [ft.sup.2]-F (W/m-K)


1 (25.4)                 39      Neoprene           0.0003 (0.00005)
1.5 (38.1)               78      Neoprene           0.0017 (0.00026)
2 (50.8)                140      Silicone            0.003 (0.00080)
1 (25.4)                154      Silicone            0.001 (0.00025)
Manufacturer              -             -           0.0006 (0.00014)

Cellular glass   [k.sub.pipe,insulation] =
thickness inch     a * T + b Btu-in/hr-
(mm)               [ft.sup.2]-F (W/m-K)


1 (25.4)          0.2695
1.5 (38.1)        0.1909
2 (50.8)          0.0946
1 (25.4)          0.2425
Manufacturer      0.2498

Polyisocyanurate (PIR) Pipe Insulation

A 1 inch (25.4 mm) and a 2 inches (50.8 mm) wall thickness of Polyisocyanurate (PIR) pipe insulation systems were selected in the second series of tests aimed to validate the novel test apparatus. Unfortunately PIR is subjected to aging phenomena and the thermal conductivity is also a function of time. To obtain a fair comparison, a batch of the material was prepared and radial C-shells and flat blocks were taken together from the same batch and were tested in a laboratory at Oak Ridge National Laboratory at approximately the same time. The results of the experiments are given in Figure 9 and the correlations for the pipe insulation thermal conductivity are given in Table 3. The 2 inches (50.8 mm) wall thickness insulation specimen was very sensitive to the temperature when compared to the 1 inch (25.4 mm) wall thickness PIR insulation specimen. A possible reason might be due to the fact the PIR had extremely low thermal conductivity and the radial heat flux for the 2 inches (50.8 mm) wall thickness was too low to be accurately measured with the current PIT device if the Aluminum pipe surface temperature must be maintained at 40 [degrees] F (4.5 [degrees] C). In order to increase the accuracy of the measurements a larger radial temperature gradient would be necessary but changing the Aluminum pipe surface temperature would introduce more complexity for the comparison and it was outside the scope of this work. As shown in Figure 9 there is some difference between the measured value of thermal conductivity of the 1 inch (25.4 mm) wall thickness PIR pipe insulation and the value obtained with the flat block configuration. The difference was up to 8.8% and it might be due the edge effects of the C-shell longitudinal split joints and the presence of joint sealants along longitudinal joints of the C-shells. These factors were investigated further in this work as it will be explained in more detail later in this paper.

Table 3. Validation Experiment Results of PIR

Cellular     Days from   Thermocouple  [k.sub.pipe,
glass       calibration   attachment   insulation] =
thickness,      test                       a * T + b
inch (mm)                                  Btu-in/hr-
                                         ft2-F (W/m-K)

                                               a            B

1 (25.4)            108      Silicone         0.0004    0.1748
                                           (0.00009)  (0.0270)

2 (50.8)            120      Silicone         0.0032   -0.0284
                                           (0.00083)  (0.0107)

Thermal Conductivity of Mechanical Pipe Insulation Systems in Dry Non-Condensing Ambient Conditions

The thermal conductivity of fiberglass, flexible elastomeric and phenolic pipe insulation systems was measured in dry non-condensing ambient conditions. The nominal insulation wall thickness was 2 inches (50.8 mm) for all three specimens and the average insulation temperature was varied from 57 to 74 [degrees] F (14 to 23 [degrees] C). Correlations of the pipe insulation thermal conductivity were developed based on insulation specimen average temperature and wall thicknesses. The thermal conductivity of these three insulation systems varied linearly as shown in Figure 10 and the coefficients of the linear correlations are given in Table 4. All three sets of experimental data fit well a linear interpolation curve fit with a maximum deviation of less than [+ or -]2.2%. Increasing thermal conductivity was observed if the temperature increases. This observation was in agreement with previous observation in the literature (Saxena et al. 1989; Budaiwi and Abdou 2002; Abdou and Budaiwi 2005; Litovsky et al. 2008). Compared to ASTM standards and manufacturer or catalog data presented for fiberglass, flexible elastomeric and phenolic pipe insulation, we observed some differences which could be explained with radial configuration, aging effects, and the presence of joint sealant on the longitudinal butt joints as explained next.

Table 4. Thermal Conductivities of Pipe Insulations Under Dry

Test Samples  [k.sub.pipe, insulation] = a *
              T + b Btu-in/hr-ft2-F (W/m-K)

                            a                     B

Fiberglass                  0.0004 (0.00010)    0.2101

Flexible                    0.0005 (0.00014)    0.2144
Elastomeric                                   (0.0334)

Phenolic                    0.0012 (0.00032)    0.1217

Effects of the Joint Sealant Applied to the Longitudinal Butt Joints of the Pipe Insulation C-shells

We investigated further the effect of the joint sealant on the measured thermal conductivity of mechanical pipe insulation systems. For the phenolic and cellular glass pipe insulation systems we installed the C-shells on our PIT device and utilized a joint sealant for the longitudinal joints. Using the joint sealant is the proper procedure recommended by the manufacturers and we investigated the secondary effects that the compound might have on the actual thermal conductivity of the pipe insulation system. Figure 11 shows the experimental results for two cases: with and without joint sealant for phenolic pipe insulation with 1 inch (25.4 mm) wall thickness. Due to the fact that the thermal conductivity of joint sealant is higher than the one of the insulation material, part of the radial heat flow occurred through the joints, instead of across the insulation material. This could lead to an increase of the specimen thermal conductivity of cylindrical shaped pipe insulation consisting of two C-shells with respect to the thermal conductivity of the pure material from flat blocks. By adjusting the thickness of joint sealant, from 0.06 inch (1.59 mm) to 0.1 inch (2.54 mm), the two trends of Figure 11 could be over-lapped if a sealant thermal conductivity is between 2.4 Btu-in/ hr-ft2-F (0.35 W/m-K) to 3.5 Btu-in/hr-ft2-F (0.5 W/m-K). For phenolic pipe insulation system, the edge effects from the joint sealant applied on the longitudinal butt joints of the C-shells increased the actual thermal conductivity by about 15%.



A novel experimental apparatus to measure the thermal conductivity of mechanical pipe insulation systems at below ambient temperatures was developed. The test set up was validated by using two types of pipe insulation systems as bench-mark and the thermal conductivity was measured for average pipe insulation temperature varying from 50 to 75 [degrees] F (10 to 24 [degrees] C). Through a series of extensive calibration procedures the uncertainty on the pipe insulation thermal conductivity was estimated to be within [+ or -]6% and the measured thermal conductivity was within 15% with respect to the data available in the public domain. The deviation was mainly due to radial configuration, edge effects, and the presence of joint sealants along the longitudinal butt joints of the pipe insulation systems. The thermal conductivity of pipe insulation systems with 2 inches (50.8 mm) nominal wall thickness was also presented for the case of fiberglass, flexible elastomeric, and phenolic pipe insulation. Results were reported in dry non-condensing ambient conditions and linear trends of the thermal conductivity were observed with respect to the average insulation temperature.

The novel experimental apparatus was used in dry non-condensing ambient conditions but it offers significant advantages with respect to the existing test apparatus for wet condensing conditions, in which water vapor ingress into the insulation is promoted. The authors recommend that once the thermal conductivity of a pipe insulation system is known for dry non-condensing conditions, a similar test set up could be used to investigate the effect of moisture ingress on the actual thermal conductivity of pipe insulation systems under wet ambient conditions and future studies should be conducted to study further this phenomenon.


The authors gratefully acknowledge funding and support from ASHRAE as well as the help of the Project Monitor Subcommittee Members and ASHRAE Technical Committee TC 1.8 Mechanical Systems Insulation. The authors would also like to thank the Building Envelopes Research Group at Oak Ridge National Laboratory for providing comparison data in Figure 9 during the validation of the test apparatus developed in this work.


B = bias uncertainty for spatial distribution [-]

D = diameter, inches or (m)

[G.sub.f] = geometric factor [-]

k = thermal conductivity [(Btu-in)/(hr-ft.sup.2]- [degrees] F), or (W/m- [degrees] C)

n = total number of measuring points in the grid.

[Q.sub.Al, pipe] = total heat transfer rate into the Aluminum pipe, Btu/h, or (W);

[Q.sub.heater] = electric power in the tape resistor heater Btu/h, or (W);

[Q.sub.leak,in] = heat transfer rate leaking in from the ambient through the external insulation, Btu/h, or (W);

T = temperature, [degrees] F, or ([degrees] C)

U = uncertainty [-]

X = uncertainty variables [-]

Y = uncertainty variables [-]


Al = aluminum

avg = average

ins = insulation

Max = maximum

Min = minimum


Abdou, A. A., & Budaiwi, I. M. (2005). Comparison of thermal conductivity measurements of building insulation materials under various operating temperatures. Journal of Building Physics, 29(2), 171-184.

ASHRAE (2009). ASHRAE Handbook Fundamentals. Atlanta, GA, USA: ASHRAE.

ASTM (2007). ASTM C 552 - 07, Standard Specification for Cellular Glass Thermal Insulation. Philadelphia: ASTM International.

ASTM (2009). ASTM C 585 - 09, Standard Practice for Inner and Outer Diameters of Thermal Insulation for Normal Sizes of Pipe and Tubing. West Conshohocken, PA: ASTM International.

ASTM (2010a). ASTM C335 / C335M - 10e1 Standard Test Method for Steady-State Heat Transfer Properties of Pipe Insulation. West Conshohocken, PA: ASTM Inter-national.

ASTM (2010b). ASTM C177 - 10 Standard, Test Method for Steady-State Heat Flux Measurements and Thermal Transmission Properties by Means of the Guarded-Hot-Plate Apparatus. West Conshohocken, PA: ASTM Inter-national.

ASTM (2010c). ASTM C518 - 10 Standard Test Method for Steady-State Thermal Transmission Properties by Means of the Heat Flow Meter Apparatus. West Conshohocken, PA: ASTM International.

Batty, W. J., Probert, S. D., Ball, M., & O'Callaghan, P. W. (1984). Use of the thermal - probe technique for the measurement of the apparent thermal conductivities of moist materials. Applied energy, 18(4), 301-317.

Budaiwi, I., Abdou, A., & Al-Homoud, M. (2002). Variations of thermal conductivity of insulation materials under different operating temperatures: Impact on envelope-induced cooling load. Journal of Architectural Engineering, 8(4), 125-132.

Crall, G. C. (2002). The use of wicking technology to man-age moisture in below-ambient insulation systems, Insulation Materials: Testing and Applications, ASTM Special Technical Publication: 4th Volume, ASTM 1426 (pp. 326-334). West Conshohocken, PA: American Society for Testing and Materials.

Chyu, M.-C., Zeng, X., & Ye, L. (1997a). Effect of moisture content on the performance of polyurethane insulation used on a district heating and cooling pipe, Philadelphia, PA, USA.

Chyu, M.-C., Zeng, X., & Ye, L. (1997b). Performance of fibrous glass pipe insulation subjected to underground water attack, Philadelphia, PA, USA.

Fleischmann, A., Hao, H. Y., Neumann, M., & Enss, C. (2003). A new technique for measuring thermal conduc-tivity at low temperature.

Freitas, V., Crausse, P., & Abrantes, V. (1991). Moisture dif-fusion in thermal insulating materials, Gatlinburg, TN, USA.

Guldbrandsen, T., Kaplar, C. W., & Korsgaard, V. (2011). Analytical model of heat transfer in porous insulation around cold pipes. International Journal of Heat and Mass Transfer, 54(1-3), 288-292.

Gustafsson, S. E., Chohan, M. A., & Khan, M. N. (1985). Thermal transport studies using pulse transient hot strip measurements of vitreous silica produced by vacuum deposition. High Temperatures - High Pressures, 17(1), 35-39.

Gustafsson, S. E., Karawacki, E., & Khan, M. N. (1978). Transient hot - strip method for simultaneously measuring thermal conductivity and thermal diffusivity of sol-ids and fluids. Journal of Physics D: Applied Physics, 12(9), 1411-1421.

Hao, H. Y., Neumann, M., Enss, C., & Fleischmann, A. (2004). Contactless technique for thermal conductivity measurement at very low temperature. Review of Scientific Instruments, 75(8), 2718-2725.

Hay, C. C. (1984). Moisture Absorption and its Effect on the Thermal Performance of Extruded Polystyrene Insulation. University of Alaska, Faibanks, Alaska.

Howard, J. F., Coumou, K. G., & Tye, R. P. (1973). Direct reading thermal conductivity instrument with digital read-out for the measurement of heat transmission in cellular plastics. Journal of Cellular Plastics, 9(5), 226-231.

Johnson, A. B., Simonson, C. J., and Besant, R. W (1998). Uncertainty analysis in the testing of air-to-air heat/ energy exchangers installed in buildings. Paper presented at the ASHRAE Transactions.

Klein, S. A. (2006). Engineering Equation Solver (Version V7.723-3D). Madison, WI, USA: F-Chart Software.

Korsgaard, V. (1993). Innovative concept to prevent moisture formation and icing of cold pipe insulation. Paper presented at the ASHRAE Transactions.

Litovsky, E., Issoupov, V., Kleiman, J., Latham, R., Kotrba, A., & Olivier, K. (2008). Thermal conductivity of mechanically compressed fiber insulation materials in a wide temperature range: New test method and experimental results, Birmingham, AL, United states.

Log, T. (1991). Transient hot-strip method for simultaneous determination of thermal conductivity and thermal diffusivity of refractory materials. Journal of the American Ceramic Society, 74(3), 650-653.

McFadden, T. (1986). Moisture effects on extruded polystyene insulation, Anchorage, AK, USA.

Rywotycki, R. (2003). Electric sensor for prompt measurement of moisture content in solid food products. Journal of Food Process Engineering, 25(6), 473-483.

Saxena, N. S., Gustafsson, S. E., Chohan, M. A., & Maqsood, A. (1989). Temperature dependence of thermal conductivities and thermal diffusivities of composites using transient hot-strip method. International Journal of Energy Research, 13(4), 411-417.

Tarnawski, V. R., Momose, T., Leong, W. H., Bovesecchi, G., & Coppa, P. (2009). Thermal conductivity of standard sands. Part I. dry-state conditions. International Journal of Thermophysics, 30(3), 949-968.

Taylor, J. R. (1997). An introduction to error analysis: the study of uncertainties in physical measurements (pp. xvii, 327 p.). Sausalito, Calif.: University Science Books.

Wang, X., Chin, Y. D., Perera, R. M., Prescott, C. N., Raborn, G. O., & Brown, K. R. J. (2000). Convection heat losses through installation gaps between pipe and insulation and between insulation half shells, Houston, TX, USA.

Whitaker, T. E., & Yarbrough, D. W. (2002). Review of thermal properties of a variety of commercial and industrial pipe insulation materials, Charleston, SC, United states.

Wijeysundera, N. E., Hawlader, M. N. A., & Tan, Y. T. (1989). Water vapour diffusion and condensation in fibrous insulations. International Journal of Heat and Mass Transfer, 32(10), 1865-1878.

Wilkes, K. E., Desjarlais, A. O., Stovall, T. K., McElroy, D. L., Childs, K. W., & Miller, W. A. (2002). A Pipe Insulation Test Apparatus for Use Below Room Temperature Insulation Materials: Testing and Applications, ASTM Special Technical Publication: 4th Volume, ASTM 1426 (pp. 24-256). West Conshohocken, PA: American Society for Testing and Materials.

Yu, M., Sui, X., Peng, X., & Fang, Z. (2009). Influence of moisture content on measurement accuracy of porous media thermal conductivity. Heat Transfer - Asian Research, 38(8), 492-500.

This paper is based on findings resulting from ASHRAE Research Project RP-1356.

Lorenzo Cremaschi, PhD


Shanshan Cai Student


Kasey Worthington

Afshin J. Ghajar, PE

Lorenzo Cremaschi is an assistant professor, Shanshan Cai and Kasey Worthington are graduate research assistants, and Afshin J. Ghajar is a Regent Professor in the School of Mechanical and Aerospace Engineering, Oklahoma State University, Stillwater, OK.
COPYRIGHT 2012 American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Cremaschi, Lorenzo; Cai, Shanshan; Worthington, Kasey; Ghajar, Afshin J.
Publication:ASHRAE Transactions
Article Type:Report
Geographic Code:1USA
Date:Jan 1, 2012
Previous Article:Literature review of the effect of temperature and humidity on viruses.
Next Article:Passive Houses: what may happen when energy efficiency becomes the only paradigm?

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |