Measurement laboratories must be mindful of equipment limitations and increasing uncertainty as the extremes of operational range are approached. For thermal insulation measurement, the prominent issue with more highly-conductive materials is that of interface resistance. Equipment is generally based on the heat flow meter or the guarded hot apparatus, both of which measure the heat flow between flat parallel plates held at different temperatures with the test specimen located between them. Interface resistance occurs between the test specimen and the plates that it contacts. Standards prescribe lower limits of specimen thermal resistance, typically 0.1 m2.K/W. However whilst interface resistance may already be significant at this thermal resistance, measurement is often sought for more-conductive products. This dissertation considers a number of aspects of such measurements, in all cases proposing the use of flexible buffer materials at the interface between the test specimen and the apparatus plates in order to provide lower interface resistance. The use of this solution is seldom reported although it is described in some standards where it is also suggested that buffers of very low thermal resistance are required in order to minimize errors. However this would require them to be very thin and potentially ineffective. An alternative prospect has been explored, that of using thicker, softer interface materials to ensure good specimen contact. Separate measurement of the interface materials, in conjunction with an error analysis allows thermal resistance to be calculated as the difference between these measurements with known uncertainty. Chapter 2 describes a study using the difference approach to measure twelve highly-conducting specimens in conjunction with four foamed plastic buffer materials, based on PVC, silicone, EVA and nitrile. The specimens ranged from aluminium sheet to fluted plastic board. Compared with direct measurement, thermal resistance values via difference measurement were lower by between 0.003 and 0.01 m2.K/W, depending on specimen and buffer choice. Silicone sponge gave the most uniform results. Compression tests showed that it also displayed the lowest deformation hysteresis. An analysis of the difference calculation is given, showing it to be numerically inexact since there are residual interface-resistance terms that are not present in both measurement cases. Chapter 3 describes a further study of the difference technique in conjunction with flexible buffer materials, extending the procedure to materials of higher thermal resistance and to thinner, harder buffers. An alternative difference calculation is proposed to eliminate residual resistance terms through comparing results for the unknown specimen and for a reference specimen with similar surface characteristics and known properties, measured using the same buffers. Specimens of expanded polystyrene board and cast acrylic sheet were measured in the heat flow meter apparatus using two alternative silicone-based buffer materials, one solid and the other a sponge. Analysis also includes earlier measurements of twelve more highly-conducting specimens, adjusting for the residual error terms. Across all of these, thermal resistance values obtained by the difference method were lower by between 0.008 m2.K/W and 0.016 m2.K/W, attributable to removing the contribution of interface resistance. In Chapter 4, a technique incorporating buffer materials is proposed for measuring the thermal conductivity of moist earthen and granular loose fill materials. Transient methods involving needle and other probes are also reviewed but it is concluded that a steady state approach offers reliable uncertainty estimation and a test method that is widely accepted in industry. Variations to the standard loose-fill method are proposed, including the use of a rigid holding frame with stiff base and silicone sponge buffer sheets, in conjunction with difference measurement to factor out the contributions from base, buffers and contact resistance. Using this approach, consistent results were obtained for loose-fill earths based on scoria, terracotta and furnace-ash at different moisture contents. Thermal resistance ranged from 0.08 to 0.4 m2.K/W. Thermal conductivity fitted well to linear regression plots against moisture content. Further comparative measurements of a single specimen showed that direct measurement was less consistent than difference measurement, and that indicated thermal resistance was higher by 0.023 m2.K/W, this effectively being a measure of the interface resistance. Chapter 5 explores earlier evidence that high-conductance materials with rough surfaces, (such as many building boards), are measured to have higher thermal resistance and higher test thickness when measured with harder buffers. Results from an experimental study of nine materials and four buffer types are reported. Thermal resistance was higher by up to 0.01 m2.K/W and thickness by up to 0.5 mm using the hardest buffer relative to the softest. An analytical model was developed, allowing measured roughness to be expressed as flat high and low areas of varying height and area fraction so that thermal resistance and height variations could be predicted as a function of roughness. Predictions were consistent with optical roughness measurements. The model further predicted that interface-resistance errors are proportional to surface roughness and are always present with harder buffers, typically reaching 010 m2.K/W for a mean roughness amplitude () of 200 μm. However with softer buffers these errors are absent below an onset level, typically at an value of 60 μm. Chapter 6 describes heat flow meter measurements and transient thermal modelling using ANSYS of a webbed, hollow-cored panel with silicone sponge buffer materials chosen to provide boundary conditions comparable to standard surface coefficients. Surface temperatures were also measured at eight locations for an uninsulated configuration as well as with bulk insulation filling. Measured and modelled temperature-time plots agreed well after corrections for web and airspace thermal conductivity. Modelled spatial variation in heat flow exceeded 200 % for one insulated case but was only about 2 % for the uninsulated panel. Modelled values for heat flux and overall thermal resistance agreed well with standard analytical calculations. However heat flows indicated by the apparatus were consistently higher than the modelled and calculated values by up to 8 %, expected to be due at least partially to specimen non-homogeneity. Nevertheless results suggest a useful role for the apparatus in providing temperature measurement under controlled conditions and helping to validate thermal modelling.
展开▼
