Thermal conductivity of a food material is an essential physical property in mathematical modelling and computer simulation of thermal processing. Effective thermal conductivity of non-homogeneous materials, such as food matrices, can be determined experimentally or mathematically. The aim of the following research was to compare the thermal conductivity of short pastry biscuits, at different baking stages (60-160 min), measured by a line heat source thermal conductivity probe and estimated through the use of thermo-physical models. The measures were carried out on whole biscuits and on powdered biscuits compressed into cylindrical cases. Thermal conductivity of the compacted material, at different baking times (and, consequently at different moisture content), was then used to feed parallel, series, Krischer and Maxwell-Eucken models. The results showed that the application of the hot wire method for the determination of thermal conductivity is not fully feasible if applied directly to whole materials due to mechanical changes applied to the structure and the high presence of fats. The method works best if applied to the biscuit component phases separately. The best model is the Krischer one for its adaptability. In this case the value of biscuit thermal conductivity, for high baking time, varies from 0.15 to 0.19 Wm<sup>–1</sup> K<sup>–1</sup>, while the minimum, for low baking time, varies from 0.11 to 0.12 Wm<sup>–1</sup> K<sup>–1</sup>. These values are close to that reported in literature for similar products.
Apparatus for routine measurements of the thermal conductivity of ice cores
