Non-destructive evaluation of thermal conductivity of wood and insulation materials using permittivity measurements with ground penetrating radar

• Data-driven method predicts thermal conductivity from dielectric properties via GPR. • High accuracy in estimating thermal conductivity for various woods and insulation. • Linear correlation between dielectric properties and thermal conductivity at 2.6 GHz. • Simplifies thermal evaluations for building walls without direct conductivity tests. The accuracy and effectiveness of heat transfer predictions in construction materials are important for energy performance evaluation and provide significant insights in building inspections. However, directly establishing the thermal conductivity of building walls is challenging and time-consuming, with existing methods being highly technical and restricted by conditions, especially at low wind velocities and large temperature differences. Therefore, in this study, the thermal conductivity is derived from relative permittivity measurements made using a nondestructive technique, and compared with direct measurements. The experiments were conducted on five types of wood and two types of insulation material using a ground-penetrating radar (GPR) system at a frequency of 2.6 GHz. The results showed a linear correlation between the dielectric properties and thermal conductivity in both dry and wet states, regardless of the wood species or types. This high-performance correlation suggests that thermal conductivity properties can be accurately estimated from permittivity measurements at high frequency, demonstrating the potential for estimating thermal conductivities of 0.05–0.3 W m −1 K −1 using GPR. The relationship between the moisture content and the dielectric and thermal properties of some types of wood was discussed, which simplified the measurement process, eliminated the need for direct thermal conductivity assessments, and opened the way for thermal performance evaluation of building walls on-site.