A Computational Approach to Predicting Thermal Conductivity of Construction Material: Towards Rapid In-Situ Energy Evaluation of Existing Buildings

Thermal conductivity of the building envelope is the basis for building energy simulation and gives key information about its energy performance. Accurate assumptions about thermal conductivity contribute to accurate energy labelling of buildings, provide insights for retrofit strategies and are instrumental in energy policy making at an urban level. However, physically establishing the thermal conductivity of the external walls of a building is difficult and is often a time-consuming process. Currently available measurement methods are extremely technical with restrictive boundary conditions. To avoid this dilemma, thermal conductivity is often inferred from published standards, which gives significantly different values compared to in-situ measurement. It is therefore important not only to measure

thermal conductivity of buildings in-situ, but also to make the means for evaluating thermal conductivity ubiquitous, accessible and uncomplicated for architects, surveyors and building engineers. Aiming to bridge the identified gap, this thesis presents a study of material properties, how they interrelate, and how these relationships can be exploited to assess the building fabric. It presents a computational approach, which is a data-driven method motivated by

experimental results. In this method thermal conductivity is predicted using computation on experimental data of relevant material properties rather than direct measurement. During this work, extensive experiments were conducted to measure thermal conductivity, dielectric and mechanical material properties to determine their correlation. These experiments were conducted on two categories of materials. One that represents wood-frame construction, which is the most prevalent form of construction for residential use here in the United States. The materials studied for this category included solid wood, plywood, OSB, chipboard, MDF

and gypsum drywall. The other category of materials is the ceramic family and is representative of clay brick construction. Clay bricks, concrete, naturally occurring stone and gypsum were included in this study. Apart from the categories described above, work was also conducted on multiple-layered materials so that the effect of stacking layers together could be studied. A final study was performed to explore the effects of moisture on materials

thermal conductivity and dielectric properties. The empirical data collected during this study suggests a strong correlation between thermal conductivity and dielectric properties. This correlation has not been systematically studied previously and experimental data on the subject is extremely scarce. The correlation found through this study identifies the potential of using handheld meters or antenna-based devices that can quickly measure capacitance and dielectric properties, to measure thermal conductivity instead. If successful, this method will eliminate the need to create steady-state conditions required for thermal conductivity measurement and simplify the measurement

enough to bring a useful gadget not only for experts, but also for the layperson.