Description Measurements in loco of thermal performance of building components are essential to quality assurance of new buildings and energy performance evaluation of existing buildings. Such measurements are complex, time consuming and have large uncertainties. This project explore the use of heat and mass transfer theory, calibration techniques and machine learning algorithms to improve the evaluation in loco of thermal properties of building components. Key Objectives Investigate the use of machine learning techniques to speed-up evaluation using current measurement techniques Investigate the use of numerical simulation and genetic algorithm optimization to infer thermal properties Investigate the use of infrared imaging to infer thermal properties Investigate minimal duration of measurement campaigns to assure reliable results