Resumen:
There is a growing interest in characterising the thermal performance of building envelopes when exposed to realistic weather and indoor conditions. In this study, data from a full-scale test of four uninsulated concrete panels is analysed using (1) a steady-state model as per the standard average method, (2) a dynamic lumped resistance-capacitance model with a stochastic method, and (3) a dynamic distributed capacitance model based on an analytical solution. These have been favoured over purely data-driven methods, since their physical formulation allows the characterisation of thermal capacity alongside the usual thermal resistance. The models are applied to different data subsets, sampling times and campaign lengths. For the sole estimation of thermal resistance, winter conditions with constant indoor heating allow campaign lengths around 72 h. For a strong indoor-outdoor temperature difference (e.g. 10 degrees C) steady-state models provide reliable estimates, and lumped capacitance models are found to suit lower temperature differences or less stable conditions. However, for estimating thermal capacity, fluctuating indoor and outdoor temperatures are preferred and only the distributed capacitance model provides consistent estimates for different time steps and data subsets. The present work might be helpful in establishing future guidelines for the use of dynamic methods with physical interpretation, presenting a case study of a simple well-known wall facing a variety of winter and summer conditions. It might also provide a basis for further research, extending the application of these models to more complex multi-layer walls and/or for the assessment of design scenarios including thermal insulation.
