Experimental assessment of the similarity law for a one-dimensional coupled heat and water vapor diffusion in hemp concrete.

• Similarity laws are verified in the framework of a one-dimensional heat and water vapor diffusion problem through an experimental campaign. • Two configurations of hemp concrete samples, equivalent in terms of similarity laws, are exposed to thermal and hydric stress. • Temperature and relative humidity at equivalent positions in both samples are measured and compared. • A detailed uncertainty investigation is proposed to discuss the reliability of the observed discrepancies. • The comparison of discrepancies and entire uncertainties show a good agreement between the reference and reduced configurations. Similarity law is a mathematical mapping that aims to set up equivalences between different configurations regarding the involved physical parameters, space dimensions and the duration of the configuration. These equivalences are defined by means of the dimensionless formulation of the physical model that describes a phenomenon. This concept is widely adopted in some scientific fields, notably in computational fluid dynamics. In heat and water vapor diffusion, similarities are applicable theoretically on configurations distinct in terms of material, space dimensions, duration or even imposed stress magnitude. It is worth noting that these are slow kinetic phenomena. In this regard, this study aims to verify similarities experimentally while using hemp concrete and reducing the duration of a reference experiment. Before performing measurements, two main limitations are fixed. First, by means of an experimental design, heat and moisture transfer are assumed one-dimensional. In addition, Dirichlet boundary conditions are considered at the surface that is exposed to ambient air. Mainly, temperature and relative humidity are measured at the quarter and middle of the material in order to compare the two similar configurations. Afterwards, an uncertainty study is carried out in order to discuss and justify the observed discrepancies. Following a deterministic approach, this investigation considers five sub-uncertainties which are related to : the calibration of used sensors, their positions, time of response, randomness of an experiment and the boundary condition modeling. Results show a good agreement between the reference and reduced configurations. Indeed, the discrepancies between both are entirely within the calculated confidence intervals. [ABSTRACT FROM AUTHOR]