Moisture Dynamics in Building Envelopes

The overall scope of this Thesis "Moisture dynamics in building envelopes" has been to characterise how the various porous insulation materials investigated performed hygrothermally under conditions similar to those in a typical building envelope. As a result of the changing temperature and moisture conditions in the exterior weather and indoor climate the materials dynamically absorb and release moisture. The complexity of the impact of these conditions on the resulting moisture transport and content of the materials has been studied in this Thesis with controlled laboratory tests. The first part of the Thesis consists of a theory and literature review on the moisture storage and transport processes (Chapter 2), on the non-Fickian moisture transport (Chapter 3) and on the methods for determining the moisture properties (Chapter 4). In the second part, the conducted experimental work, results, and analysis are presented (Chapters 5-7). The major findings are discussed (Chapter 8), before the final conclusion (Chapter 9). The Appendices include the material parameters used, some additional results and the description of the simulation models. Chapter 2: Moisture transport theory The focus of this study concerns the dynamics of moisture transfer and storage in the hygroscopic range, and on the effects of temperature gradients on such transfer. An overview of the theory and literature of these mechanisms is given in this chapter. Chapter 3: Non-Fickian moisture transport The literature review presents a number of authors that have studied non-Fickian moisture transport. The assumption of immediate local equilibrium has been generally rejected and moisture transport and storage were divided into a pore air phase and absorbed moisture phase. The link between the air and absorbed phase was modelled by a sorption equation, where the moisture flux was determined by a proportional coefficient and moisture potential. The focus in these studies has been on the development of transport equations. Chapter 4: Determination of moisture properties Some experimental methods for determination of two of the most important moisture properties are described: the moisture capacity and the transport coefficients. The determination of diffusivity is also treated. These methods refer only to the hygroscopic range, i.e. ' < 0:98. Chapter 5: Isothermal, dynamic moisture transfer This chapter describes experimental and numerical approaches to quantify the time-dependence of sorption mechanisms for some insulation materials. The experimental part included measurements with two different set-ups, where small samples were exposed to ab- and desorption steps in a controlled relative humidity and temperature environment. Changes in the bulk moisture content were continuously followed as the sample was attached directly to a balance. The experimental results showed retarded sorption, most clearly for organic insulation materials. An exception was cellular concrete, whose sorption behaviour showed almost 3 4 only Fickian characteristics. The results were analysed theoretically to quantify some characteristic parameters, e.g. moisture diffusivity, penetration depth and moisture buffer capacity. The sensitivity of the results due to the used method was shown, e.g. the moisture buffer capacity was over-estimated when using steady state properties. The results were also analysed numerically and a model for non-Fickian moisture transport was developed. The traditional assumption of immediate local moisture equilibrium was rejected when modelling dynamic moisture transport. Instead, separate nodes for air moisture in the pores of porous materials and the absorbed moisture were modelled. The link between these nodes, which retard sorption, is described by a sorption equation. This is in accordance with approaches of other researches presented in Chapter 3. An approach for determining the sorption coefficient in this sorption equation experimentally was also shown. This preliminary approach for determining a sorption coefficient, which can model retarded sorption, was encouraging. Chapter 6: Non-isothermal, steady state moisture transfer An experimental investigation was conducted in order to draw some conclusions on the magnitude of moisture transport due to temperature gradient on a range of porous light-weight building materials. A special constructed non-isothermal set-up allowed the creation of a temperature gradient of 10K and given humidity gradient over the sample. The resulting moisture flux as well as the hygrothermal states around and within the material was monitored. The results showed that there exists some kind of ’other’ transport in addition to ¢p- driven one in all the materials analysed. Rather surprisingly, all the materials, including the almost non-hygroscopic materials (e.g. rock wool) and very hygroscopic materials (e.g. cellulose insulation) showed the same characteristics. The hypothesis of relative humidity being a driving force for non-isothermal moisture transport already in the hygroscopic could not be confirmed. On the contrary, indications exists that the temperature gradient itself is driving the moisture from the warm side towards the cold side. An attempt to identify and quantify the single contributions of the different transport forms involved was also presented. Chapter 7: Non-isothermal, dynamic moisture transfer The set-up (the same one as in Chapter 6) was now used to create a dynamic climate with sinusoidal oscillations of relative humidity on the cold side over a period of 24 hours. The aim of these measurements was to identify the dynamic moisture response of a material exposed to a temperature gradient. The experimental results were compared with dynamic simulations partly with a ’conventional’ model, and with a non-Fickian model. Materials like cellulose and flax insulation and cellular concrete were able to moderate the oscillations. The difference in peak values of RH between the materials, however, was not significantly unambiguous. There existed a fairly good agreement between the measurements and the conventional Fickian model used. A minor phase delay on simulated results indicated that the true moisture capacity of the materials was lower than the mathematical one, i.e. the slope of the sorption isotherm. Implementing a hysteresis model increased the agreement between measurements and simulations but was not able to remove all the deviation. Implementing the non-Fickian model for this non-isothermal set-up did not give any good agreement with the measurements as it overly underestimated the moisture capacity. The moisture buffer capacity of the materials was also assessed. It was shown that the ’measured’ buffer capacity was higher than the theoretical buffer capacity for the materials with good buffer capacity: flax and cellulose insulation and cellular concrete. There was no difference for materials with poor buffer capacity: glass wool, rock wool and perlite.