Search

The swelling/shrinkage of spruce wood samples (Picea Abies) is documented with high resolution XRay Tomography and advanced image analysis tools. We report the reversible moisture-induced global and local deformations at the cellular and sub-cellular scales. In particular, we present sophisticated methods for detecting local deformations in the cell wall. Insight is given on the hygromechanical behaviour of wood cell material and on the role of ultra-cellular components in wood, such as bordered pits and rays.

This paper presents how simultaneous phase-and-amplitude retrieval during X-ray tomographic reconstruction allows a clear segmentation of sodium sulfate solution and hydrated sodium sulfate crystals in Savonnières limestone without the need for a dopant. This technique could be very useful to study salt crystallization processes at the pore scale, in order to improve the understanding and remediation of salt damage in porous media.

This article presents coupled data on saline transport, pore filling due to salt crystallization and resulting salt damage in Savonnières limestone. This is achieved by combining the non-destructive techniques of neutron radiography - for transport imaging - and X-ray microtomography - for pore structure and fracture visualization – applied to the same sample when subjecting it to consecutive wetting-drying cycles. Capillary uptake of water, 1.4m sodium sulfate and 5.8m sodium chloride solutions is studied at 25°C. An important decrease of the penetration coefficient for the salt solutions compared to water is found. Halite crystallization fronts are located based on neutron radiographs of the dried samples. Porosity analysis of the X-ray tomographic datasets quantifies pore filling at the position of the crystallization front. The damaging character of the salts is assessed from these datasets too. Only the sodium chloride wetted samples show fractures under the specific test conditions used (drying at 40°C).

Wood is a moisture sensitive material regarding durability (wood rot, mould growth) and also the mechanical behavior is highly influenced by variations in moisture content. Examples are the hygroscopic behavior of wood, the important swelling and shrinkage, especially in radial and tangential directions, and the decrease of stiffness and strength with moisture content. It is also believed that due to the swelling of wood, the moisture capacity of wood increases and that moisture transport properties change. It is found that the moisture sensitivity most importantly is limited to the range below the fiber saturation point, which indicates the saturation of the cell walls. Determining the dynamic stiffness on wood while loaded in compression, showed that the stiffness increases with the stress level (stiffening effect) and thus depends nonlinearly on the stress or strain level. In conclusion, wood exhibits a strong coupling between moisture and mechanical behavior and behaves fundamentally nonlinearly. In this paper, we propose a new material model for wood, based on a thermo-mechanical formulation of the changes in energy when a material is mechanically or hygroscopically loaded. The energy formulation includes the mechanical energy of the solid, the energy due to adsorption of the fluid and the energy due to fluid-solid interactions in the porous material. Traditionally, these energies include second order terms in the basic variables, strain and moisture content. Using the Clausius Duhem equation, incremental constitutive equations and material properties can be obtained by second order differentiation of the energy equation. The material properties include the stiffness, the moisture capacity and a coupling coefficient, and generally depend on strain and moisture content. In second order energy formulations, the nonlinearity of the material properties has to be determined from experiments. In this paper, we present a new higher order formulation of the energy, leading to nonlinear equations for the material properties on strain and moisture content. The advantage of a higher order energy formulation is that the nonlinear dependence of the material properties is inherently taking into account, leading to consistent formulations of the dependence of the material properties on strain and moisture content. However, these formulations lead to material properties in undrained conditions (sealed specimen with no change in moisture content). In reality, material testing on wood is generally executed in drained conditions, meaning equilibrium between vapor pressure of the environment and moisture content of the wood sample. Using a Legendre transformation, the constitutive equations and material properties can be formulated dependent on stress and capillary pressure (or chemical potential). This formulation leads to adequate nonlinear equations for the material properties, which can be determined from experiments. The model is applied to reproduce adequately free swelling due to moisture uptake and strain-stress relations for different moisture contents. The model can also explain the changes in internal stresses of a wood specimen exposed to varying relative humidity conditions.

Hygrothermal analysis of building components requires the knowledge of the convective heat and mass surface coefficients at exterior and interior building surfaces. Exterior surface coefficients are to a large extent influenced by the local wind speed near the surface. In most hygrothermal models, these coefficients are obtained from empirical formulae. They are used to relate the reference wind speed at a meteorological station to a local wind speed near the building surface and to relate the local wind speed to the heat surface coefficient. The vapor surface coefficients are generally determined from the heat surface coefficients using the Chilton-Colburn analogy. The existing empirical formulae for the heat surface coefficient as a function of wind speed however are based on only a limited number of measurements at a few facade positions and for a few building configurations. Little is known about the actual value and the variability of the surface coefficients across facades of different building geometries. In the past, most research on surface coefficients was experimental. Recently, numerical simulations based on Computational Fluid Dynamics (CFD) have been employed. Up to now, most simulations however have focused on interior surface coefficients. This paper presents high-resolution CFD simulations of exterior convective heat surface coefficients at the surface of a cubic building model. The simulations involve solving the entire structure of the boundary layer near the wall, including the logarithmic layer, the buffer layer and the laminar sublayer. Because of the high Reynolds numbers associated with wind flow around buildings and because the thickness of the boundary layer decreases with increasing Reynolds number, very high resolution grids, with cells as small as a few micrometers, have to be used near the building surface. The results show very distinct gradients across the facade. They indicate that the largest values of the surface coefficients occur at the edges, which are also the positions where wind-driven rain intensities are largest. This is of interest for the wetting-drying behavior of facades. Finally, it is shown that the low-resolution CFD approach, that has been used in some previous studies, and in which not the entire boundary layer is resolved, can yield some important discrepancies.

To realize a stable hygric indoor climate, hygroscopic moisture buffering by materials plays an important role. Not only finishing materials on indoor surfaces, but also textiles (curtains, pillows) and paper (books) show an important moisture buffering capacity. In this paper, the role of books in the stabilization of indoor climates is studied. Recently, a model for hygroscopic buffering by books has been presented (Derluyn et al. 2007) introducing an effective vapour permeability and moisture capacity. This model is applied for the modelling of small scale experiments performed on book samples of 40x40x20 mm. The samples are initially in equilibrium at 0 % RH and then subsequently exposed to a step in RH to 54% and 79.5% RH. In a next step the relative humidity decreases again to 54% RH and finally, a cyclic RH is imposed, fluctuating between 54% and 79.5% RH. During the test, the weight change of the samples is monitored. As paper clearly behaves hysteretic, a hygric hysteresis model is incorporated based on the independent domains theory. A good agreement between measurement and simulation is obtained. In a second experiment, the hygroscopic buffering of books on a shelf in a room is measured. First the climatic room is conditioned at constant temperature and relative humidity. The books are in equilibrium with the relative humidity of the room. Then, a moisture production protocol is imposed and the relative humidity in the room is monitored. A fan helps to obtain well mixed air conditions. Air speed, air temperature and relative humidity is measured at different positions in the room and around the books, specifically near to the book shelf and in between the book shelves. The weight change of the books is measured. The results are compared with a room without books on the shelves. The moisture balance for the room is modeled using two different approaches. In the first approach, well mixed air conditions are assumed, while in the second approach the air flow around the books is modeled using CFD and location dependent mass surface coefficients are determined. The simulation results for the two approaches are compared with the measurements.

Convective surface coefficients are commonly used in whole building simulation to model the indoor and outdoor heat and mass transport boundary conditions. There are some standards and guidelines available for convective heat transfer coefficients, which generally include correlations for simple geometries. The convective vapour transfer coefficients are not so well documented and are often calculated with the heat and mass transfer analogy, and then subsequently applied for cases in which the analogy is not valid. As a step towards the development of guidelines on the use of convective vapour transfer coefficients in hygrothermal models, an experimental setup and numerical modeling results are presented. An experimental setup was designed for the purpose of validating two numerical models: 1) a coupled CFD-diffusion model and 2) a vapour diffusion model. In the CFD-diffusion model, heat and mass transport in the air domain is solved using CFD, while, in the material domain, vapour transport is modelled. This model allows the prediction of the convective vapour transfer coefficient. It was found that the vapour transfer coefficient changes in time evolving to constant values depending on air speed, and to a lesser degree on the relative humidity of the air. In the vapour diffusion model, a convective vapour transfer coefficient was imposed and the change in moisture content was determined. The experiment consisted of a wind tunnel placed in an environmental chamber with climate control capability. The air flow in the wind tunnel was driven by a variable-control fan that allows for a range of speeds in the laminar regime. Convective vapour transfer coefficients were determined using the experimental setup for a number of humidity and air flow conditions over gypsum samples. The results were compared with the numerical results for the purpose of validation and sensitivity analysis.

Moisture transport in wood is complex due to its anisotropic material structure, resulting from the cells anatomy, growth rings where early- and latewood alternate and the presence of vessels, rays and pits. As a result, the moisture transport properties of wooden components are highly dependent on the direction (longitudinal, radial or tangential). Moreover, due to variations in cell geometry, and growth ring structure, the moisture transport properties, such as moisture capacity and vapor diffusion coefficient depend on location. The purpose of this paper is to develop a methodology to determine location dependent moisture transport properties and to study their influence on the global hygroscopic behavior of wood. In order to derive location dependent moisture transport properties, 2D micro structural images are obtained by Scanning Electron Microscopy of a full growth ring. The image is then thresholded to obtain a binary representation of the cell wall – cell lumen structure. The geometry of the cells is extracted and the cell borders are represented by lines. Based on the line configuration a mesh is generated. Finite element analyses are performed using a moving window methodology to obtain the position dependent material properties. A vapor pressure difference is imposed at two opposite sides, while the other sides remain vapor tight and the water vapor flux and water vapor diffusion coefficient are determined. These calculations are repeated for different relative humidities. The moisture capacity is determined based on the wall volume in the window. Different sizes of moving window are examined to find a suitable size of the representative elementary volume (area in 2D). In a second step, the location dependent moisture transport properties are used to analyze the water vapor transport at meso- and macroscale. The macroscopic moisture transport properties are then compared with measurements on wood. Further validation is performed by a adsorption/desorption test where the moisture distribution in the wood is measured by X-ray tomography.

It is known that solar driven vapor transport inward to vapour open constructions such as wood frame walls with a vapour tight finishing at the inside may occur and may lead to an undesired wetting of wooden elements and gypsum board. Although the origin of this inward vapor transport may be well understood, the quantification of the moisture damage risks remains not well understood, since the effect of material properties and climate on the two-dimensional hygrothermal behaviour of the walls is not yet adequately addressed. Therefore, a research project, subsidized by ASHRAE TC4.4, studies solar-driven inward vapor flow combining small and large-scale wall assemblies tests under comparable conditions. The objectives of the tests were to produce precise measurements on wetting and drying behavior of internal parts of the back wall, when a wetted masonry cladding is loaded under cyclic temperature loading. The specimens were composed of: brick veneer, weather resistive membrane (spun-bonded polyolefin), oriented-strand board (OSB) sheathing, glass fiber insulation, gypsum board and interior finish (acrylic paint or vinyl wall covering). The small specimens measured 400 mm and 400mm and the large specimens, 1200 mm on 2400 mm. The small specimens had their brick veneer wetted at 50% saturation, were covered with an aluminum plate and then were subjected to cyclic temperature conditions for a total test period of. The large specimens saw their brick veneer rewetted daily to the same 50% saturation, and then exposed to 8 hours of simulated solar radiation for periods from 3 to 6 weeks depending on the specimens. The loading conditions are quantified to allow comparison. It is found that, due to the temperature gradient, an important vapor flow is generated to the back wall, wetting the OSB sheathing, mineral wool and gypsum board. The vapor permeance of the interior finishing determines the wetting of the gypsum board: a vapor tight vinyl wall covering leads to significant wetting of the gypsum board and very low drying rates during drying. The moisture contents in the gypsum board are compared. The moisture behavior of small and large tests is modeled using HAM models and compared. The main differences, due to the 3-dimensional effects observed in the large specimens, are demonstrated and explained by convection loops occurring in the insulation.

This paper presents a stochastic analysis of moisture buffering in a room. The relative humidity in a room depends on the ventilation rate (or supply of air with a different RH and temperature), the moisture production rate and the moisture uptake/release rate by finishing materials, also called moisture buffering. The ventilation rate, the outside vapour pressure and temperature, the moisture production rate are essentially random and can be considered as stochastic variables. In this paper, the moisture transport properties of the surface materials and the surface transfer coefficients are considered to be deterministic. As a measure for moisture damage risk, referred to as damage indicator, we consider the difference in maximal and minimal relative humidity during a day. Since the inside relative humidity variation can be considered as a stochastic process, the damage indicator is also random and can be described by a cumulative distribution function. As performance criterion in a reliability approach, we can then define a maximum risk of exceeding a limit value for the daily variation in relative humidity. The cumulative distribution of daily variation in relative humidity depends on the moisture buffering capacity of the surfaces. We found that the moisture buffering capacity of a room is related to the reduction of the moisture damage risk, when comparing a room with and without moisture buffering capacity. As long as the average moisture production rate and ventilation rate are maintained, the moisture buffering capacity seems not to be influenced by the randomness in the stochastic variables. Using a penetration depth model, an average effective surface mass coefficient for the room is defined, taking into account the vapour transport from air to the surface, through paint layer and finally to the active surface material layer. A relation between moisture buffering capacity and effective surface mass coefficient has been demonstrated for a given room, with geometric characteristics, such as the total surface to volume ratio, with a given average ventilation rate and moisture production rate.

Rain is the main source of moisture on the building envelope. Wood cladding, window wood frame and, in the case of water infiltration, even the wood frame structure can be in contact with water. In the context of the development of a multi-scale moisture transfer model for wood components of the building envelope, it is found that one element, namely the pit of the wood structure plays a critical role in water transfer. Hundreds of pits, small porous openings between adjacent wood cells, allow for water to move along the cells, across and along growth rings. It seems that the permeability of wood would depend largely on the permeability of these pits. Liquid transfer models for wood require the knowledge of the liquid permeability of the pits. Information about pit geometry in literature is rather limited, and relatively little information about pit permeability has been provided. This is at least partly due to the fact that experimental data for this small and fragile microcellular structure are very difficult to obtain. Pits are characterised by a rather complex geometry. They consist of a central impermeable torus, which is held in place by the margo, which is a loose and complex mesh of microfibrillar strands through which water can pass. As a result of this complex geometry, the dynamics of water flow is also quite complex. This paper will present Computational Fluid Dynamics (CFD) models for a range of generic pit geometries. The models take into account the complex network structure of the margo by the detailed implementation of each of the individual microfibrillar strands. In this paper, the construction of the CFD models will be presented, the calculation results in terms of liquid permeability will be provided and their applicability for different wood species will be discussed. This paper also discusses issues with modelling porous materials with CFD.

This paper presents hygrothermal modeling results simulating the drying performance of different wood frame wall configurations submitted to water infiltration in the stud cavity. As rain can be the most important potential source of moisture in wall assemblies, an experimental investigation was previously undertaken at Concordia University to compare the wetting and drying potential of 15 walls subjected to water insertion in the stud cavity within a large Environmental Chamber. The experiment considered three different types of sheathing – oriented strand board (OSB), plywood and fiberboard, two exterior claddings – wood cladding on furring and cement stucco on metallic lath (no air space), and two different interior finish permeances – paint with/without polyethylene sheet. The moisture distribution in each wall was monitored using moisture content probes and gravimetric measurement and the conditions (exterior and interior) were controlled to reproduce August through November weather conditions for Montreal. The moisture load consisted in a wetted sill plate inserted into the wall assembly. The initial conditions for the simulation include the measured moisture distribution in the wood sill. The changes in moisture distribution in the wood sill and in the rest of the wall assembly are simulated taking into account proper boundary conditions. The simulation results are compared with the measured results. It was found that the type of cladding has the most impact on the drying performance. The role of the type of sheathing and of the level of interior finish permeance are also demonstrated and explained.