NMR imaging of moisture inside heated porous building materials

Abstract

In this thesis we have presented a dedicated NMR imaging setup as a valuable tool in visualising the moisture transport processes in heated porous building materials. With the experiments on three different building materials, we were able to validate the NMR measurement technique. The validation experiments itself provided interesting results in the field of non-isothermal moisture transport in porous materials. Although concrete is one of the most fire-resistant materials as compared to, e.g., wood and iron, after several severe fires in European concrete tunnels (Channel Tunnel, Gotthard Tunnel, Mont-Blanc) it has become clear that concrete can be heavily damaged under the impact of high temperatures. The damage processes are generally known as fire spalling of concrete. The term spalling refers to the (explosive) delamination of the surface. Pieces of concrete are ejected from the surface, thereby decreasing the thickness and hence the strength of a wall or column. Moisture and vapour transport in concrete during a fire is one of the majour causes of spalling. Despite this fact, there is almost no direct experimental data available on moisture distributions and moisture transport under elevated temperatures inside concrete, or any other building material. The aim of this study was to investigate the physical mechanisms underlying the moisture transport in concrete during heating. To this purpose, we have introduced a dedicated NMR setup capable of measuring moisture and temperature profiles during one-sided heating of building materials. In Chapter 2 we have characterised the three building materials which are used in this thesis: fired-clay brick, gypsum, and concrete. MIP and sorption isotherms have been used to characterise the pore size distribution of the materials. TGA and DSC have been used to determine the behaviour of these materials at higher temperatures. The details of the new dedicated NMR setup have been presented in Chapter 3. In order to obtain a quantitative moisture content of a porous material the NMR signal needs to be corrected for the temperature at which it was measured. An overview of the temperature influence on the nuclear magnetisation and the relaxation mechanisms was presented. A temperature correction method for the NMR signal was introduced and demonstrated. A simple vapour transport model was presented in Chapter 4. The model is based on the balance between the vapour released at the boiling front and the vapour transport towards the surface. The model was able to predict very basic and well established experimental observations from large scale fire testing. NMR heating experiments on three different saturated porous materials: fired clay brick, calcium silicate brick, and concrete were used to validate the model. From the measured moisture and temperature profiles the boiling front positions and the corresponding temperatures could be obtained. The model was able to fit the boiling front movement in these materials with the vapour permeability as the only free parameter. A short theoretical overview of the vapour transport processes in concrete is given in Chapter 5. A Stefan tube was used to illustrate the different vapour transport processes: diffusion, Knudsen flow, and viscous flow. These limiting cases of vapour transport each occur at different partial vapour pressures and capillary sizes. A combined heat and mass transfer model is presented. In this model, unlike the simple sharp front model of Chapter 4, the latent heat, liquid moisture transport, and vapour transport in both directions is taken into account. The model calculations show three characteristic features of the moisture transport process in concrete. First, a boiling front is moving from the surface further into the sample. Second, an increase in moisture content in the ‘cold’ region behind the boiling front. Third, a significant increase in vapour pressure close to the boiling front. The heating power, and stability of the NMR setup over a large range of moisture contents was verified in Chapter 6. Fired-clay brick was used for this purpose because it is an inert material. There is no thermal degradation and dehydration of the porous material, which significantly simplifies the moisture transport. The moisture transport during non-isothermal drying under varying heating rates of fired-clay brick was measured. In the measured moisture profiles two main drying stages can be identified, i.e., a first stage with essentially homogeneous drying followed by a second stage which is characterised by a receding drying front. It was shown that the duration of the first drying stage and the speed of the drying front in the second drying stage could both be scaled with the applied heating power. In Chapter 7 the moisture transport and dehydration of gypsum was studied. Chemically bound moisture is a significant source of vapour in a porous material such as gypsum. The dehydration reaction is a one or two step process depending on the relative humidity, taking place at temperatures between 100oC and 250oC. There are a number of important observations during the one-sided heating of gypsum. First, the chemically bound water could be measured with the NMR setup. Second, two separate dehydration fronts corresponding to the two dehydration reactions were observed. Third, the vapour released by the first dehydration reaction results in the development of a moisture peak. The vapour transport model of Chapter 5 was used to qualitatively describe the observed dehydration and moisture profiles. After a thorough characterisation and validation of the NMR setup, the one-sided heating experiments on concrete were presented in Chapter 8. Because of the important influence of moisture in the fire spalling process we have presented a set of experiments with moisture contents varying from 50 to 97% RH. In Chapter 8 we presented three important observations of the heating experiments. First, changing the moisture content showed no significant influence on the heat transfer in concrete. Second, immediately after the heating started, a boiling front was formed and moves through the sample. Third, direct and quantitative proof of the formation of a moisture peak was presented. Fourth, in two of the samples, 86 and 97% RH, the increase in moisture content resulted in the formation of a saturated layer. A process which is known as moisture clogging. The temperatures at the boiling front were used to estimate the vapour pressure, which amounts to a maximum of 1.9 MPa compared to 0.1 MPa atmospheric pressure. Finally, the simplified vapour transport model of Chapter 4 was used to describe the movement of the boiling front during the heating experiment.