Small strain elastic behaviour of unsaturated soil investigated by bender/extender element testing

The aim of this project was to investigate very small strain elastic behaviour of soils under unsaturated conditions, using bender/extender element (BEE) testing. The behaviour of soils at very small strains has been widely studied under saturated conditions, whereas much less work has been performed on very small strain behaviour under unsaturated conditions. A suction-controlled double wall triaxial apparatus for unsaturated soil testing was modified to incorporate three pairs of BEEs transmitting both shear and compression waves with vertical and horizontal directions of wave transmission and wave polarisation. Various different techniques for measuring wave travel time were investigated in both the time domain and the frequency domain and it was concluded that, at least for the current experimental testing programme, peak-to-first-peak in the time domain was the most reliable technique for determining wave travel time. An experimental test programme was performed on samples of compacted speswhite kaolin clay. Two different forms of compaction were employed (i.e. isotropic and anisotropic). Compacted kaolin soil samples were subjected to constant suction loading and unloading stages at three different values of suction, covering both unsaturated conditions (s= 50kPa and s= 300kPa) and saturated conditions (s=0). Loading and unloading stages were performed at three different values of stress ratio (η=0, η=1 and η=-1 ). In some tests a wetting-drying cycle was performed before or within the loading stage, with the wetting-drying cycles including both wetting-induced swelling and wetting-induced collapse compression. BEE tests were performed at regular intervals throughout all test stages, to measure shear wave velocity Vs and compression wave velocity Vp and hence to determine values of shear modulus G and constrained modulus M. The experimental test programme was designed to investigate how very small strain shear modulus G and constrained modulus M varied with unsaturated state variables, including how anisotropy of these parameters developed either with stress state (stress-induced anisotropy) or with previous straining (strain-induced anisotropy). A new expression has been proposed for the very small strain shear modulus G of an isotropic soil under saturated and unsaturated conditions. This expression relates the variation of G to only mean Bishop’s stress p* and specific volume v, and it converges to a well-established expression for saturated soils as degree of saturation approaches 1. The proposed expression for G is able to predict the variation of G under saturated and unsaturated conditions at least as well as existing expressions from the literature and it is considerably simpler (employing fewer state variables and fewer soil constants). In addition, unlike existing expressions from the literature, the values of soil constants in the proposed new expression can be determined from a saturated test. It appeared that, in the current project at least, any strain-induced anisotropy of very small strain elastic behaviour was relatively modest, with the possible exception of loading in triaxial extension. It was therefore difficult to draw any firm conclusion about evolution of strain-induced anisotropy and whether it depended upon the same aspects of soil fabric as evolution of anisotropy of large strain plastic behaviour. Stress-induced anisotropy of very small strain elastic behaviour was apparent in the experimental test programme. An attempt was made to extend the proposed expression for G to include the effect of stress-induced anisotropy. Interpretation of the experimental results indicated that the value of shear modulus was affected by the values of all three principal Bishop’s stresses (in the direction of wave transmission, the direction of wave polarisation and the third mutually perpendicular direction). However, prediction of stress-induced anisotropy was only partially successful, and it was concluded that the effect of Lode angle was also significant.

The structural and mechanical integrity of historic wood

Little is known about historic wood as it ages naturally. Instead, most studies focus on biological decay, as it is often assumed that wood remains otherwise stable with age. This PhD project was organised by Historic Scotland and the University of Glasgow to investigate the natural chemical and physical aging of wood. The natural aging of wood was a concern for Historic Scotland as traditional timber replacement is the standard form of repair used in wooden cultural heritage; replacing rotten timber with new timber of the same species. The project was set up to look at what differences could exist both chemically and physically between old and new wood, which could put unforeseen stress on the joint between them. Through Historic Scotland it was possible to work with genuine historic wood from two species, Oak and Scots pine, both from the 1500’s, rather than relying on artificial aging. Artificial aging of wood is still a debated topic, with consideration given to whether it is truly mimicking the aging process or just damaging the wood cells. The chemical stability of wood was investigated using Fourier-transform infrared (FTIR) microscopy, as well as wet chemistry methods including a test for soluble sugars from the possible breakdown of the wood polymers. The physical properties assessed included using a tensile testing machine to uncover possible differences in mechanical properties. An environmental chamber was used to test the reaction to moisture of wood of different ages, as moisture is the most damaging aspect of the environment to wooden cultural objects. The project uncovered several differences, both physical and chemical, between the modern and historic wood which could affect the success of traditional ‘like for like’ repairs. Both oak and pine lost acetyl groups, over historic time, from their hemicellulose polymers. This chemical reaction releases acetic acid, which had no effect on the historic oak but was associated with reduced stiffness in historic pine, probably due to degradation of the hemicellulose polymers by acid hydrolysis. The stiffness of historic oak and pine was also reduced by decay. Visible pest decay led to loss of wood density but there was evidence that fungal decay, extending beyond what was visible, degraded the S2 layer of the pine cell walls, reducing the stiffness of the wood by depleting the cellulose microfibrils most aligned with the grain. Fungal decay of polysaccharides in pine wood left behind sugars that attracted increased levels of moisture. The degradation of essential polymers in the wood structure due to age had different impacts on the two species of wood, and raised questions concerning both the mechanism of aging of wood and the ways in which traditional repairs are implemented, especially in Scots pine. These repairs need to be done with more care and precision, especially in choosing new timber to match the old. Within this project a quantitative method of measuring the microfibril angle (MFA) of wood using polarised Fourier transform infrared (FTIR) microscopy has been developed, allowing the MFA of both new and historic pine to be measured. This provides some of the information needed for a more specific match when selecting replacement timbers for historic buildings.