Ultrasonic Characterisation of the Structure and Properties of Wood.

Wood is a versatile and strong engineering material which displays a high level of anisotropy. Modelled as an homogeneous orthotropic solid, its natural variability and unique micro structure presents a serious challenge to the application of standard nondestructive evaluation techniques for quality assessment. In this work the characterisation of wood using ultrasonic techniques is considered. Addressing the problem is seen as requiring knowledge of three distinct areas - the structure of wood, the theoretical basis of elastic wave propagation in an orthotropic solid and ultrasonic measurement science. This forms the basis of a critical review of the literature pertaining to ultrasonic characterisation of wood. A systematic study of ultrasound velocity and attenuation over a range of frequencies is reported in chapter 6 of this work. From this study which examines three species of wood with different microstructures a number of significant results emerge. The species are seen to display different levels of anisotropy. Velocity is seen to increase significantly with frequency, particularly for some species. Attenuation is seen to be frequency dependent. The results point to the significance of the inhomogeneity of wood. The variation of ultrasound properties from pith to bark and at different levels within the tree is also measured. The orthotropic model is extended to include the effect of the inhomogeneity due to the annual ring structure in wood. This is modelled as a quasi-periodic structure. A novel set of measurements of acoustic inhomogeneity in the longitudinal and tangential directions, which can resolve within ring variations in wavespeeds, is presented. These measurements are used, along with data from the literature to show that the potential stop and pass bands may exist for waves travelling in the radial direction. A phase cancellation artefact and other interference effects are seen to exist for propagation in the longitudinal and tangential directions. Finally artificial neural network analysis is used to show that ultrasonic signals may be classified by the species through which they have propagated. This method is seen as having potential for engineering applications