The work described in this thesis consists of three phases: Phase I - Uniaxial analysis: A computer program has been written in a form suitable for calculating the failure loads for a large number of eccentrically-loaded, circular or rectangular concrete-filled columns, using the uniaxial material properties and the part cosine-wave assumption. The experimental failure loads of 22 eccentrically-loaded square and rectangular columns are compared with the computed loads, and a satisfactory agreement is found, i.e. triaxial effects are negligible for the columns tested. Phase II - Stub columns: The elasto-plastic biaxial stresses in the steel are calculated from observed strains for 14 of the available tests on concentrically-loaded stub columns, using the generalised flow-law for plastic solids. The triaxial stresses in the concrete are then calculated from simple statics. It is found that near failure the longitudinal compression in the steel is approximately equal to threequarters of the uniaxial yield stress, and the hoop tension is half the longitudinal compression in magnitude, while the longitudinal compression in the concrete core is twice the strength of uncontained concrete. The equivalent longitudinal 'stress-strain relationships' for the steel and concrete are represented by equations, and a formula is given for predicting the failure load of concentrically-loaded stub columns. Phase III - Triaxial analysis: Moment-load-curvature characteristics are experimentally determined for 35 circular columns comprising 7 tube thicknesses with 5 levels of axial load for each thickness0 Triaxial effects are found to be insignificant when the axial load is less than 40 per cent of the sum of the uniaxial compressive strengths of the steel and concrete. For higher axial loads, triaxial effects are taken into account by using the equivalent stress-strain relationships of Phase II in a semi-rational analysis. All the experimental moment-load-curvature characteristics are numerically simulated on a digital computer.

The propagation of stress waves in long polymer rods forms the basis of two major experimental techniques. The first is a modified Split-Hopkinson pressure bar (SHPB) arrangement that employs polymer Hopkinson bars (as opposed to metallic bars) in order to determine the high strain-rate mechanical properties of soft materials. The second experimental technique consists of a group of methods for determining the viscoelastic properties of polymer rods within a frequency range of 20 Hz to 30 kHz. An experimental, analytical and finite element study of stress waves propagating in viscoelastic rods is reported. A propagation coefficient is used to account for the attenuation and dispersion of stress waves propagating in polymer rods. Through experimental investigations, an optimal experimental arrangement is used to determine the propagation coefficient of a PMMA rod with an improved level of accuracy in comparison with results available in the open literature. Analytical investigations show difficulties associated with experimental arrangements as well as the numerical procedure adopted which tend to reduce the accurate frequency range of the determined propagation coefficient. The FE analysis of stress waves propagating in polymer rods suggests end effects are important; these end effects are not accounted for in any analytical bar wave theory. The high strain-rate mechanical properties of Hydroxyl-terminated polybutadiene (HTPB) are measured via a viscoelastic SHPB set-up. A scheme for processing the strain signals from the tests that allows for large strain measurement (approximately 60%) is presented. The use of viscoelastic SHPB set-up is able to produce a more sensitive measurement when compared with test results in the literature which are obtained using conventional metallic bars. A Finite element model of a viscoelastic Hopkinson bar set-up is developed. The applicability of the model in viscoelastic SHPB testing is validated.

The telescoping cantilever beam structure is applied in many different engineering sectors to achieve weight/space optimisation for structural integrity. There has been limited theory and analysis in the public domain of the stresses and deflections involved when applying a load to such a structure. This thesis proposes (a) The Tip Reaction Model, which adapts classical mechanics to predict deflection of a two and a three section steel telescoping cantilever beam; (b) An equation to determine the Critical buckling loads for a given configuration of the two section steel telescoping cantilever beam assembly derived from first principles, in particular the energy methods; and finally (c) the derivation of a design optimization methodology, to tackle localised buckling induced by shear, torsion and a combination of both, in the individual, constituent, hollow rectangular beam sections of the telescopic assembly. Bending stress and shear stress is numerically calculated for the same structure whilst subjected to inline and offset loading. An FEA model of the structure is solved to verify the previous deflection, stress and buckling predictions made numerically. Finally an experimental setup is conducted where deflections and stresses are measured whilst a two section assembly is subjected to various loading and boundary conditions. The results between the predicted theory, FEA and experimental setup are compared and discussed. The overall conclusion is that there is good correlation between the three sets of data.