No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

A general scheme for time-delay control systems is presented so that with particular choice of system elements and loop gain many of the existing time-delay systems can be recovered. The effects of imperfections in the system model for time-delay systems are described and analysed. Sensitivity expressions are obtained and their relationship with the structure of time-delay systems are explored. Sensitivity Analysis for a time-delay system is extended to Adaptive Control of time-delay systems. An adaptive control system for Smith's Predictor and an LOP Optimal Control scheme are investigated analytically and by simulation. Delay-free optimal control methods are extended to systems with delay for LQP and parametric optimisation. For parametric optimal control, we explore, in particular, cases where deliberate mismatch leads to improvement, and the improvement of delay-free systems by addition of delay is also investigated. Various cases of LQP are discussed and the sub-optimality of published schemes, induced by mismatch is exposed. The method is also extended to time-varying systems with time-delay.

This thesis contains a study of improvement and application of strain measurement technique using neutron diffraction. A dedicated neutron strain scanner - ENGIN, funded by Ee. has been developed at ISIS. It has two radial collimators, allowing for the first time, measurements taken simultaneously at two orthogonal directions; uses Pawley refinement permitting both whole pattern and individual peak profile analysis and its positioner allow's specimens weighing 250 kg to be placed with an accuracy of 100 um. Strain measurement using neutron diffraction has been investigated under two extreme circumstances: very shallow (within 1 mm near surface) and very deep in materials (hundred of mm). Near-surface measurement requires both the accurate determination of the effective measurement position associated with precise location of specimen, calculation of centroid and correction for an anomalous near-surface effect. The strain measurements on a shot-peened surface in titanium alloy were carried out using the methodology mentioned above. The result is comparable to that obtained from X-ray diffraction. The latter was exploited by studying the effect of wavelength-dependent attenuation in materials. Experiments and theoretical analysis on aluminium and iron show that the effect is small for a strain scanner using radial collimator and time-of-flight technique. However, the reduction of the diffraction peak intensity, as a function of the amount of material in the beam path, reveals that great care should be taken when measuring texture as a function of depth in materials. The neutron diffraction technique was applied to several sets of engineering strain measurements. Firstly, 3-D residual stresses surrounding a cold expanded hole in a high strength aluminium alloy plate were measured. The result agrees well with that from modified Sachs' method after taking the effect of gauge volume averaging into account. Second, measurements of strain distributions in a Q-joint under zero and 30 kN load have been carried out, which provides a confirmation of the excellence of the joint design as the fastener system has completely shielded the fastener hole from deleterious tensile strains under operational loading. Finally, calibration strain measurement for an energy dispersive neutron transmission spectrometer was performed for the first time and it is found that the transmission spectrometer is feasible for strain and phase transformation measurement.