Search

CRESST has an established analysis procedure to evaluate the energy of the events it detects, in an attempt to detect WIMP dark matter. It was shown that unless eight classes of contaminant event were removed prior to this analysis, the output energy spectrum would be significantly biased. For both scientific and practical reasons, the removal process should be blind, and a series of cuts were developed to flag these events automatically, without removing any true events. An event simulation package was developed to optimise these cuts. It was shown that noise fluctuations could also reduce CRESST’s sensitivity, so a noise-dependent acceptance region was introduced to resolve this. The upgraded CRESST experiment included a new electronics system to provide heating and bias currents for 66 detectors. This system was integrated into the CRESST set-up, and it was shown that the electronics contributed no extra noise to the detectors. Data with an exposure of 50 kg days were analysed using the cuts and the noise-dependent acceptance. The cuts were successful, with no contaminant event retained and a live time reduction of just 2.3%. The data were used to set an upper limit on the WIMP-nucleon cross section for elastic scattering with a minimum of 6.3 × 10^(−7) pb at a WIMP mass of 61 GeV. This is a factor of 2.5 better than the previous best CRESST limit.

The dynamics of a flexible air vehicle are typically described using an aeroservoelastic model which accounts for interaction between aerodynamics, structural dynamics, rigid body dynamics and control laws. These subsystems can be individually modeled using a theoretical approach and experimental data from various ground tests can be combined into them. For instance, a combination of linear finite element modeling and data from ground vibration tests may be used to obtain a validated structural model. Similarly, an aerodynamic model can be obtained using computational fluid dynamics or simple panel methods and partially updated using limited data from wind tunnel tests. In all cases, the models obtained for these subsystems have a degree of uncertainty owing to inherent assumptions in the theory and errors in experimental data. Suitable uncertain models that account for these uncertainties can be built to study the impact of these modeling errors on the ability to predict dynamic instabilities known as flutter. This thesis addresses the methods used for modeling rigid body dynamics, structural dynamics and unsteady aerodynamics of a blended wing design called the Body Freedom Flutter vehicle. It discusses the procedure used to incorporate data from a wide range of ground based experiments in the form of model uncertainties within these subsystems. Finally, it provides the mathematical tools for carrying out flutter analysis and sensitivity analysis which account for these model uncertainties. These analyses are carried out for both open loop and controller in the loop (closed loop) cases.

Phased arrays have many applications such as Radar Communication, Satellite Communication, and Wireless Local Area Networks (WLAN). For the traditional phased array, a phase shifter is used with each antenna element to establish a constant phase progression along the antenna array. A constant phase progression will force the electromagnetic wave to add up so that the energy would radiate at a particular angle with respect to the array. However, it is difficult to integrate the bulky phase-shifters in the monolithic module, especially when the application involves a large number of elements. This dissertation studies an alternative phase beam-scanning technique using arrays of coupled oscillators (COA), which avoids the use of phase shifters. This technique of COA may reduce the complexity of phase control circuits and provide for a phased array of lower volume and weight. Consequently, it simplifies the architecture of the T/R module and reduces the overall cost. In this work, dynamic equations of the nonlinear COA with arbitrary coupling networks are derived using both time and frequency domain methods. From the dynamic analysis, it is shown that the phase distribution along the array, and hence the beam scanning angle of the array, can be controlled by free running frequencies of the coupled oscillators. The stability and nonlinear behaviors of synchronized coupled oscillators are studied via the nonlinear control theory and applied to radar beam scanning arrays. Analysis indicates that a stable, unique equilibrium point exists when choosing a specific set of free running frequencies, and it is associated with the desired phase shift but within a given range. By means of previous dynamic analysis, effects of amplitude dynamics are studied for COAs with uniform, triangular and Chebyshev amplitude distributions. The array with different coupling strengths, nonlinear parameters, and synchronization frequencies are considered. Results demonstrate that beam shapes and SLLs can be controlled for the coupled oscillator array using strong coupling. The influence of the random, free-running frequency distribution of the phase error in COAs, which causes the phase shift error and hence the error of main beam scanning angle (EMBSA), is also investigated through a Monte Carlo analysis. It is found that strongly COAs are more robust than weakly COAs under the same level of randomness in free running frequencies. Furthermore, when random deviations become larger, the robustness of strongly COA is especially obvious.