During the cure of polymer matrix composites, induced stresses develop due to shrinkage of the matrix material. Consequences of this can lead to shifting of the reinforcement, adversely affecting final properties of the material, or the induced stresses can alter the final geometry of the part. With the use of a new closed loop feedback program developed, residual stresses built up during cure were minimized. Experiments were performed using the EPON 828 resin with two types of reinforcement, carbon and glass fiber. The residual stress built up during the optimized cure cycle was compared with that produced during the lPanufacturer recommended 2-step cure cycle and isothermal cure cycles. Results for both fibers show a large reduction in stresses endured during cure for the optimized cure compared to typical stresses seen under isothermal and standard cure cycles. Static and dynamic testing were done on specimens and showed that the modulus and the glass transition temperatures of cured specimens were not significantly affected by the optimized cure cycles. Results also show that optimized cure cycles were of shorter duration compared to the standard cure cycles.

Thesis (M.App.Sc.) - Carleton University, 2009. / Includes bibliographical references (p.148-151). Also available in electronic format on the Internet.

<p>In recent years there has been a significant increase in interest in smaller satellites as lower cost alternatives to traditional satellites, particularly with the rise in popularity of the CubeSat. Due to stringent mass, size, and often budget constraints, these small satellites rely on making the most of inexpensive hardware components and sensors, such as coarse sun sensors (CSS) and magnetometers. More expensive high-accuracy sun sensors often combine multiple measurements, and use specialized electronics, to deterministically solve for the direction of the Sun. Alternatively, cosine-type CSS output a voltage relative to the input light and are attractive due to their very low cost, simplicity to manufacture, small size, and minimal power consumption. This research investigates using coarse sun sensors for performing robust attitude estimation in order to point a spacecraft at the Sun after deployment from a launch vehicle, or following a system fault. As an alternative to using a large number of sensors, this thesis explores sun-direction estimation techniques with low computational costs that function well with underdetermined sets of CSS. Single-point estimators are coupled with simultaneous nonlinear control to achieve sun-pointing within a small percentage of a single orbit despite the partially underdetermined nature of the sensor suite. Leveraging an extensive analysis of the sensor models involved, sequential filtering techniques are shown to be capable of estimating the sun-direction to within a few degrees, with no a priori attitude information and using only CSS, despite the significant noise and biases present in the system. Detailed numerical simulations are used to compare and contrast the performance of the five different estimation techniques, with and without rate gyro measurements, their sensitivity to rate gyro accuracy, and their computation time. One of the key concerns with reducing the number of CSS is sensor degradation and failure. In this thesis, a Modified Rodrigues Parameter based CSS calibration filter suitable for autonomous on-board operation is developed. The sensitivity of this method's accuracy to the available Earth albedo data is evaluated and compared to the required computational effort. The calibration filter is expanded to perform sensor fault detection, and promising results are shown for reduced resolution albedo models. All of the methods discussed provide alternative attitude, determination, and control system algorithms for small satellite missions looking to use inexpensive, small sensors due to size, power, or budget limitations.

<p> Since its invention at the University of Stuttgart, Germany in the mid-1960, scientists have been trying to understand and explain the mechanism of the plasma interaction inside the magnetoplasmadynamics (MPD) thruster. Because this thruster creates a larger level of efficiency than combustion thrusters, this MPD thruster is the primary cadidate thruster for a long duration (planetary) spacecraft. However, the complexity of this thruster make it difficult to fully understand the plasma interaction in an MPD thruster while operating the device. That is, there is a great deal of physics involved: the fluid dynamics, the electromagnetics, the plasma dynamics, and the thermodynamics. All of these physics must be included when an MPD thruster operates. </p><p> In recent years, a computer simulation helped scientists to simulate the experiments by programing the physics theories and comparing the simulation results with the experimental data. Many MPD thruster simulations have been conducted: E. Niewood et al.[5], C. K. J. Hulston et al.[6], K. D. Goodfellow[3], J Rossignol et al.[7]. All of these MPD computer simulations helped the scientists to see how quickly the system responds to the new design parameters. </p><p> For this work, a 1D MPD thruster simulation was developed to find the voltage drop between the cathode and the plasma regions. Also, the properties such as thermal conductivity, electrical conductivity and heat capacity are temperature and pressure dependent. These two conductivity and heat capacity are usually definded as constant values in many other models. However, this 1D and 2D cylindrical symmetry MPD thruster simulations include both temperature and pressure effects to the electrical, thermal conductivities and heat capacity values interpolated from W. F. Ahtye [4]. Eventhough, the pressure effect is also significant; however, in this study the pressure at 66 Pa was set as a baseline. </p><p> The 1D MPD thruster simulation includes the sheath region, which is the interface between the plasma and the cathode regions. This sheath model [3] has been fully combined in the 1D simulation. That is, the sheath model calculates the heat flux and the sheath voltage by giving the temperature and the current density. This sheath model must be included in the simulation, as the sheath region is treated differently from the main plasma region. </p><p> For our 2D cylindrical symmetry simulation, the dimensions of the cathode, the anode, the total current, the pressure, the type of gases, the work function can be changed in the input process as needed for particular interested. Also, the sheath model is still included and fully integrated in this 2D cylindrical symmetry simulation at the cathode surface grids. In addition, the focus of the 2D cylindrical symmetry simulation is to connect the properties on the plasma and the cathode regions on the cathode surface until the MPD thruster reach steady state and estimate the plasma arc attachement edge, electroarc edge, on the cathode surface. Finally, we can understand more about the behavior of an MPD thruster under many different conditions of 2D cylindrical symmetry MPD thruster simulations.</p>

<p> Crack propagation in a solid rocket motor environment is difficult to measure directly. This experimental and analytical study evaluated the viability of real-time radiography for detecting bore regression and propellant crack propagation speed. The scope included the quantitative interpretation of crack tip velocity from simulated radiographic images of a burning, center-perforated grain and actual real-time radiographs taken on a rapid-prototyped model that dynamically produced the surface movements modeled in the simulation. The simplified motor simulation portrayed a bore crack that propagated radially at a speed that was 10 times the burning rate of the bore. Comparing the experimental image interpretation with the calibrated surface inputs, measurement accuracies were quantified. The average measurements of the bore radius were within 3% of the calibrated values with a maximum error of 7%. The crack tip speed could be characterized with image processing algorithms, but not with the dynamic calibration data. The laboratory data revealed that noise in the transmitted X-Ray intensity makes sensing the crack tip propagation using changes in the centerline transmitted intensity level impractical using the algorithms employed.</p>

The generation and development of large 2D vortical disturbances (coherent structures) in forced transitional and turbulent wall jets is investigated using several numerical techniques. For the early and late transition stages, 2D Numerical Simulation (2D-NS) and Direct Numerical Simulation (DNS) are employed, while for the forced turbulent flow Unsteady Reynolds-Averaged Navier-Stokes (URANS) calculations are used including a new, simplified approach called "Stability" RANS (SRANS) which substantially reduces the computational effort when compared to URANS. As base flows for the investigations, three prototypical wall jets are considered: Low and high Reynolds number laminar wall jets, represented by the Glauert similarity solution, and a turbulent wall jet (Rej = 10,000), modeled using a nearly self-preserving RANS solution starting at a virtual nozzle. The investigations of 2D vortical disturbances in both the transitional and the turbulent wall jet follow the 2D stages of shear flow transition, beginning with receptivity to harmonic forcing, followed by linear and nonlinear disturbance development, and 2D secondary instability. It is shown that the disturbance development in the turbulent flow parallels the one in the transitional flow in many respects. In particular, a 2D subharmonic resonance is found in both flows leading to a subharmonic resonance cascade with repeated vortex merging. Competing 3D fundamental and subharmonic resonances in the transitional wall jet are studied using a linearized Navier-Stokes code and 3D DNS. These 3D secondary instabilities weaken or diminish the 2D disturbances and lead to turbulent breakdown. Yet, for large amplitude forcing, the 3D resonances are surpassed by the 2D subharmonic resonance which leads to vortex merging upstream of the breakdown. With a 3D DNS of bypass transition, where a high Reynolds number laminar wall jet is tripped with large amplitude 3D forcing, it is demonstrated that 2D vortical structures persist in the presence of 3D turbulent fluctuations. In this simulation, 2D vortical structures emerge during transition and undergo repeated merging in the turbulent flow downstream.

The effect of high amplitude forcing on laminar and turbulent wall jets over a heated flat plate is analyzed. Highly accurate Direct Numerical Simulations (DNS) are used in the laminar case to investigate the dominant transport mechanisms. When forcing is applied, the skin friction is reduced markedly and the wall heat transfer is increased, in contrast to the prediction of the Reynolds analogy, which states proportionality between both quantities. Detailed examination of the unsteady flow field showed that the concepts of eddy viscosity and eddy thermal diffusivity can be applied to analyze unsteady laminar flows and to explain the effect of highly unsteady phenomena. For the investigation of the turbulent wall jet, a new Flow Simulation Methodology (FSM) is employed in the limit of unsteady BANS (Reynolds averaged Navier-Stokes) simulations. With this novel approach, the simulation of large, coherent structures in the turbulent flow field very closely parallels the laminar simulations. Following the idea of Large Eddy Simulation (LES), the large coherent motion is computed directly, while the effect of the small scale, random motion is modelled. In FSM, a state-of-the-art two-equation turbulence model is used. Forcing the turbulent wall jet results in a reduction of the skin friction and an increase in wall heat transfer. The mechanisms responsible for these mean flow changes show a remarkable similarity to the mechanisms found in the laminar case. This is confirmed by close examination of the large coherent motion and its effect on the turbulent mean flow. Using this approach, several questions regarding the character of the turbulent wall jet could be answered.

This research aims at modifying, improving and calibrating the Stress Wave Factor Nondestructive Evaluation (SWF NDE) technique and applying it to a fiber epoxy composite material and other composite structures. In order to access the composite's integrity the Energy of SWF within a selected Sub Frequency Range (SFR) instead of the whole measured frequency range as of conventional SWF is used. This technique, introduced and examined herein and is termed the Sub Frequency Range Stress Wave Factors (SFR-SWF) and is tailored to improve the conventional SWF technique with respect to sensibility and accuracy. A series of controlled damage tests were performed, and relevant acousto-ultrasonic observations were conducted. The overall property of the composites subjected to hygrothermal degradation, the localized defects such as the surface crack and the historical damage were assessed with conventional SWF and SFR-SWF. The two methods are compared in detail. The hygrothermal degradation and surface crack experiments were also simulated using the finite element method. Dynamic numerical analysis was conducted to simulate the wave propagation process, both in time domain and frequency domain using the commercial finite element code ABAQUS. The numerical results were also evaluated via both SWF and SFR-SWF, and were compared with the results of experiments. Thus, the potential of SFR-SWF was evaluated. A general conclusion from this research is that the SFR-SWF has the better capability than that of the conventional SWF in assessing the composite's overall condition, localized defects and historical damage. Since there are still open questions regarding the physical understanding of the SWF and SFR-SWF, the finite element analysis provides confirmation for certain observed behaviors of the Acousto-Ultrasonic and SFR-SWF technique.

<p> The most important and crucial design feature while designing an Aircraft Electric Power Distribution System (EPDS) is reliability. In EPDS, the distribution of power is from top level generators to bottom level loads through various sensors, actuators and rectifiers with the help of AC &amp; DC buses and control switches. As the demands of the consumer is never ending and the safety is utmost important, there is an increase in loads and as a result increase in power management. Therefore, the design of an EPDS should be optimized to have maximum efficiency. This thesis discusses an integrated tool that is based on a Need Based Design method and Fault Tree Analysis (FTA) to achieve the optimum design of an EPDS to provide maximum reliability in terms of continuous connectivity, power management and minimum cost. If an EPDS is formulated as an optimization problem then it can be solved with the help of connectivity, cost and power constraints by using a linear solver to get the desired output of maximum reliability at minimum cost. Furthermore, the thesis also discusses the viability and implementation of the resulted topology on typical large aircraft specifications.</p>

<p> This study presents and investigates a novel hybrid fuel grain that reacts pyrophorically with gaseous oxidizer to achieve restart of a hybrid rocket motor propulsion system while reducing cost and handling concerns. This reactive fuel grain (RFG) relies on the pyrophoric nature of finely divided metal particles dispersed in a solid dicyclopentadiene (DCPD) binder, which has been shown to encapsulate air-sensitive additives until they are exposed to combustion gases. An RFG is thus effectively inert in open air in the absence of an ignition source, though the particles encapsulated within remain pyrophoric. In practice, this means that an RFG that is ignited in the vacuum of space and then extinguished will expose unoxidized pyrophoric particles, which can be used to generate sufficient heat to relight the propellant when oxidizer is flowed.</p><p> The experiments outlined in this work aim to develop a suitable pyrophoric material for use in an RFG, demonstrate pyrophoric relight, and characterize performance under conditions relevant to a hybrid rocket thruster. Magnesium, lithium, calcium, and an alloy of titanium, chromium, and manganese (TiCrMn) were investigated to determine suitability of pure metals as RFG additives. Additionally, aluminum hydride (AlH₃), lithium aluminum hydride (LiAlH₄), lithium borohydride (LiBH₄), and magnesium hydride (MgH₂) were investigated to determine suitability of metals hydrides as RFG additives or as precursors for pure-metal RFG additives. Pyrophoric metals have been previously investigated as additives for increasing the regression rate of hybrid fuels, but to the author&rsquo;s knowledge, these materials have not been specifically investigated for their ability to ignite a propellant pyrophorically.</p><p> Commercial research-grade metals were obtained as coarse powders, then ball-milled to attempt to reduce particle size below a critical diameter needed for pyrophoricity. Magnesium hydride was ball-milled and then cycled in a hydride cycling apparatus to attempt to fracture the particles through hydrogen sorption and thermal stresses. These powders were then tested for pyrophoricity with atmospheric and pure concentrations of oxygen. The TiCrMn powder was chosen as the material for evaluation of propellant performance, and was mixed with DCPD in various weight ratios to determine the required additive loading needed for pyrophoricity of the bulk propellant. Weight percentages of 10, 20, 30, and 50 wt.% TiCrMn were used to evaluate relight capability and propellant performance, and weight loadings of 50, 70, and 90 wt.% TiCrMn were used to evaluate approximate maximum loading possible without rendering the propellant structurally unsound. Propellant tests were conducted in an opposed flow burner apparatus for sub-scale regression rate and relight experiments, and an optically accessible cylindrical combustion chamber (OCC) that allows high speed cameras to record the regressing propellant surface during combustion. Gaseous oxygen (GOX) was used as an oxidizer for all tests due to its ready availability and common use as a hybrid rocket oxidizer. Opposed flow burner experiments are an inexpensive means of rapidly testing various propellant formulations at different conditions, whereas OCC tests are useful for obtaining realistic data on how an RFG would likely operate as part of a propulsion system.</p><p> Relight in the opposed flow burner was attempted by cycling oxygen and nitrogen flows with carefully timed solenoid valves to initiate and extinguish combustion, and to control the slow diffusion of oxygen to the surface of the propellant, which would render the TiCrMn non-pyrophoric. The opposed flow burner experiments did not conclusively demonstrate the pyrophoric relight capability of the RFG propellant due in part to the persistence of hot spots between oxygen and purge nitrogen cycles, as determined by high-speed imaging in the near infrared range. An opposed flow burner apparatus was then constructed within a vacuum chamber assembly thus preventing atmospheric oxygen from diffusing to the propellant surface, but these tests did not demonstrate pyrophoric relight. Future work is proposed to evaluate the effect of pyrophoric particle size in order to determine the role ignition delay of each particle has in the relight capability of RFGs.</p><p> OCC experiments were conducted at a low and high GOX mass flux of approximately 150 and 300 kg/s/m², respectively, at a nominal chamber pressure of 150 psia. Four strand compositions were used: pure DCPD, 30 wt.% pyrophoric TiCrMn powder with average particle diameters of approximately 1-10 microns, 30 wt.% oxidized TiCrMn powder with average particle diameters of approximately 1-10 microns, and 30 wt.% TiCrMn powder with average particle diameters of approximately 1-4 mm. Regression rate was measure by weight loss, average web thickness change at three axial locations on the strand, and through time-resolved tracking of the regressing propellant surface via high speed video. While visual observations suggest that the addition of TiCrMn significantly increases regression rate, initial data do not show a significant trend. Additionally, it is observed that the oxidized TiCrMn strands regress at the same rate as those loaded with pyrophoric TiCrMn, suggesting that erosive burning and heat addition of the added metal may be the cause of the observed increase in regression rate. The data are too sparse to make conclusions about the effect of particle size on regression rate, so further tests are recommended to develop a significant data set for the effect of pyrophoricity and particle size on regression rate. The test article was damaged at the end of the regression rate experimental campaign, which precluded the collection of relight data that was planned for strands loaded with 50 wt.% TiCrMn particles with an average diameter of approximately 1-4 mm. Though further tests are needed to demonstrate pyrophoric relight of an RFG, the current work establishes a baseline for RFG performance and suggests that pyrophoric relight is possible by tailoring the particle size of the pyrophoric metal additive to control heat release and ignition delay.</p>