621 |
Many-Light Real-Time Global Illumination using Sparse Voxel OctreeSun, Che 18 December 2015 (has links)
"Global illumination (GI) rendering simulates the propagation of light through a 3D volume and its interaction with surfaces, dramatically increasing the fidelity of computer generated images. While off-line GI algorithms such as ray tracing and radiosity can generate physically accurate images, their rendering speeds are too slow for real-time applications. The many-light method is one of many novel emerging real-time global illumination algorithms. However, it requires many shadow maps to be generated for Virtual Point Light (VPL) visibility tests, which reduces its efficiency. Prior solutions restrict either the number or accuracy of shadow map updates, which may lower the accuracy of indirect illumination or prevent the rendering of fully dynamic scenes. In this thesis, we propose a hybrid real-time GI algorithm that utilizes an efficient Sparse Voxel Octree (SVO) ray marching algorithm for visibility tests instead of the shadow map generation step of the many-light algorithm. Our technique achieves high rendering fidelity at about 50 FPS, is highly scalable and can support thousands of VPLs generated on the fly. A survey of current real-time GI techniques as well as details of our implementation using OpenGL and Shader Model 5 are also presented."
|
622 |
Incremental Learning and Online-Style SVM for Traffic Light ClassificationLiu, Wen 28 January 2016 (has links)
Training a large dataset has become a serious issue for researchers because it requires large memories and can take a long time for computing. People are trying to process large scale dataset not only by changing programming model, such as using MapReduce and Hadoop, but also by designing new algorithms that can retain performance with less complexity and runtime. In this thesis, we present implementations of incremental learning and online learning methods to classify a large traffic light dataset for traffic light recognition. The introduction part includes the concepts and related works of incremental learning and online learning. The main algorithm is a modification of IMORL incremental learning model to enhance its performance over the learning process of our application. Then we briefly discuss how the traffic light recognition algorithm works and the problem we encounter during training. Rather than focusing on incremental learning, which uses batch to batch data during training procedure, we introduce Pegasos, an online style primal gradient-based support vector machine method. The performance of Pegasos for classification is extraordinary and the number of instances it uses for training is relatively small. Therefore, Pegasos is the recommended solution to the large dataset training problem.
|
623 |
Pathogenesis of light chain-induced dysfunction in cardiac amyloidosisSnyder, Christina AnnaMarie 22 January 2016 (has links)
Although a rare disease, light chain (LC) amyloidosis (AL) is the most common systemic amyloidosis in developed countries. It is caused by an overproduction of immunoglobulin LC proteins in bone marrow plasma cells. In AL amyloidosis, LCs that are prone to misfolding and insolubility will aggregate, form fibrils, and deposit themselves in various tissues, thereby causing organ dysfunction. The most fatal manifestation of AL amyloidosis is associated with cardiac involvement, defined by the presence of extracellular AL amyloid deposits within the heart. Cardiac amyloid infiltration typically leads to diastolic dysfunction followed by heart failure and has a median survival of approximately 6 months from the time of diagnosis if untreated.
Clinical observation suggests that a reduction in circulating LCs results in an improvement in heart failure symptoms despite minimal changes in amyloid deposition. This has led to the concept that LCs themselves are cytotoxic to cardiomyocytes. Recent studies indicate that AL LCs induce oxidative stress, cellular dysfunction, and apoptosis (programmed cell death) in cardiomyocytes via a p38α mitogen-activated protein kinase (MAPK) mechanism. They may therefore be a target for amyloidosis therapy. By understanding how LCs cause cardiac dysfunction, we can target this process with therapies and utilize downstream measures of LC activity as diagnostic and prognostic tools.
The objective of this study was to determine the role of autophagy in AL amyloidosis. Autophagy is the intracellular process of degrading aging or dysfunctional cellular components. Autophagy can be beneficial by preventing proteotoxicity and providing nutrients, amino acids, and other necessities during times of cellular stress. On the other hand, increased autophagy, like apoptosis, may mediate cellular death depending on the type of stimulus and its duration. Autophagy is induced by a variety of stimuli, including oxidative stress. AL has been demonstrated to increase reactive oxygen species (ROS), and it is unknown if autophagy mediates cardiomyocyte dysfunction in AL cardiac amyloidosis. We thus sought to determine if it is a factor in amyloid cardiotoxicity. We explored the ERK1/2, p38, and JNK MAPK pathways in particular, since MAPK signaling cascades regulate several transcription factors involved in the cell cycle and p38α has been implicated in ROS-induced cardiac AL amyloidosis.
Adult rat ventricular myocytes (ARVM) were harvested from healthy adult male rats and exposed to a variety of experimental conditions in vitro. ARVM were treated with vehicle control, human LC obtained from a patient without cardiac involvement, a positive control (aldosterone), and human AL light chains obtained from a patient with AL cardiac amyloidosis in the presence or absence of UO126, SB203580, or SP600125 (specific inhibitors of ERK1/2, p38, and JNK, respectively). The resulting protein expression levels of autophagy indicators LC3II and ATG4B in cardiomyocytes were analyzed by Western blotting. The ratio of phosphorylated to total ERK1/2 protein expression was also explored.
We found that AL light chains did not contribute to autophagy via the ERK1/2, p38, or JNK pathways. In contrast to our previous unpublished findings, the protein levels of autophagy indicators in AL-treated ARVM did not differ from vehicle control levels, suggesting that AL did not activate autophagy. However non-cardiomyopathic light chains (LC) did increase LC3II expression in ARVM, despite their human source exhibiting no clinical indications of cardiac involvement. This implies that autophagy induced by non-cardiomyopathic LCs may be beneficial and protect against the development of the cardiotoxicity seen in AL cardiac amyloidosis.
Further studies are necessary to understand the effect of autophagy in the heart and its role in cardiac amyloidosis. Continuing to explore the underlying mechanisms of AL light chain toxicity will contribute to the development of diagnostic, prognostic, and treatment strategies for AL amyloidosis.
|
624 |
Halide perovskites for photovoltaics and light-emitting diodesZhao, Baodan January 2019 (has links)
Halide perovskite solar cells, with rapid efficiency improvements from ~10% to ~23% in 6 years, have attracted significant attention due to their remarkable performance, low processing cost and their potential to become a strong alternative candidate to silicon solar cells. Significant development has also been achieved in halide perovskite-based LEDs with EQE improved from below 1% to ~20% in less than 4 years. This remarkable progress can mainly be attributed to the optimisation of halide perovskite properties. This dissertation focuses on the correlation between optical and electrical properties of halide perovskites and their remarkable performance. Bandgap tunabilities of halide perovskites in blue to green regions through mixing Br-and Cl-and in near infra-red region by substituting Pb2+ with Sn2+ are demonstrated. The absorption and PL spectra are consistent with each other supporting the bandgap tunability. Corresponding EL spectra, which are consistent with their PL spectra, are also demonstrated for blue to green regions. Terahertz measurements coupled with PLQE and transient PL results reveal that the high carrier mobilities are the main reason behind the high efficiency of tin-rich samples. A novel perovskite-polymer-bulk heterostructure is introduced and studied comprehensively. Correlations between their optoelectronic properties and remarkable performance on timescales ranging from femtosecond to microsecond are presented. Transient optical spectroscopy reveals the energy transfer from 2D regions to 3D regions happens in 1 ps. The 20% EQE of the LEDs based in this structure is consistent with conventional thin-film optical models giving internal quantum efficiency of ~100%. This in agreement with near-unity PLQE value of the pristine emissive layer material and the dominant bimolecular recombination process observed in nanosecond-scale transient PL measurements. Two typical interfacial engineering methods to improve the quality of halide perovskite and device performance are then presented. Optimised NiOx is adopted to improve the anode interface. From transient photovoltaic measurements, we find the charge collection ability of NiOx is superior to that of PEDOT:PSS. This is also the main reason behind their better photovoltaic device performance. A unique anti-solvent treatment with additive modifies both the bulk and surfaces of halide perovskites and improves the device performance significantly. Transient PL and PLQE measurements demonstrate that non-radiative recombination pathways are significantly reduced.
|
625 |
Light in scattering media : active control and the exploration of intensity correlationsPaniagua Diaz, Alba Maria January 2018 (has links)
When light encounters scattering materials such as biological tissue, white paint or clouds, it gets randomly scattered in all directions, which traditionally has been seen as a barrier for imaging techniques (reducing their resolution) or sensing, due to the reduction of the penetration depth of light. However, in recent years it has been shown that scattering might not necessarily be an impediment, and that the knowledge of the properties of multiple scattering can be indeed useful for imaging, sensing and other applications. In the first part of this thesis (Chapters 2 to 5) we study the implications of manipulating the light incident on a multiply scattering material. We experimentally show how by actively controlling the output light of a bad quality laser we manage to not only improve its beam quality, but also in an energy-efficient way, in comparison with traditional methods. In a different experiment presented in this thesis, we show how the active control of the light incident on a scattering material can be useful to improve sensing through scattering media, by means of increasing the transmission and energy deposited inside (Chapter 5). In the final part of the thesis we present the first experimental observation of intensity correla- tions between transmitted and reflected patterns from a scattering material (Chapter 6), exploring how it depends on the parameters of the scattering medium. In the last part of the thesis (Chapter 7) we present a new imaging technique based on the use of the intensity correlations described in the previous chapter, opening new possibilities to non-invasive imaging through highly scattering materials.
|
626 |
Harmonious StormDOU, JIAYUN January 2011 (has links)
The intention with this imaginary storm in the form of a decorative lamp ‘Harmonious Storm’ is meant to hang from the ceiling, sculptural and often glowing with artificial light. Meanwhile it enhances the attractive value in the specific public space. The relation between importance of aesthetics and interior public environment is discussed in this thesis, with a purpose of achieving emotional and social needs that are excitement and happiness. / Program: Konstnärligt masterprogram i mode- och textildesign
|
627 |
Assessment of high-burnup LWR fuel response to reactivity-initiated accidents / Assessment of high-burnup Light Water Reactor fuel response to reactivity-initiated accidentsLiu, Wenfeng, Ph.D. Massachusetts Institute of Technology January 2007 (has links)
Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Nuclear Science and Engineering, 2007. / Includes bibliographical references (p. 263-273). / The economic advantages of longer fuel cycle, improved fuel utilization and reduced spent fuel storage have been driving the nuclear industry to pursue higher discharge burnup of Light Water Reactor (LWR) fuel. A design basis accident, the Reactivity Initiated Accident (RIA), became a concern for further increase of burnup as simulated RIA tests revealed lower enthalpy threshold for fuel failure associated with fuel dispersal, which may compromise the core coolability and/or cause radiological release should this happened in LWRs. Valuable information on the behavior of high burnup fuel during RIA are provided by the simulation tests. However atypical design and operating conditions in simulated tests limited the application of experimental data directly to evaluate the failure potential of LWR fuels. To better interpret the experimental results and improve the capability of the fuel performance codes to predict high burnup fuel behavior, this thesis developed mechanistic models of high burnup fuel during an RIA and implemented models in a transient fuel performance code FRAPTRAN 1.3. Fission gas release (FGR) and swelling were systematically modeled to quantify gaseous loading effects. The grain boundary fission gas inventory is simulated prior to the transient using a diffusion model in FRAPCON 3.3 code. The restructuring of high burnup fuel in rim region is described in terms of porosity, pore size distribution, fission gas concentration, and pore overpressure. The model assumes the fragmentation of fuel upon the separation of grain boundary or when a threshold temperature is exceeded in the rim region. The fission gas in fragmented fuel is assumed to release instantaneously to the free volume when the fuel expansion and swelling creates sufficient pellet-clad gap. / (cont.) The relaxation of rim pore at rapid temperature increase and the thermal expansion of fission gas in fragmented fuel are considered as additional loads on the cladding besides the contact force due to fuel thermal expansion. An analytical approximation is made to calculate the clad radial displacement subjected to fission gas expansion accounting for the constraint of the cladding on the fission gas which would otherwise be neglected in a rigid pellet model FRACAS-I in the FRAPTRAN code. In comparison to the measured FGR from CABRI, NSRR and BIGR test facilities, this mechanistic model can reasonably predict fission gas release fraction for most of the test cases covering a burnup range of 26-64 MWd/kgU and enthalpy deposit of 37-200 cal/g. It reveals the effects of burnup and enthalpy deposit on the fission gas release: burnup is an important parameter affecting fission gas inventory and fuel micro-structure evolution during base irradiation; enthalpy deposit is directly connected to the availability of fission gas release via the grain boundary separation by the intergranular bubble over-pressurization. Analysis of the fission gas radial profile is made with the aid of the neutronic code MCODE to validate the fission gas release from the rim of UO2 fuel. The analysis indicates fission gas release is partly from the rim region and the majority of fission gas release is from grain boundaries for burnup up to 50 MWd/kgU. Fission gas induced hoop strain is predicted to be less than 0.3% in the early phase of RIA with peak fuel enthalpy less than 145 cal/g. Given the fact that the concerned failure mode is the PCMI failure at low energy deposit, the pellet thermal expansion is still considered as effective in analyzing the PCMI failure. However at high level of enthalpy deposit, when clad yield strength is decreased at escalated temperature due to film boiling, the fission gas either released into the plenum or retained in the fuel pellet might strain more the cladding. / (cont.) This is observed in the large deformation of the cladding in some test cases in NSRR and BIGR due to pressure load. A new set of heat transfer correlations were selected and implemented in the FRAPTRAN code to model the cladding-coolant heat transfer of high burnup fuel at room temperature and atmospheric pressure condition. This new set of correlations addressed the effects of subcooling and oxiation on the heat transfer characteristics at pool boiling conditions. They reflect the increase of rewetting temperature and increase of Critical Heat Flux (CHF) due to subcooling. They account for oxidation effects on the transition and film boiling regime and heat conduction through thick oxide as the oxidation is considered as a prominent feature of surface condition change of high burnup fuel. In addition to high burnup fuels tested in NSRR, several fresh fuel tests with different degree of subcooling and a few separate-effects RIA tests are also included to validate the applicabilty of this set of correlations. For fuel enthalpy up to 190 cal/g and oxidation up to 25 micron, the predicted peak cladding temperature (PCT) and duration of DNB achieves generally good agreement with the experimental data. The analysis of high burnup fuel heat transfer reveals that the surface oxidation could cause an early rewetting of high burnup fuel or suppression of DNB. Surface oxidation can delay the heat conducting to the surface while keeping the surface heat transfer in the effective nucleate boiling regime. It also raises the miniumum stable film boiling temperature by lowering the interface temperature during liquid-solid contact resulting from vapor breaking down. Also modeled was Pellet-Cladding Mechanical Interaction (PCMI) failure of irradiated and hydrided cladding. The hydride rim accumulated at outer clad is assumed to cause the crack initiation. The fracture toughness of irradiated and hydrided cladding is obtained by fitting experimental data at different temperature range. / (cont.) The model sets forth a simple criterion for failure associated with crack growth based on the J integral approach. The simplification is that for the thin clad, failure is assumed to occur at the onset of crack tip growth. In comparison to CABRI and NSRR test results and other failure models, the model shows a good capability to separate the failure cases from non-failure cases. These models have been applied to LWR conditions to determine the failure potential of high burnup fuel. It shows that, at high burnup (and therefore high hydride levels in the cladding), the failure enthalpy is smaller than at low burnup. The pulse width is an important parameter in the burnup up to 50 MWd/kg, but starts to become less important for higher burnup with highly corroded cladding. / by Wenfeng Liu. / Ph.D.
|
628 |
Fine structure of hydrogen using quantum beat spectroscopySchulz, Peter Anfaenger January 1975 (has links)
Thesis. 1975. B.S.--Massachusetts Institute of Technology. Dept. of Physics. / Includes bibliographical references. / by Peter A. Schulz. / B.S.
|
629 |
Measurements of thermally activated relaxation times in amorphous poly(methyl methacrylate) using photon correlation spectroscopyGangwere, George Henry January 2011 (has links)
Typescript (photocopy). / Digitized by Kansas Correctional Industries
|
630 |
An investigation of some properties of supercooled fluids using photon correlation spectroscopyHalfpap, Bradford Lee January 2011 (has links)
Typescript (photocopy). / Digitized by Kansas Correctional Industries
|
Page generated in 0.0573 seconds