Hurricanes and earthquakes have caused extensive property damage to wood-frame residential construction in the past two decades in the United States. In order to improve residential building performance and mitigate losses from hurricane and earthquake hazards, there is an urgent need for better understanding of building performance and improvements in design and evaluation tools. In this study, a fragility analysis methodology is developed for assessing the response of light-frame wood construction exposed to extreme hurricane winds and earthquakes. The fragility is a conditional limit state probability, presented as a function of the 3-second gust wind speed (hurricanes) or spectral acceleration at the fundamental period of the building (earthquakes), leading to a relation between damage state probability and the hazard stipulated in ASCE Standard 7. A fully coupled probabilistic framework is proposed to assess reliability of the residential construction through convolution of the structural fragility model with hazard models. Finally, a comparative risk assessment addresses the similarities and differences in competing hurricane and earthquake hazards. The tools above can be used to evaluate new and existing building products, model the uncertainties that are inherent to the prediction of building performance, and manage the risk that is consequent to these uncertainties economically

In a typical printed circuit board assembly, over 70 percent of the electronic components are passives such as resistors, inductors, and capacitors, and these passives could take up to 50 percent of the entire printed circuit board area. By embedding the passive components within the substrate instead of being mounted on the surface, the embedded passives could reduce the system real estate, eliminate the need for surface-mounted discrete components, eliminate lead based interconnects, enhance electrical performance and reliability, and potentially reduce the overall cost. Even with these advantages, embedded passive technology, especially for organic substrates, is at an early stage of development, and thus a comprehensive experimental and theoretical modeling study is needed to understand the fabrication and reliability of embedded passives before they can be widely used. This thesis aims to fabricate embedded passives in a multilayered organic substrate, perform extensive electrical and mechanical reliability tests, and develop physics-based models to predict the thermo-mechanical reliability of embedded capacitors. Embedded capacitors and resistors with different geometric shapes, planar dimensions, and thus different electrical characteristics have been fabricated on two different test vehicles. Capacitors are made with polymer/ceramic nanocomposite materials and have a capacitance in the range of 50 pF to 1.5 nF. Resistors are carbon ink based Polymer Thick Film (PTF) and NiCrAlSi and have a resistance in the range of 25 to 400 k. High frequency measurements have been done using Vector Network Analyzer (VNA) with 2 port signal-ground (S-G) probes. Accelerated thermal cycling (-55 to 125oC) and constant temperature and humidity tests (85oC/85RH) based on JEDEC and MIL standards have been performed. Furthermore, physics-based numerical models have been developed and validated using the experimental data. By focusing on the design and fabrication as well as the experimental and theoretical reliability assessments, this thesis aims to contribute to the overall development of embedded passive technology for Digital and Radio Frequency (RF) applications.

Abstract The goal of this research is to advance the state of the art in bulk power system reliability assessment. Bulk power system reliability assessment is an important procedure at both power system planning and operating stages to assure reliable and acceptable electricity service to customers. With the increase in the complexity of modern power systems and advances in the power industry toward restructuring, the system models and algorithms of traditional reliability assessment techniques are becoming obsolete as they suffer from nonrealistic system models and slow convergence (even non-convergence) when multi-level contingencies are considered and the system is overstressed. To allow more rigor in system modeling and higher computational efficiency in reliability evaluation procedures, this research proposes an analytically-based security-constrained adequacy evaluation (SCAE) methodology that performs bulk power system reliability assessment. The SCAE methodology adopts a single-phase quadratized power flow (SPQPF) model as a basis and encompasses three main steps: (1) critical contingency selection, (2) effects analysis, and (3) reliability index computations. In the critical contingency selection, an improved contingency selection method is developed using a wind-chime contingency enumeration scheme and a performance index approach based on the system state linearization technique, which can rank critical contingencies with high accuracy and efficiency. In the effects analysis for selected critical contingencies, a non-divergent optimal quadratized power flow (NDOQPF) algorithm is developed to (1) incorporate major system operating practices, security constraints, and remedial actions in a constrained optimization problem and (2) guarantee convergence and provide a solution under all conditions. This algorithm is also capable of efficiently solving the ISO/RTO operational mode in deregulated power systems. Based on the results of the effects analysis, reliability indices that provide a quantitative indication of the system reliability level are computed. In addition, this research extends the proposed SCAE framework to include the effects of protection system hidden failures on bulk power system reliability. The overall SCAE methodology is implemented and applied to IEEE reliability test systems, and evaluation results demonstrate the expected features of proposed advanced techniques. Finally, the contributions of this research are summarized and recommendations for future research are proposed.

Two research topics in next-generation optical networks with wavelength-division multiplexing (WDM) technologies were investigated: (1) scalability of network management and control, and (2) resilience/reliability of networks upon faults and attacks. In scalable network management, the scalability of management information for inter-domain light-path assessment was studied. The light-path assessment was formulated as a decision problem based on decision theory and probabilistic graphical models. It was found that partial information available can provide the desired performance, i.e., a small percentage of erroneous decisions can be traded off to achieve a large saving in the amount of management information. In network resilience under malicious attacks, the resilience of all-optical networks under in-band crosstalk attacks was investigated with probabilistic graphical models. Graphical models provide an explicit view of the spatial dependencies in attack propagation, as well as computationally efficient approaches, e.g., sum-product algorithm, for studying network resilience. With the proposed cross-layer model of attack propagation, key factors that affect the resilience of the network from the physical layer and the network layer were identified. In addition, analytical results on network resilience were obtained for typical topologies including ring, star, and mesh-torus networks. In network performance upon failures, traffic-based network reliability was systematically studied. First a uniform deterministic traffic at the network layer was adopted to analyze the impacts of network topology, failure dependency, and failure protection on network reliability. Then a random network layer traffic model with Poisson arrivals was applied to further investigate the effect of network layer traffic distributions on network reliability. Finally, asymptotic results of network reliability metrics with respect to arrival rate were obtained for typical network topologies under heavy load regime. The main contributions of the thesis include: (1) fundamental understandings of scalable management and resilience of next-generation optical networks with WDM technologies; and (2) the innovative application of probabilistic graphical models, an emerging approach in machine learning, to the research of communication networks.

The infrequent nature of earthquakes in the Central and Eastern United States (CEUS), and the fact that none with intensity comparable to the New Madrid sequence of 1811 12 or the Charleston earthquake of 1886 has occurred in the past century, have caused the earthquake hazard in the region to be ignored until quite recently. The seismic performance of reinforced concrete (RC) frames in the CEUS, which have primarily been designed for gravity load effects, is expected to be deficient when subjected to earthquakes that are judged, in recent seismological research, as being plausible in the New Madrid Seismic Zone (NMSZ). The objective of this study is to develop a set of probability-based tools for efficient uncertainty analysis and seismic vulnerability and risk assessment of such gravity load designed (GLD) RC frames and to use these tools in evaluating the seismic vulnerability of RC frames that are representative of the building inventory in Memphis, TN the largest population center close to the NMSZ. Synthetic earthquake ground motions for the CEUS that are available from two different Mid-America Earthquake (MAE) Center projects were used in the finite element-based simulations for determining the seismic demand on the GLD RC frames by nonlinear time history analysis (NTHA). A beam-column joint model was developed to address the deficiencies in the joints of GLD frames and was incorporated in the finite element structural models. Seismic fragilities were derived for low-, mid-, and high-rise GLD RC frames. Various sources of uncertainty were propagated through the analysis, and their significance for fragility assessment was examined. These fragilities were used to evaluate the vulnerability of the RC frame inventory in Memphis, TN with regard to performance-based design objectives, defined in terms of performance levels associated with reference earthquake hazard levels. This performance appraisal indicated that GLD RC frames do not meet the life safety and collapse prevention performance objectives that are found in recent building codes and guidelines for performance-based earthquake engineering.

High thermal stability and humidity resistance of phosphor-converted white-light-emitting diodes (PC-WLEDs) using Ce:YAG doped glass, instead of conventional Ce:YAG doped silicone, as a phosphor-converted layer is proposed and fabricated. The glass has five times higher glass transition temperature (Tg) of 750¢J compare with silicone of 150¢J, that could exhibited better performance than silicone, including lumen loss, chromaticity shift, transmittance loss, and peak emission intensity undergoing three industry-standard reliability tests at either high (8wt%) or low (2wt%) doping concentrations of Ce:YAG. The proposed glass phosphor possesses host stability as glass and retains desired fluorescence as Ce:YAG. In thermal aging, thermal shock, and damp heat reliability results, the thermal aging has the largest degradation of lumen loss, but the results showed better thermal stability that the glass phosphor with 22~30% lumen loss improvement for 2~8 wt% Ce:YAG doping than silicone phosphor. The damp heat test has the largest degradation of chromaticity shift, but the results showed excellent humidity resistance that the glass phosphor with highest 49~65% chromaticity shift improvement for 2~8 wt% Ce:YAG doping than silicone phosphor. But under thermal shock test, there isn¡¦t a large difference between glass and silicone phosphor. In this study, we demonstrate the feasibility of adapting glass as a phosphor-converted layer in PC-WLED module that can potentially provide higher reliability and better performance for high-power LEDs, particularly in the area where strict reliability is highly required and in the environment where silicone does not stand for long.

Renewal process has been often employed as a mathematical model of the failure and repair cycle of components in power system reliability assessment. This implies that after repair, the component is assumed to be restored to be in as good as new condition in terms of reliability perspective. However, some of the components may enter an aging stage as the system grows older. This thesis describes how aging characteristics of a system may impact the calculation of commonly used quantitative reliability indices such as Loss of Load Expectation (LOLE), Loss of Load Duration (LOLD), and Expected Energy Not Supplied (EENS). To build the history of working and failure states of a system, Stochastic Point Process modeling based on Sequential Monte Carlo simulation is introduced. Power Law Process is modeled as the failure rate function of aging components. Power system reliability analysis can be made at the generation capacity level where transmission constraints may be included. The simulation technique is applied to the Single Area IEEE Reliability Test System (RTS) and the results are evaluated and compared. The results show that reliability indices become increased as the age of the system grows.

In this dissertation we study the problem of allocating computational resources and managing applications in a data center to serve incoming requests in such a way that the energy usage, reliability and quality of service considerations are balanced. The problem is motivated by the growing energy consumption by data centers in the world and their overall inefficiency. This work is focused on designing flexible and robust strategies to manage the resources in such a way that the system is able to meet the service agreements even when the load conditions change. As a first step, we study the control of a Markovian queueing system with controllable number of servers and service rates (M=Mt=kt ) to minimize effort and holding costs. We present structural properties of the optimal policy and suggest an algorithm to find good performance policies even for large cases. Then we present a reactive/proactive approach, and a tailor-made wavelet-based forecasting procedure to determine the resource allocation in a single application setting; the method is tested by simulation with real web traces. The main feature of this method is its robustness and flexibility to meet QoS goals even when the traffic behavior changes. The system was tested by simulating a system with a time service factor QoS agreement. Finally, we consider the multi-application setting and develop a novel load consolidation strategy (of combining applications that are traditionally hosted on different servers) to reduce the server-load variability and the number of booting cycles in order to obtain a better capacity allocation.

Design options have been evaluated for the Modular Helium Reactor (MHR) for higher temperature operation. An alternative configuration for the MHR coolant inlet flow path is developed to reduce the peak vessel temperature (PVT). The coolant inlet path is shifted from the annular path between reactor core barrel and vessel wall through the permanent side reflector (PSR). The number and dimensions of coolant holes are varied to optimize the pressure drop, the inlet velocity, and the percentage of graphite removed from the PSR to create this inlet path. With the removal of ~10% of the graphite from PSR the PVT is reduced from 541 0C to 421 0C. A new design for the graphite block core has been evaluated and optimized to reduce the inlet coolant temperature with the aim of further reduction of PVT. The dimensions and number of fuel rods and coolant holes, and the triangular pitch have been changed and optimized. Different packing fractions for the new core design have been used to conserve the number of fuel particles. Thermal properties for the fuel elements are calculated and incorporated into these analyses. The inlet temperature, mass flow and bypass flow are optimized to limit the peak fuel temperature (PFT) within an acceptable range. Using both of these modifications together, the PVT is reduced to ~350 0C while keeping the outlet temperature at 950 0C and maintaining the PFT within acceptable limits. The vessel and fuel temperatures during low pressure conduction cooldown and high pressure conduction cooldown transients are found to be well below the design limits. The reliability and availability studies for coupled nuclear hydrogen production processes based on the sulfur iodine thermochemical process and high temperature electrolysis process have been accomplished. The fault tree models for both these processes are developed. Using information obtained on system configuration, component failure probability, component repair time and system operating modes and conditions, the system reliability and availability are assessed. Required redundancies are made to improve system reliability and to optimize the plant design for economic performance. The failure rates and outage factors of both processes are found to be well below the maximum acceptable range.

This dissertation focuses on developing inspection schedules to detect non-self- announcing events which can only detected by inspections. Failures of protective sys- tems ,such as electronic equipments, alarms and stand-by systems, incipient failures and the emergence of certain medical diseases are examples of such events. Inspec- tions are performed at pre-determined times to detect whether or not the event has occurred, and necessary actions are taken upon the detection. In this research, my interest is in developing effective inspection schedules to detect non-self-announcing events that balance system downtime and inspection effort. To evaluate the quality of an inspection schedule, I use the availability (for re- placeable) and the detection delay (for non-replaceable systems) as performance mea- sures. When the monetary cost of inspection and the cost of delay are difficult to determine, non-monetary performance measures become more meaningful. In this research, the focus is on maximizing availability or minimizing detection delay given a limited number of inspections or a limited inspection rate. I show that for replace- able and non-replaceable systems, it is possible to construct inspection schedules that perform better than periodic inspection with respect to our performance measures. The occurrence of the event I would like to detect may be influenced by certain individual characteristics. For instance, the risk of developing a certain type of dis- ease might be different for different subgroups within the population. In this case, because of the non-homogeneity in the population, benefits of performing screening tests may not be fully achieved for each sub-group by using an inspection strategy developed for the entire population. Thus, it may be of value for an individual to learn more information about his/her likehood to have the disease. To address this issue, I analyze the change in the expected delay if schedules are based on the whole population information or the individual information and provide numerical results.