In the last decade, reliability problems have become a critical subject in power modules. Understanding design weakness and failure mechanisms of thick wire bond are two critical steps in managing the risk of wire bond heel crack which is the topic of this thesis. Although this thesis does not target a specific type of power modules, we note that thick wire bond heel crack failures occur in Insulated Gate Bipolar Transistors (IGBTs). In fact, our aim is to understand failure mechanism in 300μm thick wire bonds with different geometries and materials. Since these wires experience harsh environmental conditions and high load transients, the wires undergo repetitive flexural movement which causes heel crack due to fatigue. For the purpose of understanding this failure mechanism, two experimental setups are built and utilized. The first experimental setup loads the wires using constant currents and observes the response using a scanning laser vibrometer to measure the displacement. The second experimental setup applies repetitive prescribed displacement to the first foot of the wire and detects fatigue failure using a Wheatstone bridge. It is realized that wires have different displacement property depending on their geometry and material. Maximum displacements are observed for Al-H11 instead of CuCorAl and PowerCu.

The Millennium Tower in San Francisco, which was the tallest building in western United States, recorded excessive settlements shortly after construction. There were two possible reasons for the abnormal behavior, inadequate foundation design to support excessively heavy building load, and changes in subsurface conditions due to dewatering at the nearby construction site. This research aims to identify the most reasonable causes of the settlement of Millennium Tower. In this study, a centrifuge model was constructed to simulate the settlement of Millennium Tower and to predict the settlement trend in the future. A loading device using airbag was used to apply incremental load in-flight. Dewatering was also simulated during consolidation. The centrifuge modeling was performed based on the field subsurface data obtained from geotechnical reports. Ground materials were prepared by mixing different kinds of soil, and the foundation system was modeled in detail based on the information of foundation design. The model was tested under a centrifugal acceleration of 120-g. An airbag loading system was used to simulate the multi-stage construction sequence. After simulating the construction sequence, long-term consolidation settlement over a period of 20 years was conducted. The groundwater level was lowered to study the change in settlement caused by dewatering at the nearby construction site. Eight laser transducers measured the settlements of building and each ground layer, and the pore water pressure transducers measured the pore water pressure in the clay layers at four different depths. The average settlement obtained from the centrifuge model test showed good agreement with the field measurements. The centrifuge model testing showed that the temporary change in groundwater level did not affect the long-term settlement.

This study presents the effect of the stacking sequence and fiber orientation on a composite sandwich panel subjected to static in-plane bolt loading. Six plates were constructed with laminates of unidirectional carbon fiber and cross ply weaves of fiberglass. The orientations that were examined included 0, +/- 45, and 90 degrees. Half of the plates had fiberglass lamina on the outside of the laminate while the other three plates had the carbon fiber on the outside. Experimental and analytical tests were performed to determine the best orientations and stacking sequence. For the numerical analysis, plates with fibers oriented at +/- 45 degrees showed the highest strength. The experimental data also showed high strengths for the +/- 45 degree plates. However the experimental data also showed high strengths for the 90 degree laminate but with very high displacements. These high displacements would not allow the joint to maintain its relative position to the adjacent part. The discrepancy between the strength of the FEA models and the experimental data is attributed to inaccurate strength properties. The effect of in situ strength and compression strength was found to have a significant effect on the accuracy of the FEA solution. Good correlation was found between the FEA and experimental data in predicting the trend of the stiffness of the plates.

ITC/USA 2010 Conference Proceedings / The Forty-Sixth Annual International Telemetering Conference and Technical Exhibition / October 25-28, 2010 / Town and Country Resort & Convention Center, San Diego, California / Current failure analysis practices use diagnostic technology developed over the past 100 years of designing and manufacturing electrical and mechanical equipment to identify root cause of equipment failure requiring expertise with the equipment under analysis. If the equipment that failed had telemetry embedded, prognostic algorithms can be used to identify the deterministic behavior in completely normal appearing data from fully functional equipment used for identifying which equipment will fail within 1 year of use, can also identify when the presence of deterministic behavior was initiated for any equipment failure.

ITC/USA 2011 Conference Proceedings / The Forty-Seventh Annual International Telemetering Conference and Technical Exhibition / October 24-27, 2011 / Bally's Las Vegas, Las Vegas, Nevada / Spacecraft and launch vehicle reliability is dominated by premature equipment failures and surprise equipment failures that increase risk and decrease safety, mission assurance and effectiveness. Large, complex aerospace systems such as aircraft, launch vehicle and satellites are first subjected to most exhaustive and comprehensive acceptance testing program used in any industry and yet suffer from the highest premature failure rates. Desired/required spacecraft equipment performance is confirmed during factory testing using telemetry, however equipment mission life requirement is not measured but calculated manually and so the equipment that will fail prematurely are not identified and replaced before use. Spacecraft equipment mission-life is not measured and confirmed before launch as performance is but calculated using stochastic equations from probability reliability analysis engineering standards such as MIL STD 217. The change in the engineering practices used to manufacture and test spacecraft necessary to identify the equipment that will fail prematurely include using a prognostic and health management (PHM) program. A PHM includes using predictive algorithms to convert equipment telemetry into a measurement of equipment remaining usable life. A PHM makes the generation, collection, storage and engineering and scientific analysis of equipment performance data "mission critical" rather than just nice-to-have engineering information.

ITC/USA 2007 Conference Proceedings / The Forty-Third Annual International Telemetering Conference and Technical Exhibition / October 22-25, 2007 / Riviera Hotel & Convention Center, Las Vegas, Nevada / Telemetry Prognostics is Failure Prediction using telemetry for launch vehicle and satellite space flight equipment to stop launch failures, launch pad delays, satellite infant mortalities and satellite on orbit failures. This technology characterizes telemetry behaviors that are latent, transient, and go undetected by the most experienced engineering personnel and software diagnostic tools during integration and test, launch operations and on orbit activities stopping launch pad delays, launch failures, infant mortalities and on orbit failures. Telemetry prognostics yield a technology with state-of-the-art innovative techniques for determining critical on-board equipment remaining useful life taking into account system states, attitude reorientations, equipment usage patterns, failure modes and piece part failure characteristics to increase the reliability, usability, serviceability, availability and safety of our nation’s space systems.

The aim of this thesis is the stiffness analysis of cable-driven parallel robots. Cable-driven parallel robots have drawn considerable attention because of their unique abilities and advantages such as the large workspace, light weight of cable actuators, easy disassembly and transportation of the robot. The mobile platform of a cable-driven parallel robot is attached to the base with multiple cables. One of the parameters that should be studied to make sure a robot is able to execute a task accurately is stiffness of the robot. In order to investigate the stiffness behaviour of a robot, the stiffness matrix can be calculated as the first step. Because cables act in tension, keeping the positive tension in cables becomes a challenge. In order to have a fully controllable robot, an actuation redundancy is needed. These complexities are addressed in the thesis and simulations. In this thesis, the complete form of the stiffness matrix is considered without neglecting any terms in calculation of the stiffness. Some stiffness indices such as single-dimensional stiffness based on stiffness ellipse, directional stiffness and condition number of the stiffness matrix are introduced and calculated and stiffness maps of the robot are developed. In addition, the issue of unit inconsistency in calculating the stiffness index is addressed. One of the areas which is also addressed in this thesis is failure analysis based on the stiffness of robot. The effect of the failure in one or more cables or motors is modelled and stiffness maps are developed for the failure situation. It is shown that by changing the anchor position and mobile platform orientation, the lost stiffness after failure of a cable or motor can be retrieved partially. Optimum anchor position and mobile platform orientation are identified to maximize the area of the stiffness map. Condition number of the stiffness matrix while robot is following a trajectory is optimized. In addition, when one cable fails during the path planning, the recovery of the robot is studied. Finally, these analyses on stiffness and failure provide the designer with the necessary and valuable information about the anchor positions and actuator toques. / Thesis (Ph.D, Mechanical and Materials Engineering) -- Queen's University, 2013-04-27 08:47:26.297

The ever increasing demand of petroleum in the transport sector has led to depletion of low cost/low risk reserves, increased levels of pollution, and greenhouse gas emissions that take a heavy toll on the environment as well as the national economy. There is an urgent need to use alternative energy resources along with an efficient and affordable energy conversion system to arrest environmental degradation. Polymer electrolyte fuel cells (PEFCs) show great promise in this regard - they use hydrogen gas as a fuel that electrochemically reacts with air to produce electrical energy and water as the by product. In a fuel cell electric vehicle (FCEV), these zero tail pipe emission systems offer high efficiency and power density for medium-heavy duty and long range transportation. However, PEFC technology is currently challenged by its limited durability when subjected to harsh and adverse operating conditions and transients that arises during the normal course of vehicle operation. The hydrogen-based fuel cell power train for electric vehicles must achieve high durability while maintaining high power efficiency and fuel economy in order to equal the range and lifetime of an internal-combustion-engine vehicle. The technology also needs to meet the cost targets to make FCEVs a commercial success. In this dissertation, one of the degradation phenomena that severely impede the durability of the system has been investigated. In scenarios where the cell becomes locally starved of hydrogen fuel, “cell reversal” occurs, which causes the cell to consume itself through carbon corrosion and eventually fail. Carbon corrosion in the anode disrupts the original structure of the electrode and can cause undesirable outcomes like catalyst particle migration, aggregation, loss of structural and chemical integrity. A comprehensive study using advanced electrochemical diagnostics and high resolution 3D imaging was performed and a new understanding to extend PEFC life time and robustness, by implementing engineered materials solutions, has been achieved. This will eventually help in making fuel cell systems more efficient, durable and economically viable, in order to better harness clean energy resources.

Diagnosis is the first analysis step for uncovering the root cause of failure for a defective chip. It is a fast and non-destructive approach to preliminarily identify and locate possible defects in a failing chip. Despite many advances in diagnosis techniques, it is often the case, however, that resolution, i.e., the number of locations or candidates reported by diagnosis, exceeds the number of actual failing locations. To address this major challenge, a novel, machine-learning-based resolution improvement methodology named PADRE (Physically-Aware Diagnostic Resolution Enhancement) is described. PADRE uses easily-available tester and simulation data to extract features that uniquely characterize each candidate. PADRE applies machine learning to the features to identify candidates that correspond to the actual failure locations. Through various experiments, PADRE is shown to significantly improve resolution with virtually no negative impact on accuracy. Specifically, in simulation experiments, the number of defects that have perfect resolution is increased by 5x with little degradation of accuracy. An important investigation that typically follows diagnosis is Physical Failure Analysis (PFA), which can also provide information that is helpful for improving diagnosis. PADRE influences PFA within a novel, active learning (AL) based PFA selection approach. An active-learning based PADRE (AL PADRE) selects the most useful defects for PFA in order to improve diagnostic resolution. Experiments show AL PADRE can reach an accuracy of 90% with 60% less PFA, on average, compared to conventional defect selection for PFA. In addition, during the yield learning process, the failing mechanisms that lead to defective chips may change due to perturbations in the fabrication process. It is important for PADRE to perform robustly through the entire yield learning process. Therefore, additional techniques are developed to monitor the effectiveness of PADRE in real time, as well as to update PADRE efficiently and stably to cope with changing failure mechanisms.

This thesis deals with a common cause failure analysis (CCF) for space devices. This analysis belongs among dependability analyses, which have not been sufficiently developed in a field of space industry in corresponding technical and normative documents. Therefore, we focused on devising a new procedure of a qualitative and quantitative common cause failure analysis for the space applications herein. This new procedure of the qualitative and quantitative CCF analysis was applied on redundant systems of a special space device microaccelerometer (ACC), which was developed in VZLÚ. Performance of the qualitative CCF analysis can lead to recommendations to change design of the system, making the system less susceptible to the common cause failures. Performance of the quantitative CCF analysis and its inclusion into the computation of the system reliability can lead to a more accurate estimation of the reliability (in most cases it leads to decreasing the system reliability). During the development of the ACC there were not defined any requirements to perform the CCF analysis within general dependability requirements (defined by the customer and by ECSS standards). Hence, we compared computations of the ACC reliability with and without considering the CCFs. When the CCFs were considered, the reliability of the ACC was decreased according to our assumption. On our example of the ACC we showed advantages of the performance of the CCF analysis within the dependability analyses during development of the space devices.