Browsing by Subject "Accelerated life testing"
Now showing 1 - 3 of 3
Results Per Page
Sort Options
Item Accelerated life testing with several type II censored samples(1989-08) McCoun, Kelly Lynn; Kolarik, William J.; Davenport, James M.; Duran, Benjamin S.; Lewis, Truman O.; Ghiassi, Hossein MansouriThe asymptotic normality of the maximum likelihood estimator for a likelihood function based on several independent Type II censored samples is established. Examples from the accelerated Hfe class of models are considered with regard to this result. This extends previous work for the single sample case. Conditional confidence interval estimation is discussed for the parameters in the Arrhenius and Eyring models based on data from several Type II censored samples. Conditional confidence interval estimates for the average life of a component under user-specified stress, the acceleration factor or the factor which expresses how many times longer a component should last at one stress relative to another, and the 7th quantile of the failure time distribution of a component under user-specified stress are also provided. Approximate confidence intervals for these parameters are given for comparison with the conditional confidence intervals.Item Design of an electro-mechanical hexapod for accelerated life testing of optical fiber assemblies(2010-05) Soukup, Ian Michael; Fahrenthold, Eric P.; Beno, Joseph H.The quantity and length of optical fibers required for the Hobby-Eberly Telescope Dark Energy eXperiment (HETDEX) create unique fiber handling challenges. More than 33,000 optical fibers will enable the Hobby-Eberly Telescope (HET) to collect data on at least one million galaxies that are 9 billion to 11 billion light-years away, yielding the largest map of the universe ever produced [1,2]. The design advantages made possible by optical fibers also forms challenges to prevent damage to the fragile fibers that can lead to Focal Ratio Degradation (FRD) [3]. Therefore, a life cycle test must be conducted to study fiber behavior and measure FRD as a function of time. This thesis describes the design and design methodology of an electro-mechanical test apparatus for accelerated life testing of optical fiber assemblies. The design methodology summarizes the development of functional requirements and constraints that drove the design. The test apparatus design utilizes six linear actuators to replicate the movement of the fiber system deployed on HETDEX for over 65,000 accelerated cycles, simulating five years of actual operation. The electro-mechanical test apparatus will provide insight into the effects of load history on the performance of optical fibers which published data has thus far been lacking. Performance of the electro-mechanical test apparatus will be demonstrated through simulation, modeling and calculations. The test results that will be generated from the accelerated life test will be of great interest to designers of robotic fiber handling systems for major telescopes.Item Predicting in-service fatigue life of flexible pavements based on accelerated pavement testing(2007-12) Guo, Runhua, 1975-; Prozzi, Jorge AlbertoPavement performance prediction in terms of fatigue cracking and surface rutting are essential for any mechanistically-based pavement design method. Traditionally, the estimation of the expected fatigue field performance has been based on the laboratory bending beam test. Full-scale Accelerated Pavement Testing (APT) is an alternative to laboratory testing leading to advances in practice and economic savings for the evaluation of new pavement configurations, stress level related factors, new materials and design improvements. This type of testing closely simulates field conditions; however, it does not capture actual performance because of the limited ability to address long-term phenomena. The same pavement structure may exhibit different response and performance under APT than when in-service. Actual field performance is better captured by experiments such as Federal Highway Administration's Long-Term Pavement Performance (LTPP) studies. Therefore, to fully utilize the benefits of APT, there is a need for a methodology to predict the long-term performance of in-service pavement structures from the results of APT tests that will account for such differences. Three models are generally suggested to account for the difference: shift factors, statistical and mechanistic approaches. A reliability based methodology for fatigue cracking prediction is proposed in this research, through which the three models suggested previously are combined into one general approach that builds on their individual strengths to overcome some of the shortcomings when the models are applied individually. The Bias Correction Factor (BCF) should account for all quantifiable differences between the fatigue life of the pavement site under APT and in-service conditions. In addition to the Bias Correction Factor, a marginal shift factor, M, should be included to account for the unquantifiable differences when predicting the in-service pavement fatigue life from APT. The Bias Correction Factor represents an improvement of the currently used "shift factors" since they are more general and based on laboratory testing or computer simulation. By applying the proposed methodology, APT performance results from a structure similar to an in-service structure can be used to perform four-point bending beam tests and structural analysis to obtain an accurate estimate of the necessary Bias Correction Factor to estimate in-service performance.