Home
Resource
Reliability modeling, testing and optimization of systems with mixtures of one-shot units
PDF
PDF format is widely accepted and good for printing.
Plug-in required
PDF-1(2.80 MB)
Citation & Export
View Usage Statistics
Staff View

Citation & Export
Hide

Simple citation

Cheng, Yao. Reliability modeling, testing and optimization of systems with mixtures of one-shot units. Retrieved from https://doi.org/doi:10.7282/T3QC06Q9

Export

Click here for information about Citation Management Tools at Rutgers.

Statistics
Hide

Description

TitleReliability modeling, testing and optimization of systems with mixtures of one-shot units
NameCheng, Yao (author); Elsayed, Elsayed (chair); Albin, Susan (internal member); Pham, Hoang (internal member); Guo, Weihong (internal member); Kolassa, John (outside member); Rutgers University; School of Graduate Studies
Date Created2018
Other Date2018-01 (degree)
SubjectIndustrial and Systems Engineering
Extent1 online resource (xiv, 239 p. : ill.)
DescriptionOne-shot units are usually produced and stored (or used as standby) in batches until retrieved. These units can be defined as a system which may experience degradations or sudden failures during its storage period. To assess the reliability performance of the units, reliability tests are repeatedly (in non-identical pattern) and randomly conducted across the lifetime of the units; where corresponding actions are taken afterwards. The continuous arrival of batches and conduct of tests induce the system contains a mixture of nonhomogeneous units, which is defined as a general “k-out-of-n: F system” with k and n nonhomogeneous and time-dependent. In this dissertation, we propose models to investigate the reliability metrics of the system under a variety of scenarios. Extensive simulation studies are performed to validate the models. Failure or degradation caused by thermal fatigue is a pervasive phenomenon during the one-shot units’ storage period. The Birnbaum-Saunders (BS) distribution is specifically developed for describing mechanical fatigue failures, but limited in describing a variety of hazard functions. Hence it is reasonable to investigate whether the generalized form of BS (GBS) distribution can be extended for modeling the plastic deformation induced by thermal cyclic stresses and providing reliability metrics of units subject to thermal fatigue. In this dissertation, we investigate system reliability metrics when subjecting to thermal fatigue failure by adopting the GBS accelerated model. The one-shot units might experience competing failure modes during its storage period. Specifically, repeated thermal cyclic tests (TCTs) are randomly conducted; at the end of an arbitrary TCT, the unit’s failure is observed either when any of its failure modes occurs suddenly or when any of its degradation modes reach its “failure threshold”. Under such circumstances, unit’s failure data cannot be described by a single failure time distribution; instead, a competing failure model which considers multiple failure modes is adopted to assess unit’s reliability metrics. The units’ potential failure modes as well as the reliability metrics of the system under competing failure modes are investigated in this dissertation. Due to the characteristics of one-shot units and recent advances in technology and materials, one-shot units are usually highly reliable and it is impractical to obtain one-shot units’ failure (degradation) data under operating conditions. Accelerated life testing (ALT) is an efficient approach to obtain failure/degradation observations in a much shorter time period and utilize the test data to predict reliability metrics under normal operating conditions. We develop physics-statistics-based models and obtain optimal sequential accelerated non-destructive test (NDT) plans under different scenarios. The efficiency of the NDT plans is validated by comparing the system reliability metrics obtained under accelerated and normal conditions. NDT assesses unit’s functionality without permanent damage in order to demonstrate the unit’s reliability. However, one cannot make decisions regarding system reliability by only depending on NDT results because NDT does not fully perform the unit’s functionality. In contrast, destructive testing (DT) fully tests the unit’s functionality but destroys the units. This intensifies the need to investigate hybrid reliability tests that include both NDT and DT. In this dissertation, an optimal sequential hybrid reliability testing plan is designed and the results of the tests are utilized to improve the accuracy of the system reliability metrics estimation. We validate that by conducting hybrid reliability test, the unit’s lifetime parameters approach their true values. As the reliability estimation converge, we decrease the number of units tested in DTs and eventually perform NDT only. There exists many situations that a specific number of one-units are used consecutively when put into operational use. Therefore, it becomes interesting and challenging to determine the characteristics and sequence of the one-shot units to be launched such that the operational use of the launched units is optimized. Defining the launched one-shot units as a system, we investigate the reliability metrics of the system to optimize the system’s operational use at arbitrary time by formulating an optimization problem which is applicable to a variety of objectives. We also provide the bounds of the system’s successful operational probability estimation.
NotePh.D.
NoteIncludes bibliographical references
Noteby Yao Cheng
Genretheses, ETD doctoral
Persistent URLhttps://doi.org/doi:10.7282/T3QC06Q9
Languageeng
CollectionSchool of Graduate Studies Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Version 8.5.5
Rutgers University Libraries - Copyright ©2024