Reliability analysis and maintenance modeling of multi-component systems subject to multiple dependent competing failure processes

System reliability analyses, involving multiple failure processes, are important and challenging research topics, particularly when failure processes, such as degradation processes and random shocks, are competing and dependent. When component degradation models are extended to complex systems with multiple components, different perspectives of dependency should be considered in system reliability modeling. In this research, potential dependence patterns are investigated among multiple failure processes within and among components in systems and probabilistic models are developed to assess system reliability performance. For the reliability modeling of complex systems, if one component in the system degrades or fails prematurely, it is possible that other components will also degrade or fail prematurely given the shared working environment, which means component failure times are dependent. Furthermore, if a shock arrives to the system, and its impact on one component is large, it is likely that the shock impact on other components is also large. Therefore, the models are extended to perform quantitative analyses for system reliability considering that the damages to the two failure processes caused by shocks are dependent. From a multi-component system level perspective, the dependent characteristics from shocks are complex and detailed, so the research is organized into several scenarios, i.e., dependent failure processes are considered in different ways. Stochastically dependent component degradation processes are also studied, and extended gamma process models are used to model the dependent degradation process. Based on these new reliability models, different maintenance policies are derived to provide cost effective maintenance plans. Different maintenance policies are investigated, including: (1) age replacement policies; (2) periodic inspection maintenance policies; (3) Condition-based maintenance policies; and (4) Individual component maintenance policy. This research has many meaningful research contributions. The dependent scenarios described above for shocks are studied for the first time in system reliability modeling, and it is also the initial system model based on stochastically dependent component degradation using the gamma process. Additionally, the on-condition maintenance plans and the individual component maintenance based on steady state system behavior are investigated for the first time. They represent more practical and cost effective policies for many engineering applications.Ph.D.Includes bibliographical referencesby Sanling Son