Thesis
Reliability analysis for systems with stochastic and spatial dependence
- Creator
- Rights statement
- Awarding institution
- University of Strathclyde
- Date of award
- 2022
- Thesis identifier
- T16417
- Person Identifier (Local)
- 201675698
- Qualification Level
- Qualification Name
- Department, School or Faculty
- Abstract
- This thesis develops models for system reliability prediction considering unobserved heterogeneity and spatial dependence. Three problems are considered in this thesis. The first problem concerns models to predict system and component level reliability for systems subject to minimal repair and unobserved heterogeneity. Existing models assume that unobserved heterogeneity is constant and methods for component-level prediction had not been considered. However, there are situations where unobserved heterogeneity changes and becomes homogeneous over time. This thesis develops a new reliability model that accounts for the case where unobserved heterogeneity changes and becomes homogeneous over time. We develop a new frailty model using Inverse Gaussian (IG) distribution and develop a method using Empirical Bayes that enable component-level reliability prediction. The second problem concerns models for reliability assessment for load sharing systems with spatially dependent components and proximity effects. Existing models assume equal load sharing for systems subject to load-sharing. However, there are systems that operates in a way that the load of a failed component is transferred to the working proximate components. Existing models do not account for the proximity effect. This thesis develops a new reliability model that accounts for load sharing and proximity effect between components. We introduce a function that captures the effect of each load change on the failure rate of a working proximate component. Numerical examples are presented to illustrate the developed model. The third problem concerns models for reliability assessment and preventive maintenance for load sharing systems with spatially dependent components, proximity effects and shocks. Existing models assume equal-load sharing for systems subject to load-sharing and external shocks. However, models that account for external shocks and proximity effect in the system has not been considered. This thesis develops a function that captures the effect of each load change and shock on the failure rate of a working component. We introduce a Modified failure sequence diagram and an algorithm for system reliability assessment based on Monte Carlo method. In addition, we develop an extension of the age-replacement model for preventive maintenance of the load sharing system. Numerical examples are presented to illustrate the developed models.
- Advisor / supervisor
- Liu, Bin
- Revie, Matthew
- McIntyre, Stuart
- Resource Type
- DOI
Relations
Items
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
|
PDF of thesis T16417 | 2023-02-07 | Public | Download |