Additional Calendars
Calendar Views
All
Athletics
Conferences and Meetings
Law School
Special Events

Master of Science Thesis Defense by: Bentolhoda Jafary

When: Monday, August 3, 2015
2:00 PM - 4:00 PM
Where: > See description for location
Cost: Free
Description: TOPIC: QUANTIFYING THE IMPACT OF CORRELATED FAILURES ON DISCRETE AND CONTINUOUS COMPONENT-BASED SYSTEMS

LOCATION: Lester W. Cory Conference Room. Science & Engineering Building (Group II), Room 213A

ABSTRACT:
This thesis presents three contributions to the reliability engineering research literature: (i) continuous reliability models for systems of non-identically distributed correlated components, (ii) discrete and continuous reliability models for correlated consecutive k-out-of-n Failed systems, and (iii) universal generating function (UGF)-based discrete and continuous reliability models for correlated multi-state systems. Contribution (i) extends traditional discrete methods based on a vector of independent Bernoulli random variables, introducing a quadratic number of correlation parameters to characterize correlations between the failures of each pair of components. The generalization to continuous time allows each component to follow a unique life distribution. Contributions (ii) and (iii) extend reliability models that assume component failures are statistically independent to the case where the components are correlated but identically distributed, meaning that the components possess a common reliability parameter, but also exhibit correlated failures according to a shared correlation parameter.

Models to explicitly consider correlation are especially important for the consecutive k-out-of-n Failed and multi-state systems modeling paradigms because the systems modeled by these methods are susceptible to correlated failure in real life. For example, the consecutive k-out-of-n Failed can model telecommunication networks, where the failure of any k components in sequence leads to system failure. Natural disasters and intentional attacks are often geospatially correlated and therefore models to quantify the impact of correlated failure will be beneficial to quantify the vulnerability of critical infrastructure. The multi-state systems modeling approach is also used to model the performance of power generation and distribution systems and thus also susceptible to accidental and intentional disruptions.

For each of the three reliability modeling paradigms considered, we examine the impact of correlation on several measures and metrics, including system reliability, failure distribution, hazard rate, mean time to failure (MTTF), availability, and mean residual life (MRL). The results indicate that the extended modeling methodologies can succinctly capture the impact of correlation on these measures and metrics.

NOTE: All ECE Graduate Students are ENCOURAGED to attend.
All interested parties are invited to attend. Open to the public.

Advisor: Dr. Lance Fiondella

Committee Members: Dr. Liudong Xing, Department of Electrical & Computer Engineering and Dr. Firas Khatib, Department of Computer and Information Science

*For further information, please contact Dr. Lance Fiondella at 508.999.8596, or by via email at lfiondella@umassd.edu.
Topical Areas: General Public, University Community, Electrical and Computer Engineering