Section: Business and Industry Section Meeting
Location: Council Chamber, Royal Statistical Society, Errol Street, London
Date: 13th June 2007
Time: 4.00
Tea/Coffee: 3.30.
Speaker: Professor Martin Newby
Title: Reliability through the life cycle: linking reliability arguments to the
design process (A Thomas, M Douglas, Rolls-Royce PLC
M Newby, V Tsachouridis, City University)
Abstract
The talk describes the development and implementation of a "Designer's workbench"
for reliability. As more emphasis is placed on reliability throughout the life cycle and
on lifecycle costs it becomes necessary to encourage reliability evaluations at all
stages of the design process, from conceptual design, through technical design and
into the usage phase. This project links the stages of the design process as defined
within Rolls-Royce to the systematic evaluation of reliability at each stage. The
workbench provides project members with guidance on reliability issues through
queries to databases such as FRACAS, Lessons Learned, Design Rules and Protocols
and service data. Analysis is at the component and failure mode level and allows
lifetime distributions to be determined from expert opinions and to be updated within
a Bayesian framework by experimental and service data. Discrepancies between
prior opinions and service experience are indicated. The workbench establishes a
traceable process for the reliability evaluations passed from one stage of a project to
the next. The workbench provides a user friendly graphical interface and query
system.
Speaker: Dr Michael Phillips
Title: Systems Reliability Modelling and Analysis (J Ansell, The University of
Edinburgh, and M J Phillips, University of Leicester)
Abstract
The modelling and analysis of systems reliability is critical in many industries. The
costs associated with failure and its consequences provide a need for a clear
description of the system in terms of a mathematical model that adequately reflects
the nature of the system. As well as modelling it is necessary to consider the analysis
of data from a system to assess the overall rate of the system failure. A smoothing
approach described allows for the incorporation of covariates into the modelling in
order to assess the overall performance of the system.
|