UNSPECIFIED (2000) An event based database for the support of a holistic reliability assessment tool. In: Conference on Foresight and Precaution, MAY 15-17, 2000, EDINBURGH, SCOTLAND.

Full text not available from this repository.

Abstract

In recent years the use of classical methods of reliability prediction has been discouraged because of the observed inaccuracies obtained from such prediction processes. In many cases the companies using such classical techniques, possibly drawn from the many available prediction handbooks, used a "fiddle factor" that adjusted the predicted reliability to a value near what the companies' engineers knew to be about right. This "fiddle factor" was obtained from the past history of the companies' products and the experience of the engineers performing the prediction. The REMM (Reliability Enhancement Methodology and Modeling) programme is about improving the prediction of equipment reliability by using a holistic technique, which draws on expert knowledge, backed up by data collected throughout the entire product lifecycle. The obvious candidate for implementing such a process is to use an expert system that can evaluate system similarity and relate the, use of reliability techniques and collected data to the reliability of systems. In order to support such an expert system the REMM database programme will extract data from the company's records and store it in a manner that can be mined by the expert system. The databases will consist of a number of tables that will include information about the products manufactured by the company, the manufacturing processes used and more importantly a series of tables that will contain a list of the events that have occurred for each piece of equipment. An event is any action performed on equipment or any happening that may effect or contain information pertaining to the reliability of said equipment. An event is therefore a field failure, a performance test failure, or indeed a test pass, or a passage of the system through a particular manufacturing process. It is by considering the event histories of equipment that a number of useful reliability conclusions may be drawn. This paper describes the format of these databases, as implemented for the aerospace partners within the REMM programme. Each database is described and the use and derivation of event based data is explained and some indications of how the event data may be used to support the decision making process is given. Other possible uses for the data collected are postulated, such as looking for common mode failures.

Item Type:	Conference Item (UNSPECIFIED)
Subjects:	T Technology > TA Engineering (General). Civil engineering (General)
Journal or Publication Title:	FORESIGHT AND PRECAUTION, VOLS 1 AND 2
Publisher:	A A BALKEMA PUBLISHERS
ISBN:	90-5809-140-6
Editor:	Cottam, MP and Pape, RP and Harvey, DW
Date:	2000
Number of Pages:	7
Page Range:	pp. 331-337
Publication Status:	Published
Title of Event:	Conference on Foresight and Precaution
Location of Event:	EDINBURGH, SCOTLAND
Date(s) of Event:	MAY 15-17, 2000
URI:	http://wrap.warwick.ac.uk/id/eprint/12679

Data sourced from Thomson Reuters' Web of Knowledge

Request changes to a record

Actions (login required)

View Item