From: Paul Gruhn (pgruhn(at)ix.netcom.com)
Date: Thu 05 Feb 2004 - 12:49:46 GMT
If you don't include test intervals, then you're just doing fault tree math of multiplying or adding simple probabilities (series or parallel, depending upon failure mode). That's my understanding of the very original development of RBDs. I believe there is an IEC standard on RBDs. Here are other sources I know of (but they all include TI). 1) IEC 61508 part 6 (www.iec.ch) 2) ISA TR 84.00.02 (Technical report on system modeling) (www.isa.org) 3) ISA textbook "Safety Shutdown Systems - Design, Analysis & Justification" (ISBN ISBN 1-55617-665-1) 4) "Reliability, Maintainability and Risk (Practical Methods for Engineers)", 6th Edition, David J. Smith, Butterworth-Heinemann, 2001, ISBN 0-7506-5168-7 Paul Gruhn, PE, CFSE L&M Engineering Houston, TX www.landmengineering.com pgruhn(at)landmengineering.com -----Original Message----- From: safety-critical-request(at)cs.york.ac.uk [mailto:safety-critical-request(at)cs.york.ac.uk] On Behalf Of Amann Michael FRD TE-PS Sent: Thursday, February 05, 2004 3:11 AM To: 'safety-critical(at)cs.york.ac.uk' Subject: AW: [sc] Reliability Block Diagrams (RBD) Hi, sorry for the late response but I have tried to find RBD tools and formulaes which have modeles/formulae without Proof Test (T1) and I have failed to do so. Does anybody know where to find such formulae? Regards, Michael Amann TE-PS ZF Friedrichshafen AG 88038 Friedrichshafen, Deutschland Telefon + 49 7541 77-7357, Telefax +49 7541 77-907357 m.amann(at)zf.com -----Ursprüngliche Nachricht----- Von: Tony Foord [mailto:Tony.Foord(at)4-sightConsulting.co.uk] Gesendet: Donnerstag, 15. Januar 2004 22:38 An: safety-critical(at)cs.york.ac.uk Betreff: [sc] Reliability Block Diagrams (RBD) Michael I suspect the difference you mean is not the method of modelling (Markov, RBD, etc) but whether failures are hidden/covert/unrevealed. The purpose of the Proof Test is to identify such failures so that the failed element may be repaired or replaced and then the PFD is directly related to the proof test interval T1 (ignoring the human factors associated with proof tests and assuming the repair time is very short compared to T1 and also that T1 is very short compared to the MTTF). If a proof test is not possible and thus the failed unit is not repaired or replaced, then different formulae apply. If failures are known/revealed as soon as they occur then the failed element may be repaired or replaced without waiting for a proof test to discover the failure and again different formulae apply. The RBD tools I have used have models for all the scenarios above. Regards Tony Tel: +44 (0)1582 462 324 Fax: +44 (0)1582 623 470 Mailto: Tony.Foord(at)4-sightConsulting.co.uk Web page: http://4-sightConsulting.co.uk > -----Original Message----- > From: Michael Amann [mailto:michael.amann.brz(at)t-online.de] > Sent: 15 January 2004 18:02 > To: safety-critical(at)cs.york.ac.uk > Subject: [sc] Reliability Block Diagrams > > > Hi, > > In Part 6 of the IEC 61508 there a several formulas given for > calculation of PFD and PFH and also some "lookup tables" are provided > for several "configurations" (1oo1, 2002, ...). In all these formulas > the "Proof Test Interval T1" is required. But how can I use the > formulas if for my system no proof test is planned or a proof test is > not possible? > > With a Markov-model it is "easy" to leave away the proof test, but > what do I have to do for RBDs? > > Best regards, > Michael Amann >