A. BobbioReggio Emilia, June 17-18, Dependability & Maintainability Theory and Methods 3. Reliability Block Diagrams Andrea Bobbio Dipartimento di Informatica Università del Piemonte Orientale, “A. Avogadro” Alessandria (Italy) - IFOA, Reggio Emilia, June 17-18, 2003
A. BobbioReggio Emilia, June 17-18, Model Types in Dependability Combinatorial models assume that components are statistically independent: poor modeling power coupled with high analytical tractability. Reliability Block Diagrams, FT, …. State-space models rely on the specification of the whole set of possible states of the system and of the possible transitions among them. CTMC, Petri nets, ….
A. BobbioReggio Emilia, June 17-18, Reliability Block Diagrams Each component of the system is represented as a block; System behavior is represented by connecting the blocks; Failures of individual components are assumed to be independent; Combinatorial (non-state space) model type.
A. BobbioReggio Emilia, June 17-18, Reliability Block Diagrams (RBDs) Schematic representation or model; Shows reliability structure (logic) of a system; Can be used to determine dependability measures; A block can be viewed as a “switch” that is “closed” when the block is operating and “open” when the block is failed; System is operational if a path of “closed switches” is found from the input to the output of the diagram.
A. BobbioReggio Emilia, June 17-18, Reliability Block Diagrams (RBDs) Can be used to calculate: –Non-repairable system reliability given: Individual block reliabilities (or failure rates); Assuming mutually independent failures events. –Repairable system availability given: Individual block availabilities (or MTTFs and MTTRs); Assuming mutually independent failure and restoration events; Availability of each block is modeled as 2-state Markov chain.
A. BobbioReggio Emilia, June 17-18, Series system of n components. Components are statistically independent Define event E i = “component i functions properly.” Series system in RBD A1A1A2A2AnAn P(E i ) is the probability “component i functions properly” the reliability R i (t) (non repairable) the availability A i (t) (repairable)
A. BobbioReggio Emilia, June 17-18, Reliability of Series system Series system of n components. Components are statistically independent Define event E i = "component i functions properly.” A1A1A2A2AnAn Denoting by R i (t) the reliability of component i Product law of reliabilities:
A. BobbioReggio Emilia, June 17-18, Series system with time-independent failure rate Let i be the time-independent failure rate of component i. Then: The system reliability Rs(t) becomes: R s (t) = e - s t with s = i i=1 n R i (t) = e - i t 1 1 MTTF = —— = ———— s i i=1 n
A. BobbioReggio Emilia, June 17-18, Availability for Series System Assuming independent repair for each component, where A i is the (steady state or transient) availability of component i
A. BobbioReggio Emilia, June 17-18, Series system: an example
A. BobbioReggio Emilia, June 17-18, Series system: an example
A. BobbioReggio Emilia, June 17-18, Improving the Reliability of a Series System Sensitivity analysis: R s R s S i = ———— = ———— R i R i The optimal gain in system reliability is obtained by improving the least reliable component.
A. BobbioReggio Emilia, June 17-18, The part-count method It is usually applied for computing the reliability of electronic equipment composed of boards with a large number of components. Components are connected in series and with time- independent failure rate.
A. BobbioReggio Emilia, June 17-18, The part-count method
A. BobbioReggio Emilia, June 17-18, Redundant systems When the dependability of a system does not reach the desired (or required) level: Improve the individual components; Act at the structure level of the system, resorting to redundant configurations.
A. BobbioReggio Emilia, June 17-18, Parallel redundancy A system consisting of n independent components in parallel. It will fail to function only if all n components have failed. E i = “The component i is functioning” E p = “the parallel system of n component is functioning properly.” A1A1 AnAn
A. BobbioReggio Emilia, June 17-18, Parallel system Therefore :
A. BobbioReggio Emilia, June 17-18, Parallel redundancy F i (t) = P (E i ) Probability component i is not functioning (unreliability) R i (t) = 1 - F i (t) = P (E i ) Probability component i is functioning (reliability) A1A1 AnAn — F p (t) = F i (t) i=1 n R p (t) = 1 - F p (t) = 1 - (1 - R i (t)) i=1 n
A. BobbioReggio Emilia, June 17-18, component parallel system For a 2-component parallel system: F p (t) = F 1 (t) F 2 (t) R p (t) = 1 – (1 – R 1 (t)) (1 – R 2 (t)) = = R 1 (t) + R 2 (t) – R 1 (t) R 2 (t) A1A1 A2
A. BobbioReggio Emilia, June 17-18, component parallel system: constant failure rate For a 2-component parallel system with constant failure rate: R p (t) = A1A1 A2 e - 1 t + e - 2 t – e - ( ) t MTTF = —— + —— – ————
A. BobbioReggio Emilia, June 17-18, Parallel system: an example
A. BobbioReggio Emilia, June 17-18, Partial redundancy: an example
A. BobbioReggio Emilia, June 17-18, Availability for parallel system Assuming independent repair, where A i is the (steady state or transient) availability of component i.
A. BobbioReggio Emilia, June 17-18, Series-parallel systems
A. BobbioReggio Emilia, June 17-18, System vs component redundancy
A. BobbioReggio Emilia, June 17-18, Component redundant system: an example
A. BobbioReggio Emilia, June 17-18, Is redundancy always useful ?
A. BobbioReggio Emilia, June 17-18, Stand-by redundancy A B The system works continuously during 0 — t if: a)Component A did not fail between 0 — t b)Component A failed at x between 0 — t, and component B survived from x to t. x 0 t A B
A. BobbioReggio Emilia, June 17-18, Stand-by redundancy A B x 0 t A B
A. BobbioReggio Emilia, June 17-18, A B Stand-by redundancy (exponential components)
A. BobbioReggio Emilia, June 17-18, Majority voting redundancy A1A1 A2 A3 Voter
A. BobbioReggio Emilia, June 17-18, :3 majority voting redundancy A1A1 A2 A3 Voter