SENG 421 Reliability Demonstration Chart

SENG 421 Reliability Demonstration Chart
paly

This is a demonstration chart created by the Department of Electrical & Computer Engineering at the University of Calgary for the SENG 421 Reliability course. The chart, developed by Professor B

About SENG 421 Reliability Demonstration Chart

PowerPoint presentation about 'SENG 421 Reliability Demonstration Chart'. This presentation describes the topic on This is a demonstration chart created by the Department of Electrical & Computer Engineering at the University of Calgary for the SENG 421 Reliability course. The chart, developed by Professor B. The key topics included in this slideshow are . Download this presentation absolutely free.

Presentation Transcript


Slide1SENG 421 Reliability  Demonstration Chart SENG  421 Reliability  Demonstration Chart Department of Electrical & Computer Engineering, University of Calgary B.H. Far  ( far@ucalgary.ca ) http://www.enel.ucalgary.ca/People/far/Lectures/SENG421/

Slide2SENG421 (Winter 2005)far@ucalgary.ca 2 Reliability  Demo  Chart   /1 Reliability  Demo  Chart   /1  An efficient way of checking whether the FIO (  F ) is met or not.  It is based on collecting failure data at time points.  Vertical axis:  failure number ( n )  Horizontal axis: normalized failure data ( T n ), i.e., failure time      F Figure from Musa’s Book

Slide3SENG421 (Winter 2005)far@ucalgary.ca 3 Parameters  Involved   /1 Parameters  Involved   /1  Discrimination  ratio  (  ):  Discrimination  ratio  (  ):  Acceptable error in estimating failure intensity.  Customer  risk  (  )  :  Customer  risk  (  )  :  Probability that the developer is willing to accept of falsely saying the failure intensity objective is met (i.e.,  acceptance ) when it is not.  Developer  risk  (  )  :  Developer  risk  (  )  :  Probability that the developer is willing to accept of falsely saying the failure intensity objective is  not  met (i.e.,  rejection ) when it is.

Slide4SENG421 (Winter 2005)far@ucalgary.ca 4 Parameters  Involved   /2 Parameters  Involved   /2    For    =10%  and    = 10%  and    =2  )   There is 10% risk (  )  of wrongly accepting the software when its failure intensity objective is actually equal or greater than twice (   =2) the failure intensity objective.    There is 10% risk (  ) of wrongly rejecting the software when its failure intensity objective is actually equal or less than half (   =2) the failure intensity objective.

Slide5SENG421 (Winter 2005)far@ucalgary.ca 5 Reliability  Demo  Chart   /2 Reliability  Demo  Chart   /2  A  changes rapidly with  customer risk  but very slightly with  developer risk  and it determines the intercept of  accept   boundary with the horizontal line n=0  B  changes rapidly with  developer risk  but very slightly with  customer risk  and it determines the intercept of  reject   boundary with the vertical line T n =0

Slide6SENG421 (Winter 2005)far@ucalgary.ca 6 Reliability  Demo  Chart   /3 Reliability  Demo  Chart   /3  Boundary between  accept   and  continue  regions  (    is the discrimination ratio)  (    is the discrimination ratio)  Boundary between  reject  and  continue  regions

Slide7SENG421 (Winter 2005)far@ucalgary.ca 7 Reliability  Demo  Chart   /4 Reliability  Demo  Chart   /4 Values of intercepts of boundaries with various horizontal and vertical axes

Slide8SENG421 (Winter 2005)far@ucalgary.ca 8 Reliability  Demo  Chart   /5 Reliability  Demo  Chart   /5 Values of A and B for various consumer and supplier risk levels Table from Musa’s Book

Slide9SENG421 (Winter 2005)far@ucalgary.ca 9 Reliability  Demo  Chart   /6 Reliability  Demo  Chart   /6  When risk levels (   and   ) decrease, the system will require more test before reaching the accept or reject regions,  i.e ., the continue region becomes wider.  When discrimination ratio (  ) decreases, the system will require more test before reaching the accept or reject regions,  i.e ., the continue region becomes wider.

Slide10SENG421 (Winter 2005)far@ucalgary.ca 10 RDC:  Example    /1 RDC:  Example    /1  Consumer risk   = 5%  Supplier risk     = 5%  Discrimination  ratio    = 2 Figure from Musa’s Book

Slide11SENG421 (Winter 2005)far@ucalgary.ca 11 RDC:  Example    /2 RDC:  Example    /2  Consumer risk     = 1%  Supplier risk     = 1%  Discrimination  ratio    = 2 Figure from Musa’s Book

Slide12SENG421 (Winter 2005)far@ucalgary.ca 12 RDC:  Example    /3 RDC:  Example    /3  Consumer risk     =  0.1%  Supplier risk     = 0.1%  Discrimination  ratio    = 2 Figure from Musa’s Book

Slide13SENG421 (Winter 2005)far@ucalgary.ca 13 RDC:  Example    /4 RDC:  Example    /4  Consumer risk     = 10%  Supplier risk     = 10%  Discrimination  ratio    = 1.2 Figure from Musa’s Book

Slide14SENG421 (Winter 2005)far@ucalgary.ca 14 Example  1 Example  1 Failure number Measure (million transactions) Normalized Measure (MTTF) 1 0.1875 0.75 2 0.3125 1.25 3 1.25 5

Slide15SENG421 (Winter 2005)far@ucalgary.ca 15 Example  2 Example  2 Failure number Measure (CPU hour) Normalized Measure (MTTF) 1 8 0.8 2 19 1.9 3 60 6

Slide16SENG421 (Winter 2005)far@ucalgary.ca 16 Example  3 Example  3  We are going to buy a new colour laser printer for our department. We have borrowed the printer for the test run and we are going to conduct certification test on it. Maker’s data shows that we need to change the toner every 10,000 pages. We would like to have the system running without any failure between the two consecutive toner changes and in the worst case having only one failure during the period. a) What shall be our failure intensity objective for the system?  F  = 1/10000 pages

Slide17SENG421 (Winter 2005)far@ucalgary.ca 17 Example  3  (contd.) Example  3  (contd.) b) We observe that failures occur at 4,000 pages, 6,000 pages, 10,000 pages, 11,000 pages, 12,000 pages and 15,000 pages of output. Using the reliability demonstration chart, what can we conclude about this printer?

Slide18SENG421 (Winter 2005)far@ucalgary.ca 18 Example  3  (contd.) Example  3  (contd.)  Because of failing the certification test we will reject the printer.