SRC Forum - Message Replies

Forum: Reliability & Maintainability Questions and Answers

Topic: Reliability & Maintainability Questions and Answers

Topic Posted by: Reliability & Maintainability Forum ( )
Organization: System Reliability Center
Date Posted: Mon Aug 31 12:47:36 US/Eastern 1998

Back to message list Show all replies Topics List About this forum
Original Message:

Posted by: Maurizio Saponetta ( )
Date posted: Mon Mar 5 15:31:44 US/Eastern 2007
Subject: Convertion between Calendar and Operating F/R
Dear Sirs, i am making experience in predicting Failure Rate by using the PRISM model, but i am surprised by the very low figure i obtain with respect the result that can be calculated for the same item using the model MIL-HDBK-217. I know that PRISM calculates Failure Rate in calendar hours, so i tried to convert the PRISM F/R in operating hours following two different ways: the first was varing the Duty Cycles from 10 to 100% in steps of 10 and proportionally decreasing the Cycling rate from 1000 to 1. In the second i adopted the method you suggested to other users of dividing the PRISM F/R by the decimal form of the Duty Cycle (i.e. DC=50% shall be diveded by 0.5).

The result of conversion are reported here below and thay do not appear congruent:
At DC 4% Prism F/R was 2.69,Conv Op.Time(/0.04) 67,27Fpmoh
At DC 10% Prism F/R was 2,693,Conv Op.Time(/0.1) 26,931Fpmoh
At DC 20% Prism F/R was 2,696,Conv Op.Time(/0.2) 13,483Fpmoh
At DC 30% Prism F/R was 2,700,Conv Op.Time(/0.3) 9,000Fpmoh
At DC 40% Prism F/R was 2,703,Conv Op.Time(/0.4) 6,759Fpmoh
At DC 50% Prism F/R was 2,707,Conv Op.Time(/0.5) 5,414Fpmoh
At DC 60% Prism F/R was 2,710,Conv Op.Time(/0.6) 4,51Fpmoh
At DC 70% Prism F/R was 2,714,Conv Op.Time(/0.7) 3,877Fpmoh
At DC 80% Prism F/R was 2,717,Conv Op.Time(/0.8) 3,397Fpmoh
At DC 90% Prism F/R was2,7212,Conv Op.Time(/0.9) 3,023Fpmoh
At DC 90% Prism F/R was 2,724,Conv Op.Time(/1) 2,724Fpmoh

(The prediction at 4%Dc is the result of my analyis at real operating condition of my item). By only changing the DC and the CR the variation is little and does not reflect the large increasing of working time. By dividing for the decimal value of the DC, obviously, i obtain figures that decrease as the DC increases. In my understanding i expected to have an almost constant value. Surely my way of proceeding is not correct. Can you address the correct mode to convert the Calendar FR to Operating F/R when i have different DC? and can you tell me where my approach fails? I thank you in advance for your replay


Subject: Converting a PRISM analysis from calendar hours to operating hours
Reply Posted by: David Dylis ( )
Organization: System Reliability Center
Date Posted: Mon Mar 5 16:56:23 US/Eastern 2007
Traditionally, reliability prediction failure rate models have been based on the operating time of the part, and the values were typically stated in failures per million (or billion) operating hours. The PRISM component models (and the empirical data contained in the databases used in the PRISM tool) predict the failure rate in units of failures per million calendar hours. This is necessary because it represents the common basis for all the failure rate contribution terms used in the PRISM methodology (e.g., operating, nonoperating, cycling, and induced). If an equivalent operating failure rate is desired (in units of failures per million operating hours) steps need to be taken to make this conversion.

To convert a PRISM analysis from Calendar to Operating hours the resulting PRISM predicted failure rate (based on the expected duty cycle) should be divided by the duty cycle in its decimal form (e.g., A 50% duty cycle = 0.5). It would be incorrect to assume a 100% duty cycle for the system (unless the system actually operates at 100% duty cycle) as factors that influence reliability, such as cycling and non-operation, would be not be properly addressed.

This discussion was also published in the following: //

For more information on PRISM feel free to contact the PRISM team by phone (315-337-0900), by E-mail ( or through the PRISM Forum (//

Reply to this message