PRISM Forum - Message Replies


Topic: PRISM Questions and Answers
Topic Posted by: SRC (src_forum@alionscience.com )
Organization: SRC
Date Posted: Wed Jan 12 8:33:33 US/Eastern 2000
Topic Description: Welcome to the PRISM forum! Please feel free to post your questions and comments about the PRISM assessment software here.

Back to message list Show all replies About this forum

Original Message:

Posted by: Super Wong (superwang@tyanchina.com )
Organization:TYAN
Date posted: Sun May 28 23:11:03 US/Eastern 2006
Subject: About User Defination
Message:
About component failure rate: if we use some special component, I know that we can fill failure rate in "User Defination",but when we can only get information as FITs and so on from vendor, how can we transform these information to failure rate.In other words, does PRISM only provide "failure rate" one item for Component User Defination? Best regards Super


Reply:

Subject: User Defined failure rates in PRISM
Reply Posted by: David Dylis (ddylis@alionscience.com )
Organization: System Reliability Center (SRC)
Date Posted: Tue May 30 8:56:14 US/Eastern 2006
Message:
PRISM includes component level prediction models for several different part categories. When a component model does not exist a user may select a failure rate from the extensive failure rate database included within PRISM or insert a user defined failure rate. When considering a user defined failure rate, the environment and user profile for this data should be similar to the system that is being analyzed. In addition, PRISM includes a user-maintainable component library that can capture specific part information to promote ease in building up a system reliability prediction. Part of this library includes provisions to record component test or field data that is subsequently used to optimize the specific component base failure rates. As an example, the predicted component failure rates calculated by the PRISM component models yield industry average values. Failure rates more representative of individual applications can be obtained by integrating empirical data into the prediction. Empirical data typically accounts for specific variables that can impact reliability, but cannot reasonably be included and accounted for in generic models.

The PRISM user is encouraged to collect as much reliability test and field data as possible to support the prediction. This includes the use of zero failure data, as "no failures over a given time period" is equally as important to the determination of the estimated failure rate of an item as data where failures are experienced. The use of field data at the system and assembly level is addressed by "Predecessor System/Assembly Analysis". The PRISM program also uses test and field data at the component level to account for real-life experience data to enhance predicted component failure rates. This is done mathematically by combining the predicted component failure rate with empirical data. The technique is used for this purpose. This technique accounts for the quantity of relevant data by weighting large datasets more heavily than small datasets.

All failure rates in PRISM are in failures per million calendar hours. If you have a failure rate in FITS (failures per 10^9 hours) then this failure rate would need to be converted to failures per million calendar hours.

If you have any additional questions, feel free to contact me.


Reply to this message