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
Posted by: Johnson Wood
Date posted: Tue Apr 3 3:12:21 US/Eastern 2001
Subject: Failure rate prediction
There are a great deal of products operating in the field entironment. Recently, some failures were occured and caused some losses for our company. The main cause is some types of component failed, such as hard disks, memories, power modules. Now, the company leaders wonder:
1. How long will the remainder products that not failed to this day last?
2. and if there will be a moment that most products will fail?
By these, we can make some preparatives before they fail, such as preventive replacement. Any advices will be welcome.
Subject: Failure Rate Prediction
Reply Posted by: Patrick Hetherington
Date Posted: Tue Apr 3 6:30:18 US/Eastern 2001
Data on the devices you mention have estimated mean time between failure values ranging from 70,000 hours for disk drives, 225,000 hours for memory boards to 300,000 hours for power supplies. To determine if schedule replacement is appropriate, a wear out pattern should be verified. To perform this, time to failure information should be collected for a sample and a Weibull distribution analysis should be perform. A detailed description of this analysis method can be found in the Weibull Handbook by Dr. Robert Abernethy. Further information on this test can be found at http://rac.alionscience.com/rac/jsp/webproducts/products.jsp?detail=WHDK The results of this analysis will render information that can easily be used to determine scheduled replacement intervals and risk to unfailed units described in the text.
Even with the methods the risk of failure may be intolerable. Prior to performing this analysis, it is recommended that the cost and severity of an outage caused by failure of these or other devices be understood. Redundancy or periodic backup may be more effective and certain than scheduled removal.