SPRT

[Chadi Cbridi]

[Originally Posted: 10/11/00--Transferred by ReliaSoft Moderator]

I Downloaded the SPRT program, however I don't understand exactely the consept of the sequential Probability weibull test Thanks

Chadi Cbridi@bu.edu

[RS Tech Support]

[Originally Posted: 11/11/00--Transferred by ReliaSoft Moderator]

The SPRT utility comes with a help file to aid you. In addition, you can check out sequential testing in "Reliability and Life Testing Handbook" Vol 2, by Dimitri Kececioglu, 1994 in Chapters 7-10.

[wzy]

I have the same questions. Can anyone help send me a copy of the contents mentioned in the RS TECH SUPPORT message? It is just difficult for me to find such a book. By the way I don't think the help file is clear enough for those who are not very familiar with the sequential testing plan. It looks like it is something like Weibayes testing plan, but apparently is not the same.

Maybe I should ask this way, in the example illustrated in the help file, what's the meaning of 'Run', why one item has been tested for 685hr, but the decision is 'continuing testing', whileas one item has been only tested for 682hr and the decision is 'reject hypothesis'? What is the benefits of runing a sequential testing instead of a routin testing?

I appreciate if anyone can help me out.


Emails: wangzhy@guomai.sh.cn

[wzy]

Something useful to clarify my questions posted yesteray. One can visit the following website for details:
http://www.samplingplans.com/glossary.htm#Sequential Analysis

wzy

[Discussion Moderator (Moderator)]

The sequential probability ratio test (SPRT) is a method of testing for reliability that minimizes the number of test units and test equipment needed to demonstrate a required reliability. However, the minimization of test units and resources comes at the cost of test time; sequential tests can sometimes take a great deal of time before a decision is reached as to whether the units have met the required reliability (acceptance) or are below the minimum reliability (rejection). For this reason, sequential testing may be more appropriate for evaluating design reliability than for ongoing assessment of quality or performance at the end of the manufacturing line.

Based on the parameters input by the user, a "test envelope" is generated (represented by the red and green lines in the ReliaSoft SPRT tool). Once that has been established, testing can begin, one unit at a time. Once the single unit on test has failed, the information is entered into the SPRT tool, where it is compared to the values of the "test envelope." If the trace of the value for the cumulative failures vs. cumulative test time (blue line in the SPRT tool) crosses the red line, the units have demonstrated a reliability equal to or greater than the required reliability. Likewise, if the trace crosses the green line, that indicates that the units have failed to meet the minimum required reliability. If the trace is still in the envelope, the results are inconclusive, and another unit must begin the test. Note that time values (y-axis) represent cumulative test time. That is why, in the provided example, the third failure time of 685 hr. results in a "continue testing" decision, while the fifth failure time of 682 hr. results in a "reject" decision. This is because at that point of the test, five failures have occurred in 2625 hr. (629+359+685+270+682=2625), and this is indicative that the units do not meet the minimum reliability requirements set up by the user.