(The B10 life is equivalent to the time at which reliability reaches 90%.) Then you enter 90 for the reliability and 80 for the confidence level. To specify the metric that you wish to demonstrate, you choose Reliability value at a specific time from the Metric drop-down list on the RDT sheet. Since you are able to make reasonable assumptions about the life distribution that describes bulb A’s failure rate behavior, you select the Parametric Binomial test design method on the control panel of the folio. In the Test Design Assistant window, you select Reliability Demonstration Test Design. To create a new test design folio, you choose Home > Insert > Test Design. Compare the test times that would be required given different sample sizes and different numbers.Determine the test time for a zero-failure test that would demonstrate the target metric with a sample.(In some special cases, the 1-parameter Weibull or 1-parameter exponential distribution can be suitable for analyzing a data set containing few or no failures.) Further, if our estimate of the failure rate behavior is not accurate, we will have no way of knowing based on the results of a 0 failure test. Note, however, that zero-failure tests generally cannot be used to determine the actual value of a product’s life metric. This sort of test provides a quick and efficient way of accomplishing this. Zero-Failure Tests: A zero-failure test is appropriate in this case because you only want to know, with a given confidence level, whether the life metric is greater than the specified requirement. In other words, you decide to design a test that uses the available sample size of 10 and will demonstrate the target metric if no failures occur before a certain time (i.e., you design a test where the number of "allowable" failures is zero). You are allocated 10 bulbs for the test and decide on a zero-failure reliability demonstration test. More specifically, your objective is to demonstrate that bulb A has a B10 life of at least 500 hours with 80% confidence. This is equivalent to 90% reliability at 500 hours. Reliability Demonstration Test Designīased on your past experience with bulb A, which we plan to use in an upcoming projector model, you are asked to design a demonstration/validation test to show that the bulb has the required B10 life of 500 hours. In this scenario, we have already analyzed some test data from bulb A, and the analysis shows that the failure behavior of bulb A follows a 2-parameter Weibull distribution with beta = 4.22. If there is limited or no information available, we can try several such guesses and choose the most conservative resultant test plan. This can come from field data for similar components, a test conducted on a prototype or even an educated guess. In order to design a reliability demonstration test, we generally need an estimate of the failure rate behavior. You'll start by using the Reliability Demonstration Test tool to design a zero-failure test intended to demonstrate a specified reliability for a projector bulb titled "Bulb A." Afterwards, you’ll use the Expected Failure Times Plot to compare the failure times that are expected from a design with a specified reliability to the actual failure times that are observed during a test. In this example, you will work with two tools available in the Test Design folio. Weibull++ includes several test design tools that provide ways to design reliability tests and evaluate/compare proposed test designs.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |