Intervals - How does W++ use them?

[Steve]

W++ has "Last Inspected" and "State End Time" for interval data. Say I have been looking for failures every day, but want to group my data in intervals. If my Last Inspected is 0 and State End Time is 50 and I have 4 units that failed at exactly 50 days, would the 4 points go in the 0 to 50 interval or the 50 to 100 interval? In other words, does the "Last Inspected" or the "State End Time" INCLUDE that value or does it exclude it?

I hope this makes sense.

[Pantelis]

With respect to the last item in your question when you are dealing with an interval it does include the end points ( i.e. less or equal to 50). With regards to the intro to your question, I guess I am not sure why you would want to enter data that is not censored as censored … but I guess that was not part of your question J

[Steve]

"I am not sure why you would want to enter data that is not censored as censored .."

Actually, that leads to the next question: What is the maximum number of rows that W++ can handle? I'm sure the answer starts with "It depends....", but that is one reason for the question regarding intervals.

[Pantelis]

The spreadsheet on V7 can handle up to 64,000.

[Steve]

What are the limits of version 6?

[Arai.M]

V6 allows 16,384 rows.

[Steve]

There is no reason I can't use an interval set like this is there?

0 to 10
11 to 20
21 to 30
31 to 40
etc.

Most examples I have seen would have looked like this:

0 to 10
10 to 20
20 to 30

I just want to make sure I understand how W++ is managing the intervals.

[David]

Hi Steve,

Yes, Weibull++ can handle a data set like the one you described. Weibull++ does not require the start of the interval to be equal to the end time of the previous interval.

I hope this helps.

[AromaticsThailand]

Yes! I think can be use in Weibull.

If I would like to calcuate TTF for the system from above time cycle test in Excel, I must to calculate Mean in each cycle right?

For instance:

0 - 10 = 5
11 - 20 = 15

and etc.

[David]

You can use the midpoint of the interval as an estimate of the time-to-failure within each interval. But this is an assumed approximation. If you want to be conservative you could use the start time of the interval, or be optimistic by using the end time of each interval. Weibull++ does not make either of these assumptions when analyzing interval data. Specific information on the analysis of interval data within Weibull++ can be found at http://www.weibull.com/LifeDataWeb/analysis_parameter_methods.htm#suspended_data (under ReliaSoft's Alternate Rank Method (RRM)) and at http://www.weibull.com/LifeDataWeb/appendix_b_special_analysis_methods.htm#rrm.

I hope this helps.

David