WEBVTT

1
00:00:00.000 --> 00:00:07.480
This video takes one data set through a Weibull analysis in ReliaEn: from the failure ages in the worksheet to the finished report.

2
00:00:07.625 --> 00:00:10.445
The data come from a pump-seal test rig.

3
00:00:10.875 --> 00:00:26.505
In column A, type the 10 failure ages, pressing Enter after each one: 42, 66, 87, 112, 128, 155, 176, 210, 240 and 280.

4
00:00:26.958 --> 00:00:35.138
In column B, type the ages of the 4 seals that did not fail: 50, 200, 300 and 300.

5
00:00:35.138 --> 00:00:38.898
These ages are censored: the seal was still working when it was taken off.

6
00:00:39.333 --> 00:00:47.203
So 14 seals were tested: 10 failed during the test, and 4 were still sealing when they came off the rig.

7
00:00:47.625 --> 00:00:49.405
Open the menu Life Data.

8
00:00:49.875 --> 00:00:51.605
Then Time Dependence.

9
00:00:52.083 --> 00:00:54.843
And choose Weibull (2P) distribution…

10
00:00:55.292 --> 00:00:59.662
The form Weibull Analysis (MLE) opens in the left panel.

11
00:00:59.662 --> 00:01:02.822
MLE stands for maximum likelihood estimation.

12
00:01:03.167 --> 00:01:08.477
Data situation stays at Exact times, because each row is one seal.

13
00:01:08.583 --> 00:01:11.573
Failure times column already shows column A.

14
00:01:12.042 --> 00:01:14.382
Censored times is set to column B.

15
00:01:14.833 --> 00:01:17.603
Confidence % stays at 95.

16
00:01:18.000 --> 00:01:20.020
Press OK — Calculate.

17
00:01:20.458 --> 00:01:23.418
ReliaEn fits the model and opens the report.

18
00:01:23.542 --> 00:01:32.622
The INPUT DATA card shows what ReliaEn used: 14 units on test, 10 failures and 4 survivors, fitted by Maximum likelihood (right-censored).

19
00:01:32.750 --> 00:01:44.830
In RESULTS, the shape β is 1.8157, with a 95% interval from 1.08449 to 3.03994.

20
00:01:44.958 --> 00:01:51.678
The whole interval lies above 1, so these seals wear out: the risk of failure grows with age.

21
00:01:51.792 --> 00:01:59.732
The characteristic life η is 223.278: by that age, 63.2% of the seals have failed.

22
00:01:59.875 --> 00:02:05.395
The mean time to failure, MTTF, is 198.477.

23
00:02:05.500 --> 00:02:12.160
B10, the age by which 10% of the seals fail, is 64.6524.

24
00:02:12.292 --> 00:02:16.742
The median life, B50, is 182.465.

25
00:02:17.208 --> 00:02:22.038
The Weibull probability plot shows the 10 failures against the fitted line.

26
00:02:22.038 --> 00:02:27.328
The survivors have no point of their own, but they change the order numbers of the failures after them.

27
00:02:27.792 --> 00:02:30.832
Below it are the reliability and hazard plots.

28
00:02:30.958 --> 00:02:35.698
The fit check reads Weibull NOT rejected — points follow the line.

29
00:02:35.698 --> 00:02:45.618
The correlation r is 0.9976, above the critical value of 0.9095, so the plot is straight enough to trust the Weibull model.

30
00:02:45.750 --> 00:02:54.360
The method comparison is not shown, and ReliaEn says why: the comparison methods need complete data, and here 4 seals survived.

31
00:02:54.833 --> 00:03:02.823
SERVICE INTERVALS gives the operating time that keeps reliability at or above each target, from 95% down to 5%.

32
00:03:02.958 --> 00:03:09.888
To keep 90% reliability, replace the seal before an age of 64.7, the B10 life.

33
00:03:10.333 --> 00:03:18.773
The PERCENTILE TABLE — B-LIVES WITH BOUNDS gives the age by which 1% to 99% of the seals fail, each with its 95% bounds.

34
00:03:18.917 --> 00:03:24.977
B10 is 64.65, between 31.25 and 133.8.

35
00:03:25.417 --> 00:03:30.437
That is the whole route: the ages in the worksheet, one form, one report.

36
00:03:30.437 --> 00:03:34.397
The next video shows censored data and how to choose the distribution.