Calculates a statistical Bi-Weibull distribution parameters from a data set. The calculated parameters are Beta > 0 the shape parameter and Eta > 0 the scale parameter.A third variable, r-squared is calculated in order to describe the goodness of fit. A parametrizided least squares linear regression is used to obtained the parameters mentioned above.
The following variables and steps should be completed in order to run results
Hit to add a new row
If you have a large set of data, the easiest way to import is to create an Excel file with 1 workbook and 2 columns like this:
The first column must have the text "Cycles" as its first value. Items in this column must be numerical positive values.
The first column must have the text "Censored" as its first value. Items in this column must be boolean values, either TRUE of FALSE.
Step on the row you want to edit until it turns yellow and update the field you want to edit. Then click on the checkmark to save changes.
This is the field where failure events must be added. You can use any units (hours, tonnage, cycles, days). Make all data have the same kind of units.
Check this field if the event is uncomplete (i.e. A record of a component that was replaced before failure)
Step on the row you want to remove until it turns yellow. Click the red bin to delete that row.
To clear all records and charts click Clear All
Click Run to obtain the output data which contains the density function charts, Weibull Parameters and goodness of fit.
The straight line in the CDF chart represents the Weibull Cumulative Distribution Function obtained with the Median Rank methodology for the data in the Inputs table.
This is a plot that represents the Weibull Probability Density Function with the parameters obtained after runing the tool
Running the tool output the following parameters
This value represents the shape parameter or median. In failure analysis ETA is better known as the Characteristic Life
This value represents the shape parameter.
When Beta is 0 B 1, failure behaviour is considered Infant Mortality
If Beta is B=1 (or very close to 1), failure behaviour is considered Random.
Finally, Beta>1, failure behaviour is considered Wear Out
This value represents the square of the sample correlation coefficient. The closer to 1 it is, the more the data fits a Weibull Distribution.