PDA

View Full Version : Lognormal distribution with Weibull++ Ver. 6


probsmth
October 24th, 2006, 12:03 AM
When calculating a lognormal distribution, generally the basis e is used. Is there any chance to user another basis, e.g. 10?

We have customers, who insist on using lg insted of ln. This leads to other values for the log mean and log standard deviation.

Tarik El-Azzouzi
October 24th, 2006, 10:24 AM
The lognormal distribution states that the ln of times follow a normal distribution. So what you can do is: you can transform your original times to failure using the log function, enter the transformed data into the data folio and fit the normal distribution. However, remember to transform your time inputs also if you are using the QCP. For example, if you want to calculate R(100 hours), you should enter log(100) as the time value in the QCP time input.