Reliability HotWire

Issue 124, June 2011

Tool Tips

*In Weibull++, why doesn't the calculated maximum likelihood estimator (MLE) line track the data points on my probability plot? 

When the parameters have been calculated using rank regression analysis, probability plots are often used to evaluate the fit of the distribution to the data. The plotting positions of the data points are determined by the failures/suspensions in the data set (x-axis) and the unreliability estimates (y-axis). The regression line is then fitted to the data points on the plot in order to obtain the distribution parameters. Therefore, with rank regression analysis, the closer the points are to the line, the better the model fit.

In contrast, the maximum likelihood estimation (MLE) method uses a statistical approach to estimate the parameters. The solution of the line is obtained from the likelihood function, and not by the plotting positions of the data points. Therefore, probability plots should not be used to evaluate the fit of a distribution when using MLE. (For a description of the methods that can be used for MLE analyses, see http://www.weibull.com/hotwire/issue71/relbasics71.htm.)

Since many practitioners still wish to obtain a probability plot for analyses performed with MLE, the default option in Weibull++ is to show the line that was calculated based on the likelihood function together with the points that would have been calculated for a rank regression analysis. Therefore, it is not a cause for concern if the points do not closely track the line because the line and the data points were obtained using different analysis methods.

As an alternative, you have the option to adjust the y-axis plotting positions so that the points are plotted directly on the MLE line (i.e., at the position where the failure time intersects the MLE line). To do this, choose File > User Setup. In the User Setup window, click the Plot tab and then select the Adjust MLE points on Plot Line option, as shown next.

User Setup Plot Tab

*How can I manage the records in the Xfmea 5/RCM++ 5 System Hierarchy and FMEA Hierarchy?

Xfmea 5 and RCM++ 5 include several options that make it easy to manage the records in the System Hierarchy and FMEA Hierarchy:

Drag and drop: You can move records within the System Hierarchy or the FMEA Hierarchy by dragging them to the desired location. When a record is moved, all associated analyses, attachments and lower level records will also be moved. In the System Hierarchy, the moved item will be added to the same level as the item that you drag it onto (e.g., a component dragged onto a subsystem will become subsystem), and it will be inserted above that item. In the FMEA Hierarchy you can move records to other places in the FMEA analysis at the same level (e.g., you can move a Cause to be with other Cause records, but you cannot move a Cause record to make it an Effect record).

Multi-select: There are two ways to select multiple records within the System Hierarchy or FMEA Hierarchy. You can click the first item and then press CTRL and click additional items. You can also click the first item and then press SHIFT to select all items between and including the two. You can then use the standard commands for Cut, Copy or Delete to perform the action simultaneously for all selected records. You also have the ability to drag and drop multiple selected records, if desired.

Paste Special: After cutting or copying items in the System Hierarchy, you can use the Paste Special window (Edit > Paste Special) to specify how you want the items to be pasted.

Paste Special Window

You can paste the items at the system level, at the same level as the currently selected item or at the level below the currently selected item. You can also choose whether or not to include lower level items and/or analyses associated with the items.