Volume 8, Issue 1

Reliability Edge Home

New Feature in Weibull++ for Automotive
Warranty Analysis Using Mileage Data

In an article entitled "Analysis of Automotive Warranty Data in the Mileage Domain" from the last issue of Reliability Edge, Dustin Aldridge of Delphi Corporation discussed the usefulness of considering usage (in this case, mileage) in addition to time in service when analyzing automotive warranty data. He presented an example of an analysis that used a statistical distribution to estimate the mileage for non-failed units that are still eligible under the warranty agreement (i.e. suspensions) based on the mileage readings for failed units returned under warranty. The analysis involved the extensive use of spreadsheet functions to calculate a conditional probability of failure for each month and then applying the probabilities to the population being analyzed. This approach provided some important benefits to make the analysis more realistic but it was also fairly labor-intensive and time-consuming to perform. A recent Service Release to ReliaSoft’s Weibull++ software provides a new feature in the Warranty Analysis module that can accomplish similar objectives much more quickly and easily. This article presents a very brief introduction to this new capability (using a fictional data set for illustration purposes).

After selecting the new "Usage Format" when creating a new Warranty Analysis folio in Weibull++, the user simply enters the number of units put into service for each time period. Then, when entering the available information regarding warranty returns, the user specifies the usage accumulated by each unit at the time of its return along with the corresponding date when the unit was put into service. The next step is to specify a distribution and parameters that describe the usage for the component. This could be based on the analysis of data from a previous warranty period, from early data in the current warranty period, etc. Figure 1 displays a possible usage distribution for an automotive component. Figure 2 shows the returns data entered into the Weibull++ Warranty Analysis folio along with the usage distribution defined in the Control Panel on the right side of the window.

Lognormal Ppf Usage
Figure 1: Usage distribution for the component

Warranty Returns
	Data
Figure 2: Warranty Analysis folio with Returns data
and usage distribution settings displayed

When the Warranty Analysis folio performs the calculation for data entered in this format, it will generate a full data set containing the usage for each of the observed failures and the estimated usage for each suspended data point (where the estimates are based on the warranty data and usage distribution provided). A failure distribution can then be fitted to the observed returns data and the estimated suspension data, and a warranty forecast can be performed. Figure 3 shows the forecasts generated for this example. Other useful information can also be obtained from this analysis.

Warranty Forecasts
Figure 3: Warranty forecasts based on the actual usage data
for the observed failures and the estimated usage data for the suspensions

If you are interested in more information on this analysis approach, the concepts underlying this new feature are also presented in more detail in a recent Reliability HotWire article, which is available on the weibull.com website at http://www.weibull.com/hotwire/issue73/hottopics73.htm. End Article

 

ReliaSoft.com Footer

Copyright © HBM Prenscia Inc. All Rights Reserved.
Privacy Statement | Terms of Use | Site Map | Contact | About Us

Like ReliaSoft on Facebook  Follow ReliaSoft on Twitter  Connect with ReliaSoft on LinkedIn  Follow ReliaSoft on Google+  Watch ReliaSoft videos on YouTube