iSixSigma

Charting Durability Data

Six Sigma – iSixSigma Forums Old Forums General Charting Durability Data

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • #38727

    RG
    Participant

    Is there a best method for charting durability data?  We are trying to determine if a product durability issue is actually increasing or not.  The problem is that the product(s) that have been returned may have been in the field for a few months.  Also the units shipped are increasing over time, so we can’t get a true reading if the product defect is increasing or not.  The factors that need to be considered are units returned for each month, units shipped each month, and length of time in the field for each unit returned.  We thought of using a 3 month rolling average to plot the data.  Such as plotting the 3 mth rolling average of units returned for the defect per the 3 mth rolling average of units shipped.  Any thoughts?

    0
    #116429

    Bob J
    Participant

    RG,
    If you have a date code you might be able to evaluate the Time To Failure trending by looking at how long the product was in the field before it was returned…   
    Just a thought….
    Best Regards,
    Bob J 

    0
    #116436

    Thai
    Participant

    If you are dealing with a significant number of parts shipped per month, you may want to look at failure rates at different months in service (or weeks depending on the MTF).  By plotting the failure rate at different months in service (i.e. 1MIS, 2MIS, 3MIS, etc) against the production month, your data is normalized to different monthly production levels.  You are also able to see if the failures are due to infant mortality or a wear-out function.  This type of durability measure is widely used in the automotive industry to track warranty issues.

    0
    #116465

    Bob J
    Participant

    Kirk,
    Are you a reliability guy?
    Best Regards,
    Bob J

    0
    #116469

    Thai
    Participant

    Not really, just one of those jack of all trades, master of none.  I deal alot with warranty reduction projects so I have had to pick up on reliability statistics along the way…
    Kirk

    0
    #116471

    Bob J
    Participant

    Kirk,
    No problem…  Seems like you have a good working understanding of how best to approach product reliability tracking and assessments…. 
    Good stuff…
    Best Regards,
    Bob J

    0
    #116755

    Allard Munters
    Participant

    I had similar issues some time ago and I used Minitab’s “Reliability/Survival Analysis” successfully. You could split the data in groups attributed to, for example, production month, or shipment month, and then calculate the MTTFs of the individual groups. This tool actually analyses their distributions (often Weibull) and test if the distributions are equal or not. If the distributions are equal (same shape&scale) then the data is actually coming from the same distribution, so there is no significant difference. Another very nice feature of the tool is that you can do “censoring”, a technique that allows you to not only include data of defective units (units that have actually died), but also of units that have not yet failed. Minitab’s helpfile and the user manual give a good explanation of the tool, including clear examples.

    0
    #116767

    Patrick McDermott
    Participant

    Here is an article that provides a method to do the type of assessment you need:
    http://www.weibull.com/hotwire/issue49/hottopics49.htm
    The methodology is consistent with Allard’s response.

    0
Viewing 8 posts - 1 through 8 (of 8 total)

The forum ‘General’ is closed to new topics and replies.