All Member Open Forum

 View Only
Expand all | Collapse all

Weibull analysis

  • 1.  Weibull analysis

    Posted 01-31-2019 08:46 AM
    I am looking for an overview presentation around Weibull concepts and also any tool we can start to use to do the analysis.

    ------------------------------
    Krishnan Shrikanth
    CMRP
    SINGAPORE
    ------------------------------


  • 2.  RE: Weibull analysis

    Posted 02-01-2019 06:55 AM
    Krishnan,
    Try Wes <g class="gr_ gr_29 gr-alert gr_spell gr_inline_cards gr_run_anim ContextualSpelling ins-del multiReplace" id="29" data-gr-id="29">Fultons</g> website Home of SuperSMITH software and Weibull analysis training from Fulton Findings LLC.

    I have attended Wes' <g class="gr_ gr_109 gr-alert gr_spell gr_inline_cards gr_run_anim ContextualSpelling multiReplace" id="109" data-gr-id="109">3 day</g> workshop and can highly recommend it.

    Weibullnews remove preview
    Home of SuperSMITH software and Weibull analysis training from Fulton Findings LLC.
    We are the Weibull Software Leaders, Wes Fulton and Dr. Bob Abernethy, home of SuperSMITH(TM) and The New Weibull Handbook(C)
    View this on Weibullnews >


    ------------------------------
    Stephen Cooper
    Asset Management Specialist
    GRU
    Orange Park, FL
    ------------------------------



  • 3.  RE: Weibull analysis

    Posted 02-03-2019 06:58 AM
    Dear Stephen Cooper,

    thanks for the very educative links and let me review and use as appropriate. more to follow,

    ------------------------------
    Krishnan Shrikanth
    CMRP
    SINGAPORE
    ------------------------------



  • 4.  RE: Weibull analysis

    Posted 02-01-2019 07:03 AM
    Try the link below to the package for Weibull in R. You can find many discussions and instructions for Weibull with R by googling Weibull in r. R is a free open source software available for download at https://r-project.org.





  • 5.  RE: Weibull analysis

    Posted 02-03-2019 06:50 AM
    thank you Richard for the references and let me review the same for more information,

    ------------------------------
    Krishnan Shrikanth
    CMRP
    SINGAPORE
    ------------------------------



  • 6.  RE: Weibull analysis

    Posted 02-03-2019 03:27 AM
    ​Shrikant,

    I wouldn't call them a presentation. However the following Sites might be of some help to you.

    1. Weibull.com
    Reliability Engineering, Reliability Theory and Reliability Data Analysis and Modeling Resources for Reliability Engineers

    2. Reliasoft.com
    Weibull++ - Reliability life data analysis
     
    I'd spent some time earlier on learning Weibull Analysis and performing some simple analysis (though out of touch now). Remember there was a trial option for the Software.

    Thanks

    ------------------------------
    Sundar Naranammalpuram P.
    Navi Mumbai
    Maharashtra
    India
    ------------------------------



  • 7.  RE: Weibull analysis

    Posted 02-03-2019 06:49 AM
    thanks Sundar - indeed i have attended a Reliasoft seminar in Singapore few years before where they introduced Reliasoft software capabilities for Weibull analysis. Let me review them and thank for the connect,

    ------------------------------
    Krishnan Shrikanth
    CMRP
    SINGAPORE
    ------------------------------



  • 8.  RE: Weibull analysis

    Posted 02-04-2019 08:06 AM
    What software are you using for the statistical analysis? Minitab?​

    ------------------------------
    Matthew Harris Utilities
    Mechanical Reliability Engineer
    Versoco
    Rome GA
    ------------------------------



  • 9.  RE: Weibull analysis

    Posted 02-04-2019 08:26 AM
    Here is something to consider.  We should evaluate if Weibull is a fit to the failure history or events. The book, "Event History Analysis with R," shows how to start with an empirical analysis of history and thence tests for fit with different functions such as Weibull. If there is a fit, then it is acceptable to use the function.

    Richard G. Lamb, PE, CPA, ICBB, MBA
    Houston, Texas
    Tel: 832-710-0755; Email: rchrd.lamb@gmail.com
    Educational website: https://analytics4strategy.com

    Authored books:
       
    Availability Engineering and Management for Manufacturing Plant Performance
        Maintenance Reinvented and Business Success: Everything is about business






  • 10.  RE: Weibull analysis

    Posted 04-09-2019 03:08 PM
    I just tried the WeibullR package in R. There are sorts of things possible. It seems to be based on Abernathy. This is what I got. Notice parameters and expected and limits table on left. Can see the script that generated the table a plot. If you've read Abernathy's book, all of the functions in the WeibullR package make sense to you. Now I'm going to mess around with it.

    Set up and run Weibull with r

    ------------------------------
    Richard Lamb
    Analytics4Strategy.com
    Houston TX
    832-710-0755
    ------------------------------



  • 11.  RE: Weibull analysis

    Posted 04-09-2019 03:20 PM
    Notice in the weibullR generated report the box in the middle--script. It is saved by R as a. txt file. Once built you can send the script to anyone, or make it available in a directory. Thence, pull the script into R on their computer and run it. Of course, they can adjust it for their own needs.

    ------------------------------
    Richard Lamb
    Analytics4Strategy.com
    Houston TX
    832-710-0755
    ------------------------------



  • 12.  RE: Weibull analysis

    Posted 04-18-2019 02:09 AM
    Dear Sirs,
    We are having the data on near misses and incidents happening at different locations in our facility over a period of time. Can Weibull analysis be used for predicting the probability of the most accident prone area? Then concentrated efforts can be put on those areas to reduce risk. Would be good if an example is shared.

    ------------------------------
    Srihari R ramasubramanian.srihari@gmail.com
    ------------------------------



  • 13.  RE: Weibull analysis

    Posted 04-18-2019 07:11 AM
    This reply does not answer your specific question on how to forecast the next event,  However, if you want to characterize the trend of these discrete events, consider the LaPlace Test.  Near the bottom of the website named "KSC Reliability" is a one-page write up.   I have used the score from the LaPlace Test to emulate likelihood of occurrence (along with the quantity and consequence of the past events) in a risk measure for recorded and undesired events (e.g., problem reports).  Tim Adams

    ------------------------------
    Tim Adams ASQ CRE
    Reliability & Risk, Senior Engineer
    Nasa Kennedy Space Center
    Kennedy Space Center FL
    ------------------------------



  • 14.  RE: Weibull analysis

    Posted 04-18-2019 07:17 AM
    Thank you! I will look up the resource.
    Regards, 
    R.Srihari






  • 15.  RE: Weibull analysis

    Posted 04-18-2019 07:44 AM

    Srihari,

    Take a look at the Crow-AMSAA method. Chapter 9 of Abernathy's book "The New Weibull Handbook" provides a good explanation with examples. There is the issue of life and the CA seems to step around it. Maybe most important to you, the method will allow you to look for change over time-indicated by points plotting to different lines over time. Essentially, you may be able to set it up with cumulative events versus cumulative operating time.

    Look at Karl Burnett's contribution and attachments in the string, "Has anyone measured maintenance effectiveness with Crow-AMSAA?"






  • 16.  RE: Weibull analysis

    Posted 04-19-2019 10:44 AM
    Hi Srihari,

    Dr. Larry Crow was a consultant for us that contributed to the development of our RGA software which has the Crow-AMSAA models available that Richard Lamb references below. We have free full use evaluation versions of the software that you could access to test out using your data, etc. You can find info on that here: https://www.reliasoft.com/products/reliability-analysis/rga 

    I'd be happy to connect you with one of our Application Engineers to discuss both RGA and Weibull methods in how they relate to the area of interest you note above.

    Best of luck going forward.

    Mark


    ------------------------------
    Mark Shanahan
    HBM Prenscia
    Tucson AZ
    ------------------------------