All Member Open Forum

 View Only

  • 1.  New Product Reliability

    Posted 04-13-2023 04:11 PM

    Hello all, I'm looking for advice on best practices for developing reliability models for new a new product. The product is a large industrial machine with various novel/proprietary components that don't have any historical data to mine for reliability. In this scenario, how does one predict the reliability of the overall machine without historical reliability data.



    ------------------------------
    Hartley Simpkins PEng
    Service Engineering Lead
    Svante
    North Vancouver BC
    ------------------------------


  • 2.  RE: New Product Reliability

    Posted 04-13-2023 04:48 PM
    Edited by Torbjorn Idhammar 04-13-2023 04:48 PM

    Hartley,

    when you say "reliability model", what are you trying to accomplish?  Predict the reliability of this machine calculated in %?  If so, what will you use the data for?

    Best Regards,



    ------------------------------
    Torbjorn Idhammar
    President & CEO
    IDCON, Inc.
    http://www.idcon.com
    Raleigh NC
    ------------------------------



  • 3.  RE: New Product Reliability

    Posted 04-13-2023 06:38 PM

    Hi Torbjorn,
    Yes the goal would be to calculate reliability as a %. Our KPI's are currently based on uptime percent. Downtime and total maintenance time is easier to determine (grease this often, change oil this often, expected part life, etc.).

    The actual reliability seems more difficult to determine with something that we've developed and don't have a lot of data on. How do you determine this without a lot of test articles?

    A statement that we can achieve 95% uptime with 90% reliability (for example) would be the end goal.

    Thanks in advance.



    ------------------------------
    Hartley Simpkins PEng
    Service Engineering Lead
    Svante
    North Vancouver BC
    ------------------------------

    Original Message


  • 4.  RE: New Product Reliability

    Posted 04-14-2023 09:11 AM

    Back in the day we would construct Reliability Block Diagrams of all new equipment coming into the system. Each major component or black box that made up the equipment was a block unto itself. We would go as far as considering software and firmware as blocks as well. Then, armed with this foundational model, we could apply initial reliability expectations. Even with new equipment, we can assume that the reliability will be less than 1.0 for each block. The RBD model also accommodates parallel or redundant situations that can have dramatic impact on the overall Reliability of the equipment. We would start by setting each block with a Reliability of 0.999 (given we know nothing about the device yet). Now the model will calculate for us a reasonable expected reliability right out of the box. Then, as we learn more over time about the equipment, we would update the reliability number for the blocks. RBDs are a very powerful tool. 



    ------------------------------
    John Todd
    Total Resource Management
    Alexandria VA
    ------------------------------



  • 5.  RE: New Product Reliability

    Posted 04-14-2023 09:41 AM
    Hartley - 
     
    Though your equipment may be new/novel, I think your best approach would be to gather a team of veterans from engineering, operations, and maintenance who have worked with similar equipment and discuss with them problems/failure modes that may related to this equipment, and consider potential new ones that may arise because of the specific novelty.  Make them part of the team, perhaps part time only, and take their advice as you can.  Your model will evolve from this.  I hope this helps.  
     
    Ron Moore
     
     





  • 6.  RE: New Product Reliability

    Posted 04-14-2023 11:29 AM

    Hi John and Ron,
    Thanks for these suggestions. I think the RBD would be a very good starting point that can be continually adapted as knowledge is gained. We have done a FMEA with a very large, diverse group of engineers, operators, etc. for the products. I can leverage this info for the RBD. Might be a very grey to start but should develop more as we learn.

    Would love to hear other ideas as well.



    ------------------------------
    Hartley Simpkins PEng
    Service Engineering Lead
    Svante
    North Vancouver BC
    ------------------------------

    Original Message


  • 7.  RE: New Product Reliability

    Posted 04-14-2023 11:41 AM
    Edited by Torbjorn Idhammar 04-14-2023 11:41 AM

    A very simple variable to keep in mind is the total number of components.  The more components..... typically, the less reliable.  It is crude, but if you have similar (not the same) equipment, similar environment, similar amount of components, it may give you a hint.  



    ------------------------------
    Torbjorn Idhammar
    President & CEO
    IDCON, Inc.
    http://www.idcon.com
    Raleigh NC
    ------------------------------



Related Content