Application of Six Sigma to product failure

Six Sigma – iSixSigma Forums Old Forums General Application of Six Sigma to product failure

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
  • #27392

    Jason Edwards

    Our management have requested investigation into the viability of adopting a six sigma approach to maintenance of existing wood power poles. From what i can gleem from your site, the six sigma approach applies to a manufacturing process (and control thereof) rather than to management of asset failure.

    Could you please advise if the six sigma approach can be applied as a failure target eg 3.4 failures per million assets per year, and if so, how this target is justified. (present standards specify 99.99% = 100 failures per million assets per year)



    Hi Jason
    I am not too sure what is the nature of your product or service. I guess I need more clarity to answer your query better. Nonetheless, I would like you to know that Six Sigma can and has been used in improving quality of products like jet engines to refigerators to pagers to medical life saving devices to etc etc etc and in so far as services are concerened in the IT industry in the Call centre industry in the Financial Services sector name a few.
    What I want you to appreciate is that the Tools and Techniques used in Six Sigma are not specfic to any Industry but are applicable all across, which is why in my opinion it is a highly potent tool for Problem Solving
    Regards and Have a great day



    The three factors I’ve found unique to the Six Sigma philosophy are

    1. Although statisticaly based, upper management only must comprehebd the numbers 0-6 to see progress or lack thereof.

    2. Calculation of a long-term shift in performance to insure not only capability but sustainability, thereby validating the process not requiring specific chaos manager be assigned to maintain performance.

    3. Training of mangement as well as operations to accept statistcal results instead of opinions.



    Hi Jason,

    The simple answer to your question is yes, Six Sigma can be applied to Asset Management or any other business practice of a “Transactional” nature.

    The concept is to express Defect Opportunities in terms of “DPMO” = defects per million opportunities, which can be referenced to a “Sigma” level. Six Sigma is virutal perfection (3.4 defects/million) which is useful because it sets the bar much higher than typically expected from many processes (Eg; 2%defects) Six Sigma also “levels the field” by equating all comparisons on the same basis, DPMO.

    In your case 100 defects / million represents a Sigma of 5.2 and a yield of 99.99%, as you stated. For most processes this would be an excellent performance, assuming the process is in control and results are repeatable.

    On occasion a given process may exceed six sigma, although that’s not a common occurrance. However, the same manufacturer very likely will have other processes in his businss running in the range of 3 Sigma, such was the subject of a presentation given by Motorola at a seminar I attended last year.


    Gary Cone

    This is strictly a financial decision. If the cost of doing nothing (present failure rate of 100 per million) is less than the cost of improvement, do nothing. If the present cost is significant, Six Sigma can help. It is straight forward. Analyze all failures, do a pareto of causes, and go from there. All processes can be improved, your comapny just needs to decide if this improvement is a critical business issue.



    Thank you for your reply.

    Our concern lies in determining the ‘number of oportunities’ for failure.

    The exact scenario is this.

    We have approximately 500,000 wood poles in service. From time to time these break. Regulations specify a reliability rate of 99.99% PA = 50 failures PA.

    But this is a yearly rate.

    When we came to compare our failure rate the the six sigma model, it became apparent that the six sigma model was based on each item haveing only one opportunity to fail (at time of measurement of drift). If this is the case we need to meet approx 1 pole failure per year! 50 x reduction!

    Our poles may fail this year, next year or the year after, providing many opportunities to fail. Does this mean that each pole has one opportunity of failure per year of it’s life (eg 60 years), because the unit of measure we are using is failures PER YEAR?

    Another line of thought here was that all poles are inspected every 5 years on a rotational basis. Is this event the opportunity to fail?


    Gary Cone

    The thing you need to understand is the distribution the failed poles are following. Do you know or do you have time to failure data?


    Jason Edwards

    We have diltiled three years of failure data into a normalised histogram showing ‘Number of failures per 10,000 poles in service’ with respect to a 10 year age category. The results are

    0-10 years old = 0 failures per 10,000 poles
    10-20 yrs = 0
    20-30 yrs = 0.1755
    30-40 yrs = 2.1317
    40-50 yrs = 4.0326
    50-60 yrs = 7.1712

    The average time to failure is 44.1 years, and the average failures per year is 26.3.

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

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