iSixSigma

Continuously drifting processes

Six Sigma – iSixSigma Forums Old Forums General Continuously drifting processes

Viewing 12 posts - 1 through 12 (of 12 total)
  • Author
    Posts
  • #31254

    melvin
    Participant

    I have a machining process that drifts continuously throughout the run.  My length grows as the cutters wear.
    What is the SIMPLEST method of control given that the operators are pressed for time and have limited mathematical abilities.  Would precontrol be an option?  My Cpk can be as high as I want depending on how often I want to make offsets, so I need an effective way to trigger the offset.

    0
    #82245

    Mikel
    Member

    The wear should be predictable. If so, use the lower 95% confidence on wear and institute a hard rule of offsets every x number of passes. Put a tool wear chart showing the starting point after offset and ending point with control limits based on data and sample immediately after offset and somewhere midway through the intervals. Using precontrol like thinking would be adequate, such as sample two consecutive pieces and plot each value vs. taking average and range. make the green +/- 1.5 s from the predicted wear line and so on.

    0
    #82265

    Chip Hewette
    Participant

    Control charts don’t have to have a single X double bar value.  Control can be monitored as the average moves continuously lower and lower.  The key is to understand the “expected” shift downwards, and alert the operator when the individual value or subgroup average exceeds the control limits based on the expected trend downward.
    Understanding the likely causal factors of (a)raw material metallurgy, (b)cutter metallurgy, (c)cutter geometry, (d)part rotational speed, (e)cutter feed, and (f)phase of the moon (ha!) may be of the most benefit.  One or more of these factors may cause a sudden shift away from the expected length dimension.
    If you monitor and control these factors, you may be able to simply put a rule in place to change the cutter every xxx parts, or every so many hours.  If you put a tool resharpening program in place (or insert changeover plan), you can create a kanban (container) of ready to go cutting tools so that the line operator is never waiting for tooling.
    It would be helpful to have a simple cutting tool length gage on hand, so that the required offset can be read off of a dial indicator.  If you make the gage so that it references one side of the length spec, you may be able to create all positive offset values.  Then, the work instruction would read “Cut xxx parts.  Remove cutting tool.  Obtain new cutting tool.  Gage.  Key in offset value in line 1245 of program Z.  Load tool into holder.  Make five parts.  Gage part for length.  Confirm offset.  Run xxx parts.”

    0
    #82267

    Benjamin
    Participant

    Seems like Wheeler addressed this issue in one of the columns he used to write for Quality Digest (?). Seems like he used slanting control limits. You might try that web site.
     

    0
    #82292

    melvin
    Participant

    The difficulty here is how do I estimate sigma?  The more frequently I make offsets, the lower my sigma and vice versa.

    0
    #82293

    melvin
    Participant

    I searched qualitydigest.com for Wheeler and got quite a few articles but not the one you refer to.  Do you know the title or a keyword to help refine my search or even an approximate date of the article???

    0
    #82296

    Mike Carnell
    Participant

    Bob,
    There are two sigmas in this process. You are looking to reduce the overall process sigma. The control will come from the sigma at each setting. As you make your adjust ments (driven by tool wear) is it just the mean shifting or is the within group standard deviation deteriorating as well. In most instances it is actually both. Gather some data and run a test of equal variances to find out what you are dealing with.
    Good luck.

    0
    #82297

    Loehr
    Member

    Hi Bob,
    I also was unable to find the article by Wheeler, but I came across another source of information on charting a process with tool wear.  It also explains how to estimate process capability for such a process. 
    The technique is explained on pages 759-769 of the book, Measuring Process Capability, by David Bothe.  I got my copy through ASQ’s Quality Press, but I have also seen it on Amazon.com.
    Hope it helps.

    0
    #82298

    Benjamin
    Participant

    The article was entitled “Can I Having Sloping Limits?” and appears in Quality Magazines’s May 1999 edition.
    Good luck.
     

    0
    #82323

    Summerfield
    Participant

    What abot using a EWMA chart

    0
    #82350

    Gabriel
    Participant

    Bob, I don’t usually deal with drifting processes, but after reading your questions and some answers I have a few ideas that might work. I am sending to iSixSigma a few charts to attach here to better explain the idea (see attached below).As Mike said, you have two variations: Part to part variation (that variation that you would still have if the process didn’t drift) and the variation due to the drift cause of wear.You said “the Cpk can be as high as I want, deppending on how often do I make offsets to correct wear”. From that frase, it seams that the part to part variation is really very very small compared with the drift, and also that the part to part variation is nearly negligible compared with the specifiaction.If the process had no part to part variation at all (what of course is just a mental experiment) you could let the length grow until it reaches the USL and then make the offset to take it to the LSL. That would give you a rectangular process distribution. But because you have at least some measurement variation, and because you are not measuring every part as to catch the one that reaches exactly the USL, and because the offsett can not be perfect so as to get exacly the LSL in the first part after the offset, you will probably want to put some “safety margin” arround the limits to make that rectangular process distribution somehow narrower than the specification.But, you do have part to part variation. And you need to understand how large it is and also if it is the same just after the offset and after the wear when you are about to make a new offset.An idea (please, someone corrects me if it woul not work) would be to plot the lenght in function of the number of part after the offset and then you will get a scatter diagram. Then you will need to find a curve to fit the scatter (may be a stright line?). That curve will be the wear model. If the points fit very close to that curve then you have little part to part variation. Then you can calculate the MSE (mean square error) of the point arround that curve (the residual) and it will give you an estimation of the average part to part variance. Now you still need to know if the part to part variation is constant along the wear. So you can make a second plot of the contribution of each point to the MSE in function of the number of point. If this second scatter can be adjusted by an horizontal line of value MSE (as calculated in the first scatter chart), then the part to part variation is constant.If not, you will need to find a model for the variation of the part to part variation. You can make this with the data of the scatter plot (not sure how) or, if the wear is negligible for let’s say 100 consecutive parts, you can take samples just after the offset, at the end of the wear process just before the new offset. A linear approximation with this two points should be enough unless the variation changes in a very non linear way, which seems unlikely to me. Once you have a model of how the average changes along the wear and how the process variation changes, you have your process model complete and you can define your “offset limits” (i.e. the length at which you will make the offset and the set-up value for that offset).And, even better, if you get enough data along enough time as to make a good process model and verify that it allways behaves the same, you can use it to predict to reduce the level of inspection. You can begin to measure only when the process is about to reach the point where the offset has to be made, and may be one day you can make the offset on a “hard time” basis (i.e every n parts), instead of on a measurement basis. Download: Drifting Process 1 [Gif graphic file] Download: Drifting Process 2 [Gif graphic file] Download: Drifting Process 3 [Gif graphic file] Download: Drifting Process 4 [Gif graphic file]Viewing Tip: Usually, you can click on a link to view the document — it may open within your browser using the application (in this case Microsoft PowerPoint). If you are having difficulty, try right clicking the link and selecting “Save Target As…” or “Save As…” to save it to your computer harddrive.

    0
    #82379

    zhou
    Participant

    Gabriel,
    pretty good idea, and well illustrated.
    it’s very helpful to me as i’m encountering the same problem
    thanks

    0
Viewing 12 posts - 1 through 12 (of 12 total)

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