1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
Dismiss Notice
You must be a registered member in order to post messages and view/download attached files in this forum.
Click here to register.

Attribute GRR for Visual Inspection

Discussion in 'Gage R&R and MSA - Measurement Systems Analysis' started by Pongsakorn, Mar 26, 2022.

  1. Pongsakorn

    Pongsakorn Active Member

    Joined:
    Aug 10, 2015
    Messages:
    72
    Likes Received:
    16
    Trophy Points:
    7
    My organization does Attribute GRR for Visual Inspection by using photo of good and reject samples.
    The purpose to do so because it is difficult to maintain the sample to be in the original condition for future
    GRR since the sample will have oxidation, contamination etc.

    I am not sure if this practice is acceptable in view of customers and auditors.
    Please enlighten me.
     
  2. Andy Nichols

    Andy Nichols Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    5,086
    Likes Received:
    2,553
    Trophy Points:
    112
    Location:
    In the "Rust Belt"
    If you mean your Certification Auditor, then they have no say in the matter. They don't decide what is 'acceptable".

    As far as customers, have you discussed it with them? Is it effective for you to do it this way?
     
    Pongsakorn likes this.
  3. Jennifer Kirley

    Jennifer Kirley Moderator Staff Member

    Joined:
    Jul 31, 2015
    Messages:
    1,071
    Likes Received:
    722
    Trophy Points:
    112
    Location:
    USA
    I have been in and seen numerous examples where inspectors are trained to recognize similar attributes, varying from an average reading in ultrasonic oscilloscope displays to determine overall percentage of silver soldered joints in copper-nickel piping, to a sense of smell detecting "notes" in a perfume's aroma. I have never, however seen a GR&R done to record this and I can not imagine what it would look like. As an auditor I would not reasonably expect to see it; it just isn't a commonly expected method for sensory-recognized data.
     
    Last edited: Mar 27, 2022
    Pongsakorn likes this.
  4. Bev D

    Bev D Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    605
    Likes Received:
    663
    Trophy Points:
    92
    Location:
    Maine
    In my experience using photos for a MSA is NOT advisable. A 2D image cannot match the depth of field, subtleties of color variation, etc. and is simply not the same as the actual object. Part of any visual MSA is present the objects at the actual rate utilized in production under the same lighting and other environmental conditions. You should also perform the test with enough objects that the ratio of good/bad objects is the same as in production.

    I suppose that a photo could be used IF the nature of the defect is that it is very obvious, and does not involve color variations or other characteristics that can’t be captured in a photo with very high definition…You can use photos as examples or standards (with limited success based on the nature of the defects)

    remember that defects can be eliminated and new ones can occur.

    If you have a very low defect rate and your objects degrade rapidly I can understand the temptation to use photos. But I would recommend that you either store the defective objects in such a way that they don’t degrade. If your defect rate is not that low you should be able to select new objects in real time.
     
    Pongsakorn and Andy Nichols like this.
  5. Pongsakorn

    Pongsakorn Active Member

    Joined:
    Aug 10, 2015
    Messages:
    72
    Likes Received:
    16
    Trophy Points:
    7
    Thanks Andy.
    This is the thing that I am not sure and I am not confident to discuss with customer since it seems to be not proper way to do.
    I am trying to find out the way to protect the sample from such change of condition.
     
    Andy Nichols likes this.
  6. Pongsakorn

    Pongsakorn Active Member

    Joined:
    Aug 10, 2015
    Messages:
    72
    Likes Received:
    16
    Trophy Points:
    7
    Thanks Jenifer !!
     
  7. Pongsakorn

    Pongsakorn Active Member

    Joined:
    Aug 10, 2015
    Messages:
    72
    Likes Received:
    16
    Trophy Points:
    7
    Thanks Bev.
    I agree with you that this is not proper way to do.
    However, I am finding the way to prevent the samples from such degradation.
    By the way, one more question, if there are 30 defects that inspector need to examine and make decision, should there be samples of such 30 defects included in the set of GRR sample completely?
     
  8. Bev D

    Bev D Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    605
    Likes Received:
    663
    Trophy Points:
    92
    Location:
    Maine
    Are the 30 defects currently occurring? If they are than yes they should be included.
    If they are not actually occurring then a sample unit (even if they have oxidation or contamination) or a very high photo should be used for training and reminder purposes. If you include a defect that isn’t currently occurring the MSA mix of defects and good units will not reflect actual occurrence rates and serves no purpose as the inspector will not see the defect in regular production.
     
  9. Pongsakorn

    Pongsakorn Active Member

    Joined:
    Aug 10, 2015
    Messages:
    72
    Likes Received:
    16
    Trophy Points:
    7
    Thanks Bev for the valuable comment !!