Attribute Data
Six Sigma – iSixSigma › Forums › Old Forums › General › Attribute Data
- This topic has 2 replies, 3 voices, and was last updated 15 years, 11 months ago by
k.bhadrayya.
-
AuthorPosts
-
June 27, 2006 at 3:13 pm #43852
ian LockParticipant@ian-LockInclude @ian-Lock in your post and this person will
be notified via email.Can anyone help me. I am recording the number of blow holes in an automated weld process. There are 4 idnetical machines so I have a measure of number of blow holes per machine / shift. I have constructed box plots to look at the spread but am a little stuck as to where we go next. Which hypothesis method should i be using.
0June 27, 2006 at 3:23 pm #139660Ian:We had a similar welding project and found that we had to be really clear on the operational definition of a ‘defect’. Once you have tackled that, then make sure that the number of defects is gathered in a standardized way that means something. In other words, take into account differing amounts of downtime and feed rates for each machine otherwise I could decrease the number of blow holes by shutting the machine down for an hour or so.In order to show the significance of the differences, you can do a Chi-square/proportion test on the number of defects per unit whatever.You will likely get a number of posts suggesting control charting the process. Depending on whether the feedrates are constant and identical, you should consider one of the four types of attribute control charts to monitor the process.Cheers, BTDT
0June 29, 2006 at 5:58 pm #139793
k.bhadrayyaParticipant@k.bhadrayyaInclude @k.bhadrayya in your post and this person will
be notified via email.Dear Ian
First be clear about the purpose of measuring the number of blow holes / machine /shift of 4 identical machines. Is the purpose is to find out the best performing machine out of 4 mchines ? or the purpose is to reduce the number of blow holes for a better weld. with out clear objective application of hypothesis does not yield any thing. Based on the objective the techinique can be decided.
K.Bhadrayya
[email protected]0 -
AuthorPosts
The forum ‘General’ is closed to new topics and replies.