B
I've been assigned to turn around a sheet metal forming shop. Most the the equipment is rather old (+25 years in service) and a lot of the work is performed by tribal knowledge.
We have many pieces of equipment (Shears, back presses amd turrets) that are know trying to cut and form sheet metal blanks and assemblies to a +/- 0.015 inch tolerance.
What I need to do is quickly determin if the equipment at hand is capable to repeatability produce product at the specification needed.
My problem it is very very difficult to break into the production and run 30-50 test panels to perform capability tests.
My question, since we are running extensive components with the +/- 0.015 requirements, can I just measure the deviation from nominal to everything thats being produced and use the deviation from nominal to perform the capability studies?
One issue, is the data is not normal in distribution. Second, I would have both positive and negative numbers to workd with in respect fro niminal.
Looking for advice, comments, suggestions and a general opinion if this is statistically valid approach.
Please help, I am under a tight deadline and I will be very greatful for any and all constructive advice.
We have many pieces of equipment (Shears, back presses amd turrets) that are know trying to cut and form sheet metal blanks and assemblies to a +/- 0.015 inch tolerance.
What I need to do is quickly determin if the equipment at hand is capable to repeatability produce product at the specification needed.
My problem it is very very difficult to break into the production and run 30-50 test panels to perform capability tests.
My question, since we are running extensive components with the +/- 0.015 requirements, can I just measure the deviation from nominal to everything thats being produced and use the deviation from nominal to perform the capability studies?
One issue, is the data is not normal in distribution. Second, I would have both positive and negative numbers to workd with in respect fro niminal.
Looking for advice, comments, suggestions and a general opinion if this is statistically valid approach.
Please help, I am under a tight deadline and I will be very greatful for any and all constructive advice.
