Statistic Tools to Know Capability of Sealing Machine for Seal Width

aufa_fadhli

Starting to get Involved
#1
We have a data of seal width with various sealing temp. as the variable parameter to be challenged. Meanwhile the sealing time and sealing pressure is in fix setting.
Since the seal width is not even, so we measured the seal width for the minimum and maximum of seal width.
According to this, what tools would you suggest to me to analyse the data so we will know the capability of the machine?
I have been thinking about using capability analysis but I am not sure.
Currently I am using minitab as statistical software analysis.

for detail please refer to attachment.

I am very appreciate for your help
 

Attachments

Miner

Forum Moderator
Staff member
Super Moderator
#2
You appear to have more than one objective here. You state that you want to determine the capability of the machine, but the data are from an experiment. Capability requires a stable process, but an experiment deliberately changes the process. We can recommend several ways to analyze the experimental portion of the data, but the capability portion requires collecting data from a stable process.

There are numerous ways to analyze the experimental data. I would recommend using various graphical methods such as multi-vari, scatter plots and boxplots so you can visualize the variation and potential sources. Then I would recommend more analytical approaches such as ANOVA and regression. You have a number of significant sources of variation (i.e., seal temperature, right/left side, within seal variation). I would recommend optimizing your seal temperature and determine how to minimize the right/left side and within seal variation first. Once you have done that, you can collect samples over time from a stable process for your capability.
 

aufa_fadhli

Starting to get Involved
#3
Hi Miner,

Thank you for your prompt reply,
This is an old sealing machine. Hence, the performance already not stable especially for the seal width result. Major improvement is not possible considering the machine usage for production running is very low.

We already know that this machine is not stable. However, what would you suggest? What kind of analysis should be provided that suitable for this condition? or just do non-parametric analysis using box plot or individual plot or histogram?

Thank You
Aufa
 

Miner

Forum Moderator
Staff member
Super Moderator
#4
This is an old sealing machine. Hence, the performance already not stable especially for the seal width result. Major improvement is not possible considering the machine usage for production running is very low.

We already know that this machine is not stable. However, what would you suggest? What kind of analysis should be provided that suitable for this condition? or just do non-parametric analysis using box plot or individual plot or histogram?
This is the problem. You have stated twice that the machine is not stable. Therefore, any estimate of capability, no matter what method used will be wrong. An unstable process is an unpredictable process. Its performance today will not be the same as its performance tomorrow.
 

Top