Capability Analysis for Packaging Seal Strength with spec. >0.1 Kgf using Minitab

[aufa_fadhli]

Hi All,

I am very appreciate if you can help me. I have tried to learn the related article but still don't understand.

I have project to validate blister pack machine. In this case, I have collected 30 samples for each setting parameter (each pack will have 4 location of seal strength to be measured). The spec. only one sided: >0.10 Kgf/15mm. So, the total measurement is 120.

My question is how to perform capability analysis using minitab (especially minitab 17)?

What to do if my measurement (data) is not normal?

for better understanding please find the attachment.

Thank You So Much
Aufa

 

[Miner]

re: Capability Analysis for Packaging Seal Strength with spec. >0.1 Kgf using Minitab

We would be happy to assist. Can you clarify a few things about your data set?

1. Sealing Time. Did you deliberately change sealing time as part of an experiment, or are these random variations?
2. What do you mean by "Shot"? Are the blister packs molded? Are these consecutive shots?
3. Could you attach a diagram of the positions of the cavities and the locations on each cavity?

It is likely that you have multiple process streams that would have a big impact on how the data are analyzed, so the answers to these questions are important.

 

[aufa_fadhli]

Hi Miner,

Thanks for your response,

1. Yes, I deliberately changed the sealing time, to find the optimum seal strength we used "trial and error" method to get the best result. As you saw on my data, we challenge sealing time vs sealing temperature. for the pressure we set to the maximum setting (set as constant parameter).
2. The blister pack are molded one. In one mold, it has 7 cavities. So, in 1 shot (cycle) we will have 7 pack after cut.
3. Please find the attachment for detail of seal strength location and 7 cavities, sorry if I can't give proper image since the machine is not running now.

Thanks in advance
Aufa

 

[Miner]

This was an extremely detailed data set that can be analyzed a multitude of ways. However, you are interested in the capability analysis. Since you performed a designed experiment in addition to the capability, you will first have to determine which combination of sealing temperature and sealing time is optimum for you (e.g., 140 seal temp and 4.0 seal time).

Focus on the data for this combination. First, you should verify that you do not have distinct process streams with different means. The variation across cavities is similar enough that they may safely be combined. However, there is a significant difference between test location C and the other test locations (A, B & D). A, B & D are similar, but C has a higher mean. There is an opportunity to reduce variation by bringing these means closer together.

Physical test results such as strength are frequently right skewed. You have several options available to you in Minitab 17. You can perform a nonnormal capability analysis using the Largest Extreme Value distribution, or perform a normal capability analysis and using the Box-Cox transformation. Lastly, you can perform a nonparametric capability analysis using the ECAPA macro.

All three approaches will give you a Cpk ~ 1.5 for the 140 seal temp and 4.0 seal time data.

 

[aufa_fadhli]

Hi Miner,

Thank a lot for your explanation. According to your explanation you have done some analysis regarding my data. However, I have some questions,

1. You said, "However, there is a significant difference between test location C and the other test locations (A, B & D). A, B & D are similar, but C has a higher mean. There is an opportunity to reduce variation by bringing these means closer together.". what statistical tool you used to say that C has significant different? one way or two way ANOVA?

2. You said, "You can perform a nonnormal capability analysis using the Largest Extreme Value distribution". How you determined to use largest extreme value distribution during capability analysis?. I already know from the normality test my data is non-normal. However, honestly I still don't understand how to choose the correct "fit distribution" on minitab 17. Could you explain to me?

3. Since I have 4 location of seal strength with sample size 30ea. Hence total measurement is 120ea. How to arrange the data in minitab for capability analysis? do I only need to arrange all the data in one row and distinct the data based combination of seal temp. and seal time?

Very appreciate for your assistance

Thanks in advance
Aufa

 

[Miner]

1. I used a 1-way ANOVA with paired comparisons.
2. I used Stat > Quality Tools > Individual Distribution Identification. You pick the best fits by examining the graphs, and looking for lower Anderson-Darling (AD) statistics. There were a number of reasonable fits. I chose the Largest Extreme Value because it is used to model the distribution of maximum results such as the peak water heights of rivers during floods. A tensile strength is the peak value of a tensile curve, so it will often fit this distribution.
3. Arrange the measurements in a single column for a given Sealing Temperature/Sealing Time. I recommend that you add another column to indicate the test Location and use that to indicate the subgroups.

 

[aufa_fadhli]

Hi Miner,

Thanks a lot for your complete explanation.
I will try to do what you advise to me.

Aufa