Hello all. I'm new to JMP and have been reviewing many of the use cases and webinars. I'm looking for a strategy on how I can approach optimizing the process conditions of a batch manufacturing process. There are plenty of resources about the multivariate analysis I would like to do, but I haven't found how to approach my goal because of the nature of the process. I have two data sources I'm dealing with:
I have seen videos of multivariate methods to study the effect of one continuous process condition on another (temperature vs. pressure), but I'm not sure how to set up my data so that I can profile a process condition and study its effect on the final quality. The quality (or response) data is not continuous, but it can be tied to the continuous data because the Batch IDs exist on both spreadsheets.
Any ideas on how to approach this? Thanks!
You should look into the Functional Data Explorer (FDE) in JMP Pro. Your hourly batch data generates a batch profile for each of those measures. FDE will characterize that batch profile and then allow you to relate it to the single output quality value.
I'm not really doing this approach justice with my simple explanation, but it is something that you should look into.
Thanks Dan. Do you know if there are any other approaches that are possible with JMP "standard"?
There are other approaches, but honestly, none of them are nearly as good as using the Functional Data Explorer (FDE).
The issue is how do you characterize the entire profile of a batch into a single observation to match with the quality measure. Options that I have tried in the past:
* Identify a particular time point in the profile that is "representative" of the entire profile.
* Use the maximum/minimum/median/75th percentile/25th percentile/trimmed mean/etc.. to represent the entire profile.
* Use a model form to model the profile. Then use the parameters from that model form to represent the profile.
* Use principle components to summarize the profile and then use the principle component scores to represent the profile.
These are just some ideas that I have used. I REALLY wish FDE was available back when I was encountering these problems on a regular basis.
Here are a few of my thoughts:
1. You will need to have the x's (variables during the processing of a batch) and the Y's (response variables characterizing the batch) in the SAME data file.
2. While you suggest there is one sample of the batch that is tested (or that you have a quantitative response for), is that sufficient? Have you considered variation due to the measurement systems in the lab? How is the sample obtained? With one sample from each batch it will be impossible to speared within batch variation from between batch variation. You should develop sampling plans based on hypotheses regarding the effect of x's and sample to understand those relationships.
3. Can you create intermediate response variables? Y's that may be correlated to the Y's measured at the lab, but are measured upstream from the lab? For example, maybe the Y in the lab is molecular weight and a measure of the batch could be viscosity.
4. You might not need all of the continuous data as the batch is being made. Again start with hypotheses as to what relationships you think are of interest. Sample the stream of data to select data of interest to examine relationships in JMP. For example, you might hypothesize that the rate of temperature change impacts the Y. Then you can determine the slope of the line for the rate of temp change in making the batch and use that statistic to investigate the relationship with the Y (using fit model in JMP).
5. Using Analyze>Multivariate Methods>Multivariate look for relations ships between the x's. Reduce the number of x's to include in the data set based on which x's are correlated.