I understand you are interested in a component of variation study across multiple steps in a manufacturing process. I don't understand your data table? You will need to describe your data table. What are all of the column labels? Are these different dimensions of the product? It appears you have a repeated pattern of 28 rows. Is the first 28 the dimensions before grinding and the 2nd after grinding and the third after plating? You need a column indicating this (perhaps there is one in Chinese?) How were the 28 gotten? In time series order? What variables were changing during the grabbing of the 28? How representative are the 28 of all variation in each process step? What is not changing in your study (Inference space)? Is measurement error estimated somewhere or is it confounded?
In order to assess stability and quantify the components of variation at each step in your process, you will likely use control charts. Range charts assess stability and X-bar charts compare sources of variation. (You might also use ANOIVA to quantify variance components if each component is stable.) How effective the control charts will be at assessing stability and quantifying the variation of each step is a function of how representative your sampling is of the variation in the process. For example, let's say your 28 parts in your study are from one raw material lot. It would be impossible to assess stability over multiple raw material lots.
To visualize components of variation, you can use Graph Builder and/or variability charts.
I don't understand this: "Batch statistical analysis is used instead of single-dimensional analysis"?
"All models are wrong, some are useful" G.E.P. Box
