How can I prepare my discrete batches of data for training?

I'm trying to calculate effect of parameters of an operation on the thickness of a wall.

Each operation is thinning the wall thickness and at some point the wall is replaced and operation starts again.

My operational parameters are changing daily and are collected on daily basis. However, the thickness info is measured at the end of the operation approx. after 60 days. Therefore, for each thickness value I have 60 rows of parameter data.

I'm new to Machine-Learning. I've watched various lectures and tried to google my case, but haven't come up this kind of examples.

How should I prepare my dataset for handling this kind of discrete batch data training?

I think just adding a column for batch number won't solve my problem since there is no relation between the thickness and the batch number.

Is there a way to sequentially train my model? Train for first batch, then update/improve the already trained model with the next batch of data and then the next one..

Or is there a way to handle this case quickly by just simply adding a column (i.e. batch no) to my dataset and specify the situation while training the model?

I'm trying to use Amazon SageMaker with Data Wrangler. But, i can use Python and Jupyter Notebooks..