Help on analysing industrial data


I'm a little bit new at Knime, and I'm working with it on a industrial application.

I've structured the data, and have an csv file with three columns: DateTime, Set Point and Process Value. My goal is to compute how many times within an hour the Process Value crosses the Set Point, as illustrated in the attached pdf.

Is there some node that does this work? How could I elaborate a workflow to do this analysis?

Thanks in advance.

It seems your data is neatly sorted by the logging time. That makes it quite easy.

First you need to create a new column and set a flag in that if a transition is made from above to below or the other way around. You can do that witha  java snippet, see this thread for a very short sample of something similar:
I suggest using a 1 for a transition and a 0 for no transition.

Note that this is needed as you have to compare each row with its predecessor.


Then you can extract the date+hours from the timestamp (discarding the minutes etc) so that you have 60 rows with the same date-hour timestamp.


After that you can group the rows taking that date-hour timestamp as a grouping value, and summing the flag.




Thank you Ellert!