Skip to main content
Solved

Fingerprint covering parts of a continuous process with variable duration

  • June 5, 2026
  • 3 replies
  • 17 views
yracette
Polestar
Forum|alt.badge.img+2

One of our continuous processes has a “lift” step with a flow that follows a typical pattern with a spike and a ramp down. The lifts are a relatively short duration ~20 minutes and therefore happen several times a day

I created a fingerprint from several of the most representative lifts using layers resulting from a similarity search .

Question: because lifts happen at irregular time intervals, how can I properly align the fingerprint to coincide with the beginning of a given lift on a TrendHub view? Do I need to have some kind of trigger?

Best answer by Jef Vanlaer

Hi Yves,

I assume you’re referring to a trigger for a deviation monitor on the fingerprint? Ideally, you have a clear trigger search that identifies the start of the “lift” step (e.g., a value based search on a stage tag, or a discrete 0/1 tag that becomes 1 during the lift). If you use the same search to select the layers that form the fingerprint, you get the best alignment.

If such tags are not available, these are some alternatives:

  • Use the Similarity Search as a trigger (potentially with a lower threshold). This has some disadvantages though, as lift steps that are too dissimilar from the reference one will not trigger the monitor, and the monitor will only run around the end of the lift step as that’s when the Similarity Search will yield a new result.
  • Construct a Value Based Search that identifies the start of the lift step (e.g., delta aggregation on flow > x for 2 minutes) to be used as a trigger (and ideally also to select the fingerprint layers to guarantee good alignment).

Let us know if any of these options work for you...

Yves Racette

3 replies

Jef Vanlaer
Community Manager
Forum|alt.badge.img
  • Jef Vanlaer
  • Community Manager
  • Answer
  • June 5, 2026

Hi Yves,

I assume you’re referring to a trigger for a deviation monitor on the fingerprint? Ideally, you have a clear trigger search that identifies the start of the “lift” step (e.g., a value based search on a stage tag, or a discrete 0/1 tag that becomes 1 during the lift). If you use the same search to select the layers that form the fingerprint, you get the best alignment.

If such tags are not available, these are some alternatives:

  • Use the Similarity Search as a trigger (potentially with a lower threshold). This has some disadvantages though, as lift steps that are too dissimilar from the reference one will not trigger the monitor, and the monitor will only run around the end of the lift step as that’s when the Similarity Search will yield a new result.
  • Construct a Value Based Search that identifies the start of the lift step (e.g., delta aggregation on flow > x for 2 minutes) to be used as a trigger (and ideally also to select the fingerprint layers to guarantee good alignment).

Let us know if any of these options work for you...

Jef Vanlaer
Community Manager
Forum|alt.badge.img
  • Jef Vanlaer
  • Community Manager
  • June 5, 2026

If you just want to visually align the fingerprint with a new step on a view, the trigger search also comes in handy: use the search to find the latest (or another) lift step, add it as the base layer, and then add the fingerprint at start of the view.

yracette
Polestar
Forum|alt.badge.img+2
  • yracette
  • Author
  • June 5, 2026

Jef, 

Thank you very much for the quick response and clear instructions. I do not access to a discrete signal that would change state at the beginning of a lift, but I will create a trigger based on a rate of change of the flow tag as you suggest.

Thanks again

Yves 

Yves Racette

    Need help?

    Can't find the answer to your question? Create a new topic or file a support ticket.

    New to TrendMiner?

    Discover our Industrial Analytics platform at www.trendminer.com.

    Terms & ConditionsAccessibility statement