Adjusting Axios Learning Periods for Intermittent Intervals

Change learning period schedule based on device usage.

"Is the learning period fixed for the axios sensor? Can it be adjusted? I have a case where the fan is shut down for weekends and not taking readings therefore the amount of readings that contribute to the average levels is less.

So is there a way for extend the learning period say based on total number of readings to get a "better: average?

One possible application that we are thinking runs 18 hours a day and 6 hours a day shut down. Do the shut down readings affect the average or are they ignored?"

August 11, 2023 12:14

Hi there, thank you for reaching out!

SKF Axios does not have a fixed time for the learning period, and the machine learning model was
developed with cases like the example you shared in mind. The model is ready after a pre-determined
amount of "acceptable" measurements. Readings with vibration levels below .5mm/s, or .02ips, are not
considered "acceptable" measurements - so they will not be used for the model.

The model is generally ready within 2-3 weeks, depending how long it takes to collect acceptable measurements.

Comments