This article outlines what the learning period is and why it is important.
How can it help you?
A thorough and consistent learning period for each piece of your equipment is required for the FitMachine AI, which uses machine learning to tell you the condition of your equipment. Without a proper learning period, the effectiveness of the system's Condition Alarms is significantly reduced.
How does it work?
The learning period takes place once a FitMachine sensor has been installed and onboarded. During this period, 5000 running samples are collected for the FitMachine AI to learn the regular patterns of the equipment.
These samples are used to build models that the FitMachine AI can compare with new data. The samples provide the basis for the Condition Alarm system, such as the equipment baseline temperature and the '‘normal'’ vibration boundaries for equipment operation. Essentially, the learning period gives our AI an understanding of your equipment's ''normal'' running state so that, as the equipment runs, it can detect variances from this ''normal'' state.
How long does the learning period go for?
How long the learning period takes depends on how often the equipment runs. Typically, a piece of equipment that runs 24 hours/day will take two weeks to complete learning. It will take longer to gather the required running samples for equipment that has less running time than this.
Where can I find an equipment's learning period progress?
The learning percentage is displayed on the List View (see below).
This percentage increases based on when the digital models are built and not how many samples have been taken.
It increases in jumps rather than linearly. If the FitMachine sensor is gathering running samples then the learning will progress.
Condition Alarms coming from the system are less likely during the learning period. Before 33% learned the system won't send condition alarms, between 33% and 67% the system will send yellow advisory alarms only and after 67% the system will send red alerts.
What else should I look at?
---
Thanks for reading. As always, if you have any questions or concerns please reach out to MOVUS Support here.