Understanding your data

Treon Connect turns raw sensor signals into decisions about your machines. This section explains what the platform measures, what it means, and how it decides something is wrong — so the numbers and charts in the applications actually tell you something.

From signal to decision

1. Measure. A sensor on a machine samples vibration and temperature and computes KPIs — compact numbers like velocity RMS that summarise the machine's state.
2. Trend. Those KPIs are stored over time. A rising trend is often the earliest sign of a developing fault.
3. Detect. Treon Connect flags problems two ways:
   • AI anomaly detection learns each machine's own normal behaviour and flags deviations — no thresholds to set. This is a core part of what Treon Connect does.
   • ISO / threshold rules compare KPIs against standard or custom limits.
4. Act. Detections become events your team investigates — using spectra and waveforms to diagnose the cause — and resolves.

Where to go deeper

• Overview of solutions — Treon Connect, Treon Make and Treon Flow, and what differs between them.
• The KPIs explained — velocity RMS, acceleration RMS, kurtosis and the rest.
• AI anomaly detection — how the models learn and what an AI event means.
• Reading spectra and waveforms — diagnosing the cause of a problem.
• ISO 20816 severity classes — where threshold limits come from.
• RPM – rotation speed explained — why machine speed matters, and how a Treon Industrial Node X measures it.