Platform · Analytics
Set up once, use on every machine.
A measured value, a calculated figure, a machine state and an AI model's result all work the same way in Toleron: they can be charted, recorded and used for the next evaluation. That is why new evaluations add up instead of piling up special cases.
Schematic
- Wrong setups
- Refused while building
- Released versions
- Never change
- New machine
- Assign, don't program
- New evaluations
- Test phase before going live
No plausible-looking wrong numbers
Some calculations look correct but are not — for example when very fast sensor data is fed into a step that expects one value per interval. The result is not an error message but a believable wrong number. Toleron refuses such a setup while it is being built and says what is missing.
A running evaluation never changes underneath you
Every released calculation step stays fixed. An improvement becomes a new version; what is running keeps working exactly as before.
Set up once, reuse on every machine
An evaluation is described once in machine-neutral terms — for example “the spindle load”. For each machine you assign once which of its values that is, with live values shown while you assign.
After that, every dashboard, report and evaluation works for that machine. A new controller brand needs an assignment, not programming.
Tested before it counts
- Preview runs a new evaluation on your real history while you are still editing it. Gaps in the data stay visible as gaps.
- Recalculating the past can be paused and resumed, and what was already calculated is kept.
- Going live follows fixed steps: a new evaluation first runs in a test phase in the background, without affecting your dashboards, and can only go live after an observation period. Rolling back is always possible.
- Nothing in use disappears by accident: Toleron refuses to delete a value that a running evaluation still produces.
Production figures on the same basis
Toleron derives production figures the same way on every machine: machine state (from the controller or from a retrofitted sensor), shift calendars with breaks, night shifts and holidays, part counting with traceable corrections, job detection from program changes, downtime with automatic rules and bulk manual classification, and OEE recalculated every minute.
Go deeper
Related
Production & Process Analytics
OEE, utilisation, downtime, parts and jobs per shift — with an honest gap where data is missing, never an invented number.
Learn moreProcess Anomaly Detection
Spots unusual cuts early — starting with simple, explainable limits and moving to trained models only when your data supports it.
Learn moreTool & Process Monitoring
Evaluates every single cut, so tool and process decisions rest on what really happened in the material.
Learn moreNext step
Bring intelligence to your machines.
Connect existing CNC equipment without replacing the machine or the controller. We start with one machine, real data and an honest acceptance run.