In production—Program — · —
Part —Expected batch end: —
Next—
Industrial example
A CNC machining centre, part after part.
A single machine connected with Link: states, cycle times, tool life and spindle vibration, OEE, scrap and energy per part. It works for machines that are a few years old too. Simulated data; one second is 15 seconds. Try breaking a tool.
With AI machine learning reads vibration and power draw, predicts tool wear and spindle failures, and suggests when to step in.
Loading the digital twin…
Spindle—
Tool—
Control—
State—
Coolant—
Finished parts—
CNC 3 · — · simulated data · —
Shift OEE—
Availability —Performance —Quality —
Good pieces——
Cycle time——
Scrap——
Energy per part——
Current cycle —
- Spindle speed
- —
- Feed rate
- —
- Spindle load
- —
- Override
- 100 %
State
——
The state comes from the control over OPC UA; on machines without OPC UA it comes from an input module read over Modbus TCP.
Stops this shift
—
Tools · life used and expected change
Each tool’s life is counted in parts by the control. The end mill has a sister tool in the magazine: at end of life the control picks it on its own, without stopping the machine.
Spindle · condition-based maintenance
- Vibration
- —
- Temperature
- —
—
Cycle times · latest parts
—
Energy · by machine state, today
Now—
Today—
Per part—
—
Events
Devices in Link
- CNC control · OPC UA
- Spindle vibration sensor · Modbus TCP
- Energy meter · Modbus TCP
- Coolant concentration · Modbus TCP
On machines without OPC UA, Link reads the stack light, part counter and alarms from a Modbus TCP input module.
Order from the ERP
ERP——
- Confirmations sent
- —
- Machine hours on the batch
- —
Orders, batches and confirmations go through the Studio REST API: good parts, scrap and machine hours reach the ERP without paperwork.
Anomalies · Datathesys Cloud
Ø12 H7 contouring: for three days it has taken 6 % longer with the same program, and coolant concentration has dropped from 7.2 to 5.9 %. Coolant is probably too diluted: the tool runs hotter and wears faster.
Studio sends the data to Datathesys Cloud, which compares batches and days and flags unusual behaviour.
Expected plan
Batches from the ERP, operator shifts and breaks, tools at today’s wear rate; unplanned stops not included.
Forecast
What the Industrial vertical does
Older machines too: OPC UA where available, a Modbus TCP input module where not. States and part counts arrive either way.
States, stops and OEE per machine and per shift, with stop causes and the Andon page.
Tools and cycle times: life in parts, expected change, cycle time part by part against the standard.
Condition-based maintenance: spindle vibration and temperature, with the recommended action before the failure.
Quality and energy: scrap by cause, kWh per part, per batch and per shift.
ERP/MES integration: incoming orders and outgoing production reports with Studio's REST API.
Got machines to connect? Let's talk.