How it works

What it simulates

A vision inspection cell on a bottling line, measuring fill level and closure from photographs of real bottles. One device, 15 tags, over Modbus TCP, with frames published to DataHub over NATS.

The measurement is made from the image rather than read from a sensor, so what travels north is a number, and the picture only if you ask for it.

• fill level and headspace measured per bottle
• the inspection verdict as a tag like any other reading
• frames into DataHub when NATS is configured

What it configures on Litmus Edge

On startup, and it removes exactly what it created when you are finished:

• 1 DeviceHub device on the Modbus TCP driver, 15 tags
• 1 Digital Twin model and instance for the cell
• 4 multi-input Analytics groups: fill anomaly, filler drift forecast, bottle images, headspace inference

Scenarios you can inject

• An underfilled or overfilled bottle is caught per bottle from the picture, against the same specification a checkweigher would use
• A missing closure is a separate verdict from a fill defect, because the two are different failures of different machines
• Filler drift turns a run of measurements into a forecast, so the useful number is not the bottle that failed but the trend that says the filler will