Digital shadow
We read your equipment data via OPC UA, MQTT or log files. We do not touch the controller.
The data flows into BigQuery, ready for analysis, shift comparison and AI.
Digital twin and digital shadow
Digital twin for your production
A digital twin is a digital image of your equipment that continuously receives its data. datadice builds it from controller, camera and machine log data.
A local AI detects rejects and tool wear right at the line.
Goal:Understand and control your equipment
Last reviewed: · reviewed by Thomas Knorr, Managing Director datadice
Stamping line 3Reading data
PLC · OPC UACamera · local AIEnergy meter · MQTT
Shift OEE
69 %
Reject rate
0.9 %
Regrind in about 1,400 strokes
AI suggestion: regrind at the next stop
Example with made-up values
At a glance
- What
- datadice builds a digital image of your equipment from the data your machines already produce. A local AI analyses images and logs inside the plant.
- For whom
- Your line keeps stopping. Only the shift log knows why.
- Result
- A live image of your line with OEE per VDMA 66412, downtime analysis, shift comparison, traceability and AI forecasts.
- Start
- Initial meeting, then a strategy workshop with a data source check on one line
Funding:Small companies in Bavaria receive a grant of up to 50% from the Digitalbonus for the introduction. We prepare the application with you free of charge.
Go to the DigitalbonusInteractive demo
Try the digital twin of a stamping line
The demo runs entirely in your browser. It shows a stamping line with press, conveyor and AI camera. All values are sample values.
The demo loads only when you scroll here. It sends no data.
- Switch between digital shadow and digital twin
- Local AI inspects every part at the line
- AI answers questions about downtime with log lines
- OEE waterfall, downtime by cause and energy per good part
- Timeline with shift comparison
- Traceability for every reject
- Scenario for a different stroke rate
- Data sources with OPC UA, MQTT and BigQuery
- Alerts to the shift lead with approval
- 2D view for smartphones
Try the digital twin of a stamping lineOpen in new tab
Definition
What is a digital twin?
A digital twin is a digital image of real equipment that continuously receives its data and returns set points to the controller.
Without the path back to the controller, research calls it a digital shadow. In everyday use both are called digital twin.
In a Bitkom survey from May 2026, 57 % of companies with 100 or more employees say they lack the data needed for digital twins. That is why datadice starts with the data that controllers, cameras and machine logs already deliver.
Typical situation
Your machines produce data every second. Controller, camera and energy meter store it separately or not at all. Downtime sits in the shift log, rejects in a spreadsheet. In a Bitkom survey from May 2026, 57 % of companies with 100 or more employees say they lack the data needed for digital twins.
Our answer
datadice starts with a digital shadow of your equipment
A digital shadow only reads data from the equipment. It does not touch the controller. You see downtime, rejects and wear right away. We add the return channel of a digital twin once your team trusts the alerts.
We believe the value of a digital twin lies in the data of your equipment. That is why we start by reading the data your equipment already has. The AI runs inside your plant, so your images and logs stay with you. A person approves every change to the equipment.
For your IT lead
Does datadice change our controller?
No, the digital shadow reads data via OPC UA or MQTT and writes nothing back.
datadice only builds a return channel to the digital twin if you ask for it. Your shift lead then approves every change to the equipment.
The AI runs on an edge server in your plant. Images and logs stay there. datadice stores KPIs and history in your Google Cloud in an EU data centre.
For your management
What do we see after the first project?
For each line you see how much time downtime, slow speed and rejects cost.
OEE per VDMA 66412 splits losses into availability, performance and quality. The shift comparison shows which shift loses time where.
Energy per good part shows what each stop costs. You decide with numbers which line needs attention next.
Architecture
How datadice builds your digital twin
We start with one line and one question. Once the image runs reliably, we extend it.
Step 1 of 4
01
Goal
Together we choose one line and one KPI, such as rejects or downtime.
Example: a stamping line with camera inspectionStep 1 of 4
Goal
Line: Stamping line 3
Question: Why do rejects rise on the night shift?
Outcome
What you get
You get a running image of your equipment with documented data sources. Your team extends it to more lines and questions.
- Connection of controllers, sensors and logs
- Data storage in your Google Cloud
- Live view of the line in 2D or 3D
- OEE, downtime and energy per good part
- Local AI for image inspection and forecasts
- Alerts to the shift lead with approval
Building blocks
Four building blocks depending on your project
KPIs and downtime
We calculate OEE per VDMA 66412 from availability, performance and quality. A waterfall shows where time is lost.
We sort downtime by cause. Energy per good part shows what each stop costs.
Local AI at the line
An edge server in the plant inspects images within milliseconds. Images and logs never leave your plant.
A local language model answers questions about downtime and names the log lines it relies on.
Return channel to the digital twin
The digital twin sends set points back to the controller, such as a new stroke rate.
Every change needs the shift lead's approval. Your team stays responsible for the equipment.
Comparison
Digital twin or digital shadow: what is the difference?
A digital shadow only reads data from the equipment, a digital twin also sends set points back. Research names the digital model without automatic data exchange as a third stage. In everyday use all three are called digital twin.
| Feature | Digital model | Digital shadow | Digital twin |
|---|---|---|---|
| Data flow | No automatic exchange | From the equipment into the model | Both directions |
| Changes to the controller | None | None, read only | Set points after approval |
| Typical use | Planning and simulation | KPIs, forecasts, traceability | Control and optimisation in operation |
| Requirement | CAD or process model | Read access to controller and sensors | Write access, safety concept, acceptance |
| Our starting point | Only if needed | First stage of every project | Next stage once your team is ready |
Data flow
- Digital model
- No automatic exchange
- Digital shadow
- From the equipment into the model
- Digital twin
- Both directions
Changes to the controller
- Digital model
- None
- Digital shadow
- None, read only
- Digital twin
- Set points after approval
Typical use
- Digital model
- Planning and simulation
- Digital shadow
- KPIs, forecasts, traceability
- Digital twin
- Control and optimisation in operation
Requirement
- Digital model
- CAD or process model
- Digital shadow
- Read access to controller and sensors
- Digital twin
- Write access, safety concept, acceptance
Our starting point
- Digital model
- Only if needed
- Digital shadow
- First stage of every project
- Digital twin
- Next stage once your team is ready
Application
We start with one line and one question
That might be the rejects of a stamping line or the downtime of a packaging machine. You see early which data exists and which answers it gives.
The guideline of the Bavarian Digitalbonus names digital twins explicitly under no. 5.4.2 as an example for the Digitalbonus Plus. Small companies in Bavaria receive up to 50 % funding.
Related use cases
Business Intelligence
Reporting that every department relies on
Reports and dashboards with documented definitions that show the relevant developments and allow drilling down into the details.
See the exampleOperational Steering
Metrics directly in the application
Your teams see overviews directly in their applications. The metrics are documented. From the overview, your teams jump straight to the individual case.
See the exampleFrequently asked questions
An objection we often hear
„We need new sensors before we can build a digital twin.“
Do we need new sensors for a digital twin?
Usually not. Controllers, cameras, energy meters and machine logs supply the data. We add sensors only for a specific question.
Which machines and protocols do you connect?
We connect controllers via OPC UA, sensors and energy meters via MQTT. We read older equipment through log files or gateways.
Is a digital twin eligible for funding?
The guideline of the Bavarian Digitalbonus names digital twins explicitly for the Digitalbonus Plus. Small companies in Bavaria receive up to 50 % funding. You submit the application yourself. We prepare the offer and project description with you free of charge.
Getting started
Which line causes the most downtime or rejects today?
In a free initial meeting we discuss your equipment and the data you have. Then you decide whether a strategy workshop makes sense.
Starting point:Strategy workshop with a data source check on one line