TwinLearnXR Selected as an Official AID4SME Pilot
We are thrilled to announce that TwinLearnXR (Digital Twin for process optimisation and XR training) has been selected as an official Pilot for the AID4SME project (https://aid4sme.eu). Out of a highly competitive pool of European innovators, our solution was chosen to help bridge the gap between cutting-edge AI and industrial excellence.
We applied under the “Product-production Digital Twins” category, and we are ready to put our technology to the test.
This milestone is an important step for TwinLearnXR, as it gives us the opportunity to validate our AI-powered Digital Twin approach in a real industrial context and demonstrate how advanced simulation, predictive intelligence, and immersive learning can support the next generation of manufacturing.
What Is the TwinLearnXR Digital Twin Platform?
TwinLearnXR is an AI-driven Digital Twin and XR training platform designed to help manufacturing teams understand, test, and improve complex production processes.
The platform combines:
- Digital Twin modelling to simulate production behaviour
- AI prediction to estimate product quality outcomes
- Optimisation logic to recommend better process parameters
- XR-based training concepts to support operator learning and process understanding
The goal is to make manufacturing knowledge more accessible, interactive, and data-driven.
Instead of relying only on trial-and-error on the production line, users can test process settings virtually, receive estimated quality results, and explore better parameter combinations before applying changes in the real world.
Why Digital Twin Technology Matter in Manufacturing
Modern manufacturing processes are complex. Small changes in machine settings, material ratios, or product targets can influence quality, scrap, energy use, and process stability.
A Digital Twin helps by creating a virtual representation of the production process. This allows users to ask questions such as:
- “What will happen if I change these settings?”
- “Will this recipe still meet quality tolerances?”
- “Can I increase recycled material while keeping the product within specification?”
- “What process parameters are most likely to produce an acceptable result?”
TwinLearnXR turns these questions into an interactive workflow where users can run predictions, compare outcomes, and learn from the system.
Prediction to Optimisation in Digital Twin
TwinLearnXR currently supports two main Digital Twin modes.
Prediction Mode
In Prediction Mode, the user enters target product values and process parameters. The system then estimates the expected quality outputs and returns a final result:
OK, CA, or NOK
The prediction result includes estimated values such as dimensional deviations, surface energy, ash content, and other quality indicators.
Optimisation Mode
In Optimisation Mode, the user defines the target product values and allowed parameter ranges. The system then tests multiple candidate recipes virtually and recommends the best process settings based on quality, tolerance rules, and the selected optimisation objective.
This means the user can move from:
“What happens with these settings?”
to:
“What settings should I try?”
AI With Rule-Based Validation
One of the key design principles behind TwinLearnXR is that AI should support decision-making, not replace process control logic blindly.
The AI model estimates the expected quality outputs, but the final decision is validated by backend tolerance rules. This means that if the AI predicts a positive result but the estimated values violate hard tolerance limits, the system can override the model and return the safer final decision.
This layered approach improves trust, transparency, and practical usability.
Supporting Sustainable Manufacturing
TwinLearnXR also supports sustainability goals by helping users explore better use of recycled and regrind materials while maintaining product quality.
By testing parameter combinations virtually, manufacturers can reduce unnecessary physical trials, identify more stable recipes, and support better process decisions.
Potential benefits include:
- Faster operator learning
- Better process understanding
- Reduced scrap and rework
- Improved use of recycled materials
- More consistent quality prediction
- Safer testing of process changes before production
XR Training and Operator Learning in Digital Twin
Beyond prediction and optimisation, TwinLearnXR is designed to support future XR-based training experiences. It leverages Immersive4Learning developed by ARVRtech.
The vision is to help operators learn complex production processes through interactive simulations, visual guidance, and scenario-based training. This can reduce learning time and make technical knowledge easier to transfer across teams.
By combining AI, Digital Twin logic, and immersive learning, TwinLearnXR aims to support both experienced engineers and new operators.
A Step Toward Industrial AI Adoption
Being selected as an official AID4SME Pilot is a strong validation of the TwinLearnXR concept.
It gives us the opportunity to demonstrate how AI and Digital Twins can be made practical, explainable, and useful for real manufacturing environments.
Our focus is not only on building an AI model, but on creating a complete workflow:
- Data-driven prediction
- Process optimisation
- User-friendly interaction
- Secure access
- Traceable results
- Future model retraining
- Operator learning support
What Comes Next
As part of the pilot, TwinLearnXR will continue to evolve from a working prototype into a more complete industrial decision-support platform.
The next development steps include improving the optimisation engine, expanding user management and history tracking, supporting controlled model retraining with new measured data, and preparing the system for richer XR learning scenarios.
We are excited to begin this next phase and to contribute to the future of AI-enabled, sustainable, and human-centered manufacturing.
TwinLearnXR is ready to put Digital Twin technology to the test.
