AI-Enhanced Digital Twin for CNC Milling Optimization

Use Case:

DARTEF aimed to develop a predictive AI module integrated with a digital twin of CNC milling processes. By leveraging sensor fusion physics-informed simulation and neural networks the solution enables real-time prediction of tool wear and process quality.

Outcome:

• Integration of AI model (multivariate LSTM neural network) with physical simulation engine (Simcenter Amesim);
• Use of DARTEF’s Virtual Machining Twin and retrofit sensor suite (vibration, acoustic emission, power) on a 3-axis CNC milling machine;
• Training on real sensor data collected from experimental machining cycles with various spindle speeds, feed rates, and depths;
• Evaluation of energy usage scenarios via simulated toolpaths for optimal process planning;
• Demonstration of improved accuracy in tool state prediction and early failure detection using sensor-informed LSTM models.

Ecosystem Support:

StairwAI support enabled access to AI expertise from DEPTHEN, including neural network design, training strategies, and data normalization. DARTEF utilized cloud computing for model training and benefited from the mentoring and support infrastructure provided by the StairwAI program.

AI Relevance:

This success story highlights the power of AI for industrial digital twins via:

1. Integration of physics-based simulation and AI-driven predictions;
2. Sensor fusion using low-cost retrofitting kits;
3. Improved efficiency in production lines through predictive maintenance;
4. Increased accessibility of smart manufacturing for SMEs.

Summary:

DARTEF developed an AI-powered digital twin of a CNC milling machine, integrating real sensor data and physics-based simulation to enable real-time prediction of machining conditions and tool wear. The system used a multivariate LSTM neural network trained on data from a 3-axis retrofitted machine equipped with acoustic, vibration, and power sensors. The project focused on predicting surface roughness, tool state, and energy consumption, allowing the optimization of machining parameters before and during production. The solution was successfully validated through experimental trials and simulation-based process planning. StairwAI support facilitated AI development and model training, bringing DARTEF closer to offering intelligent digital twin solutions for small and medium-sized manufacturers.