Empirical Digital Twin
Create Simulation Models from Real Hardware
An accurate empirically derived digital twin model of a real powertrain can support a more optimized verification, validation and sign-off process.
Existing test automation systems and hardware can be used to follow a transient excitation cycle to simulate a non-linear system such as a hybrid powertrain. The benefit of this is that while some level of robustness can be achieved with traditional lab-based and in-field testing, a digital replica can be used to explore all possible scenarios thereby providing greater coverage and reduced risk.
Benefits of an Empirical Digital Twin
- Trustworthy simulation results:
- Models are verified against a real system.
- Ideal for complex, non-linear systems:
- Model response accurately describes nonlinear behavior. Suits IC engine, hybrid systems, thermal systems, and fuel cells.
- Quick to make:
- Training data cycles scale with variables not time
- Highly precise, based on real data:
- Excellent R2 correlation
- Faster than real time:
- No complex calculations mean scenarios can run in seconds.
- Identify areas of concern quickly:
- Simulate difficult to test regions to check conformity and highlight issues.
- Improve global vehicle models when necessary:
- Improve vehicle plant models
- Share models across departments:
- Multiple departments can benefit from increased model accuracy.
Empirical Digital Twin Creation Process
A Digital Twin can be created from any complex vehicle system: ICE, e-motor, thermal, vehicle.
1. Experiment design
Create dynamic test points
2. Generate training data
Record test data from lab to generate training data.
3. Model creation
Use ‘training’ data to build empirical model of attributes.
4. Validate model
Validate model against real unit to confirm accuracy (R2)
5. Run simulation(s)
Simulate scenarios* with Empirical Digital Twin
30x faster than real-time!
6. Predict responses
Predict Performance, Electric Range or Emissions
7. Identify and mitigate Hotspots
Identify “hotspots” – problem areas that need to be worked on
