Yonglae Lee

Senior Engineer (Dr.-Ing.) at Hankook Tire, Europe Technical Center (ETC), expert in computational biomechanics and tyre modelling

Kenneth Richardson

Research Engineer at Hankook Tire, Europe Technical Center (ETC), specialized in vehicle dynamics and simulation

Why is virtual development essential today?

Development timelines are shrinking while vehicle complexity grows. Most importantly, tyre models for OEM vehicle simulation have become a commercial prerequisite. Since 2021, 100+ OE programs use virtual tyre models, with pure virtual pilot programs achieving mass production after just three virtual loops plus one physical loop.

What role does virtual testing play?

Virtual testing starts at the concept stage before physical tyres exist. The Tire Digital Twin is a model family connected by common data. AI systems propose specifications in a few minutes using our database based on 20+ years of history.

OEMs (Original Equipment Manufacturers) develop new vehicles using advanced simulation software. In this process, vehicle concepts and design candidates are tested virtually with dedicated vehicle simulation programs.

These simulations require not only complete vehicle models, but also accurate tyre models, as tyres play a crucial role in overall vehicle behaviour. When tyre models are provided, OEMs integrate them into their vehicle models and use them to run detailed simulations under various conditions.

Since tyres strongly influence vehicle performance, high-quality tyre models are essential for reliable results.

Different models are used depending on the objective:

• MF-Tyre: Handling and vehicle dynamics (steering, braking, stability)
• MF-Swift: Extended MF-Tyre with additional dynamics for handling and ride
• FTire: High-detail physical model for advanced dynamics, ride comfort, and chassis development
• CDTire: High-fidelity model for NVH (Noise, Vibration, and Harshness), as well as ride comfort, durability, and vibration analysis

OEMs select the appropriate tyre model based on their development goals, balancing accuracy and computational effort.

What are current limitations?

Four key challenges remain: (1) Models are parameterized on sandpaper drums, not real asphalt; (2) High-fidelity models still require physical tires for parameterization; (3) Single-specimen validation misses manufacturing variation; (4) Extension to wear state and in-service data is needed. Wet, snow, and ice simulation capabilities are expected to make significant advancements between 2026 and 2028, driven by ongoing research collaborations.

How well do driving simulators correlate with reality?

Correlation between simulator and physical evaluation is generally good, with the highest accuracy in the linear to early nonlinear range, meaning in conditions where tyre behaviour is still relatively predictable before becoming highly complex.

Characteristics such as on-centre steering feel, lane change behaviour, linear response, and understeer behaviour are well captured. Multiple test campaigns since 2021 confirm that drivers can reliably distinguish between different tyre specifications in simulator environments.

Is virtual testing replacing physical testing?

No — virtual testing complements physical testing. Off-line simulation gets close to target, but final human-in-the-loop judgment improves the design. The real measure of success is achieving OEM approval at the first physical tyre test. Physical testing remains necessary for low-friction conditions, subjective evaluation, and safety certification, but its role is shifting from discovery to verification.

What makes tire modeling challenging?

The tyre is the accuracy bottleneck of every vehicle simulation — the most nonlinear, highest-uncertainty component. Rubber properties shift with temperature, speed, load, and slip. Critical parameters depend on frequency: cornering stiffness for handling (0–10Hz), dynamic response for ride (10–100Hz), belt resonance for NVH (100–500Hz). No single model covers everything — hence the need for a model family anchored to common physics.

How early are tyre manufacturers involved with OEMs?

Earlier than ever. OEM partners now provide detailed vehicle data at concept stage, enabling tyre integration into driving simulators before physical prototypes exist. Vehicle-tyre matching begins with RFQ planning, proceeding through virtual design, simulation, and driving simulator evaluation before experimental tyres are built. This early integration is a competitive differentiator in premium OE business.

What role does AI play?

Neither pure machine learning nor pure FEM alone is enough — a hybrid delivers both accuracy and speed. AI-powered systems predict performance from hundreds of design variables across 2.48 million test results, built on Azure MLOps infrastructure. Surrogate models generate MF-Swift specifications in under one minute. AI reinforces physics-based models but does not replace them — the foundation remains fundamental tire-road physics.

What are the sustainability benefits?

Virtual development reduces prototypes, cutting material consumption, manufacturing energy, and transportation emissions. Every virtual loop is one less physical tyre built. Further CO₂ reduction is expected as wear testing is replaced by simulation. The digital twin is a genuine driver of greener innovation without compromising development quality.

What does the future hold?

Simulation-first development is now standard practice among premium OEMs. Physical winter testing will remain necessary for the foreseeable future, but hard-packed snow simulation is advancing through 2026, with wet and soft snow following through 2028. Full virtual homologation is partial reality and will expand as regulatory frameworks adapt.

AI will complement, not replace, physics-based models — a physics-plus-data hybrid is the future standard. Developers need integrated competency across tyre mechanics, FEM, data science, and software engineering. Drivers are evolving from evaluators to analytical partners who validate simulation models and quantify subjective data. The future is collaborative: no company achieves this alone.

Key Performance Metrics

• • 100+ OE programs using virtual development since 2021
• • 30% time reduction: 9 months → 6 months
• • 30%+ development efficiency gain
• • A few minutes AI specification generation per tyre size
• • Hundreds design variables, 20+ years history, 2.48M test results

Technology Roadmap Highlights

• 2026: GPU solver optimization, hard-packed snow traction, expected mileage prediction
• 2027: Wet asphalt simulation validation, soft snow simulation
• 2028: Advanced NVH models, comprehensive durability analysis, ice friction simulation
• 2029-2030: Full lifecycle digital twin with in-service data integration