Applied Dynamics Research
CFD expertise for motorsport, aerospace, and research.
Expertise proven at the highest level of motorsport, including Formula 1.
Leading teams in motorsport and research rely on our CFD methodology and simulation infrastructure. Detailed case studies available under NDA.








Our Intrascale™ Compute Cluster
Custom cooling throughout
AMD EPYC Milan — 640 cores of raw CFD compute power
Frontier models running on-premise. DeepSeek R1 for research and NDA-protected dataset analysis. Your data never leaves our facility.
No cloud. No third parties. No leaks.
Deeply coupled partnerships for chassis development, vehicle dynamics, and design engineering in Japan. Complete production capability without overhead or markup in the middle.
From 3D scan to race-ready CAD. Aerospace tolerances. Experience: over 100 vehicles including IndyCar and Formula 1.
Surface deviation analysis — scan data provided by team, reverse engineered to micrometer tolerance.
Ligier JS P320 — Creaform HandySCAN
Surface design refinement — optimized for downforce, drag, and balance. Prototype experience includes Norma, Ligier, and Wolf chassis.
Engineering Platform
Creaform scanning systems. 0.05mm accuracy standard, micrometer precision available depending on part size.
Geomagic Design X transforms scan data into feature-based CAD with automatic and guided extraction.
Licensed for Siemens NX and SolidWorks. Specialized tools for CATIA import. CFD-ready geometry output.
ANSYS with custom solvers on our Intrascale cluster.
Founding Partner · Wind Tunnel · Master Model Builder
In-house 3D printing for wind tunnel models and prototypes
Four-year collaboration with Nissan Technical Center. Yoshi Suzuka and Andrew Brilliant developed the SST (Super Small Tunnel) — a compact prototype that matched the full-scale NVH facility for accuracy and repeatability. Multi-scale methodology: start small and inexpensive, increase model scale as the design matures. Custom sting apparatus, precision measurement systems in LabVIEW, and CFD simulation validated against experimental results.
Director of Computational Physics
ANSYS with custom solver development. Complete automation system with on-demand cluster power up/down.
In-memory solutions: 1.5TB (CPU) · 400GB (GPU accelerated)
Critical results validated through multi-solver consensus rather than single-method averaging. Adversarial configurations expose assumptions; agreement across independent methods is stronger than convergence within one.
Aerodynamics · Electronics cooling · Fluidic tuned mass dampers
GPU-accelerated solvers for complex field problems. Lattice gauge theory ported to CUDA on our cluster.
Formal verification via Lean — based on Sannai group methodology, Kyoto University
Inviscid flow · Superfluid dynamics · Condensed matter · EM fields
Computational infrastructure for research groups. We implement and optimize — your team focuses on the science.
Custom pipeline for large dataset analysis. Pattern recognition across parameter spaces. Local infrastructure — no cloud dependency, no data leaves the building.
Off-the-shelf tools didn't cut it. We built our own — secure, interactive, VR-ready.
Your aero data. Securely delivered. Audit-ready.
Interactive Aero Maps
Flow Visualization
3D surface plots. Advanced interpolation between test points. No more spreadsheet guessing.
Auto-isolates turbulent from laminar flow. Selectable vortex bodies. Streamlines seeded from structures — see complex flows without the noise.
Reads what you mean, not where you click. Filters outliers from turbulence and geometry — 17% more precise than traditional sampling.
STEP, IGS, X_T, EPRT. Measure directly in-browser. No CAD software needed.
Step inside the flow field. Meta Quest, Valve Index, SteamVR.
Encrypted app. Audit-ready logging. Windows, macOS, Linux.
Wake Structure Isosurface
Structure-Seeded Streamlines
Generic CFD post-processing means random seed points, spaghetti streamlines, and manual iteration hoping to expose coherent flow structures. It's slow and non-deterministic. Our Wake Structure Isosurface methodology replaces this with automated detection of vortical regions using structure intensity metrics, separating turbulent from laminar flow algorithmically. Select individual structures by strength, generate streamlines seeded from inside each vortex core rather than arbitrary spatial points, and obtain clear insight into complex three-dimensional aerodynamics in minutes rather than days.
The Consensus Pressure Dropper complements this by filtering geometric sensitivity and turbulent outliers to provide stable surface measurements on challenging geometries: radiator faces, fin arrays, heat exchangers, cooling ducts. Internal validation shows 17% improvement in precision relative to conventional point sampling.
Throughput and decision latency. From solver completion to fully generated wake structures, streamlines, and report artifacts: approximately three minutes. Equivalent workflows in conventional ANSYS-based post-processing routinely require 20-40 minutes per iteration. This matters because solver resources are expensive and finite. While human decision-making is delayed, compute infrastructure sits idle. You're queuing up your next runs blind. Meshing starts only after you finally see results. Every second of decision latency is wasted wall-clock time. We needed reports ready almost immediately after solution. This system made the difference.
Differential analysis modes. Side-by-side comparison with automated delta computation. Changes in streamlines, surface pressures, and Cp distributions are quantified and highlighted directly. The system exposes causal effects of geometric modifications without manual A/B squinting.
Local-first architecture. VR CFD from the cloud? Good luck with latency. And the egress costs when you're moving terabyte-scale datasets. Our viewer runs entirely on your local 100 Gbps network. A 1 TB solution transfers in ~80 seconds and loads in VR instantly. No external dependency. No off-site data transfer. No throttling. Engineers work directly on the complete solution field, not down-sampled subsets. This isn't a plugin layer or scripting framework. It's purpose-built infrastructure for serious aerodynamic analysis.
Publication-quality output for journals, internal reviews, and academic dissemination. If your research team is spending more time wrestling with plotting libraries than interpreting physics, talk to us.
Results across five continents
Toyota Racing Development, Japan
Toyota Racing Development, Japan
LG Motorsports — fastest privateer team
"Best of the rest" on lowest budget in series
Vidar Frogner, Porsche 993 GT2 — Rudskogen track record
Dawie Joubert, Lotus Exige "Lotari" — Killarney track record
Paul's Automotive Engineering, Faessler Mustang — 11 track records
Lyfe GTR • Tesla
Gobstopper • Norris
Revline • J-Spec
HKS • Under Suzuki
Tanuki • Nemo
Zeelie • Joubert
AMB Aero® secures 4th championship title with Tanuki's dominant performance, missing the outright lap record by just 0.5 seconds.
Read more →Pieter Zeelie shatters hillclimb record with 37.090s, defeating Nissan GT-Rs by 1.3 seconds. Applied Dynamics Research aerodynamics deliver back-to-back Simola victories.
Read more →Roger Clark Motorsport's Applied Dynamics Research-equipped BRZ obliterates 9-year Brands Hatch record by 3.6 seconds, dominates 2024 UK Time Attack Pro Extreme championship.
Read more →Cincinnati racing team dominates 2025 season with Applied Dynamics Research aerodynamic package, breaking Pittsburgh track record by 7 seconds.
Read more →Applied Dynamics Research aerodynamic development doubles downforce on 800hp Toyota MR2, achieving fastest RWD time at South Africa's premier hillclimb.
Read more →Applied Dynamics Research-equipped Mitsubishi Eclipse claims second overall at Canada's premier hillclimb, shattering 9-year-old class record.
Read more →Ready to accelerate your motorsport or research project? Share your requirements and we'll provide a tailored solution.
USA — Berkeley
+1 415 525-6463
Japan — Sapporo
+81 11-311-6866