AI for CFD

Collection of innovative AI Models for Computational Fluid Dynamics (CFD).

Ultra-fast results, radically lower TCO. Explore new possibilities.

Key features

  • CFD solvers acceleration with AI / Artificial Intelligence
    (beyond what’s achievable with hardware upgrades or algorithms adaptation to hardware accelerators alone)
  • Replacement of numerical solvers with equivalent AI models
    (faster analysis, reduced cost of trial & error experiments)
  • Solver specific AI models for better accuracy
  • Digital twin ready
    (generalized AI models able to simulate various use cases and geometries; no need to have different models when parameters change)

Typical project deployment/scenario

  1. CFD Solver selection: pick the one you’d like to accelerate
    Typical candidate: simulations that take too long
  2. Setting the targets for AI Model(s)
    What level of acceleration do you expect? Reduce time/iterations? Anything else?
    What is the accepted accuracy? Within 10-15%? Higher? Lower?
    Should AI generate only the result of the simulation or intermediate steps as well?
  3. Solver specific information
    Consultancy to assess solver’s input/output (types, ranges, etc.)
  4. AI Model customization and training
    Based on the above, byteLAKE will train the CFD Suite to accelerate your CFD Solver(s) and generate results compatible with your existing tools

Example deliverables

AI Model(s) compatible with your existing workflow

  • No need to change any data types, replace toolchains, etc.
  • No hardware/infrastructure upgrades or changes needed
  • Cross platform compatibility (CPU-only, CPU+GPU, PC/ data center/ cloud etc.)

AI Model(s) easy to use

  • You just prepare your input data as you would normally do for your CFD Solver(s)
  • Instead of sending the data to your CFD Solver(s), you send it to the AI Model(s) and you do so in the same fashion as you would normally do for your CFD Solver(s)
  • You wait (much less!!!) for the results(s) that come in the same format as your CFD Solver(s) would produce
  • After all, you can still run your CFD Solver(s) based simulation
  • Plus, if you modify your CFD Solver(s), your AI Model(s) can adapt to changes as well!

AI Model(s) opening new possibilities

  • Much faster simulations
  • Significantly reduced cost of trial & error experiments

Benefits and new possibilities with AI

  • Ultra-fast results
  • Lower cost of trial & error experiments (i.e. much faster prototyping and parametrization)
  • AI Models outperform efforts related to hardware upgrades or algorithms adaptation to GPU/ FPGA accelerators alone
  • Real-time simulations for ever-increasing amounts of data
  • Run complex simulations within the same infrastructure
  • Lowering the entry barrier for CFD new adopters
  • AI Models can act as “CFD Calculators”, replacing CFD Solver(s)
  • Radically lower TCO (Total Cost of Ownership)
  • Easy to use. AI Model(s) work with the same data types and use the same infrastructure that you already have for your CFD Solver(s)

Roadmap

  • Replacement of solvers with AI Models
    AI Model(s) generalization (one model for various use cases, geometries)
    2D and 3D domains support for steady-state/transient solvers
  • Interactive AI Models
    Solver specific AI Models, interacting with solvers for solver’s intermediate results
  • Custom AI Model generator
    Toolchain for AI Models generation and updates
  • byteLAKE CFD Suite unleashed
    Collection of innovative AI Models for CFD across industries

byteLAKE's CFD Suite Roadmap

Product launch scheduled for November 2020.
Contact us to to learn more.

Join the CFD Suite community!

AI in CFD

Contact us at CFDSuite@byteLAKE.com to request access to byteLAKE CFD Suite. We offer a convenient and instant access in on-premise as well as Cloud configurations. For qualified partners there is also a possibility to receive access to byteLAKE CFD Suite with a complete hardware infrastructure.