top of page
Eigen Control
  • Writer's pictureIsabella Smetana

The Case for a Next Gen CFD Platform

We're excited to introduce Eigen's breakthrough product: a pioneering computational fluid dynamics (CFD) solver platform designed to accelerate solutions to engineering problems across many industries, such as energy transition, automotive, aerospace, and computer chip fabrication.


What’s the addressed fundamental challenge?

Accelerating time-to-solution in Computational Fluid Dynamics (CFD) applications presents one of the most fundamental engineering challenges to the advancement of manufacturing, transportation and access to sustainable energy. Imagine the advantage of certifying the airworthiness of an airframe by CFD analysis, validation of new chemical catalysts by CFD analysis instead of field tests, or the advantage of building a bioreactor to commercial scale and bypassing the buildout of test plants by CFD analysis. Current state-of-the-art CFD applications must be fundamentally rewritten to accelerate such engineering aspirations into reality. They suffer from:

  • software that doesn’t efficiently scale in emerging supercomputers, and

  • legacy mathematical algorithms that were dictated by compromises to leverage decades old hardware; new math is needed for new fast changing hardware.

What’s our solution?

Eigen Control accelerates CFD applications by fundamentally rewriting state-of-the-art CFD solvers on top of AI platforms, TensorFlow and PyTorch, as the linear algebra engines. We do this to A) harness the vast computational power of AI hardware to fully load the GPU kernel and B) to implement new numerical methods with tremendous speed. As a result of both advances, time-to-solution can be significantly shortened. For example, a reservoir simulation for a Carbon Capture Storage application can take over 30 days to run with current reservoir modeling CFD technology; that time-to-solution can be shortened by an order of magnitude with Eigen’s solver platform. Such speed can unlock vast economic value for our customers. Ours is a unique approach that has required new math and ingenuity. We built the mathematical advancements on the shoulders of the PhD research in Applied Mathematics and experience at Los Alamos National Labs of our co-founder and CTO, Brandon Chabaud.


Our Journey

We’re a young company in the early stages of product development. For the past year we have built our first solver of many. It solves complex turbulence modeling with a state of the art numerical method - with just one dependency: TensorFlow. (Demo below) We seek to hire talent (mathematicians, AI software engineers, software infrastructure engineers etc) to implement our products in commercialized applications. If you are one, please come join us.


Why now?

Recent breakthroughs in computational power of parallel computing for generative AI applications are technological achievements that lend themselves directly to accelerating CFD applications. The same linear algebra engine platform (Tensorflow) that powers AI breakthroughs can power the linear algebra required for complex CFD applications. Meanwhile AI hardware has made enormous improvements in the cost and performance of raw compute; and, powerful exaflop supercomputing hardware is becoming more accessible. New math and algorithms are needed to harness such computational power. The time for a new CFD platform is now.


Energy Transition

We’re a company based in Houston, the energy capital of the world. Our ecosystem is embracing the urgency for Energy Transition. However, accelerating time-to-solution in CFD applications is one of the fundamental choke points in Energy Transition today. Fast and accurate engineering toolsets are required to harness new energy. The Eigen Control cutting-edge solvers would enable tremendous speed up in time-to-solution engineering use cases:

  1. Carbon Capture and Storage Systems: CFD solvers play a crucial role in revolutionizing the efficacy of carbon capture and storage (CCS) technologies, which mitigate climate change by capturing and sequestering CO2 emissions from industrial processes and power plants.

  2. Renewable Energy: Efficient design of new bioreactors for green fuels, optimization of wind turbines and solar panels.

  3. Air Emissions Monitoring, Plume Diffusion, and Source Identification: CFD solvers play a critical role in studying air emissions, tracking the dispersion of pollutant plumes, and identifying their sources.

  4. Water Management: CFD solvers assist in optimizing water treatment facilities, flood control systems, and irrigation networks, ensuring efficient water usage and distribution while minimizing environmental impact.

  5. Electric Vehicles (EVs): As the adoption of EVs grows, optimizing battery performance, battery cooling systems, and aerodynamics becomes increasingly important for enhancing range and efficiency.

  6. Sustainable Infrastructure: CFD solvers aid in designing energy-efficient buildings, optimizing heating, ventilation, and air conditioning (HVAC) systems, and planning urban green spaces, all of which contribute to the development of sustainable cities.

  7. Marine and Ocean Engineering: By enabling the analysis of fluid-structure interactions, CFD solvers support the design of sustainable offshore structures, wave energy converters, and eco-friendly ships, promoting responsible ocean resource utilization.

  8. Fusion Plasma Modeling: CFD solvers are indispensable for advancing the development of nuclear fusion as a clean and virtually limitless energy source.

It's hard to overstate the importance of fast and accurate CFD analysis in many engineering applications. Our advanced CFD solvers, powered by machine learning platforms, have the potential to revolutionize several industries. This technology would be particularly valuable in accelerating pressing Energy Transition engineering requirements. Investing in, partnering with, and sup

porting our technology could drive significant value in the market and contribute to a more sustainable and responsible future.





Comentários


bottom of page