Vanellus is developing a machine learning-based physics simulation solver that significantly enhances the efficiency of computational fluid dynamics (CFD) for complex geometries, particularly in heat exchanger design for additive manufacturing. This technology addresses the high computational costs and lengthy simulation times that hinder product development across industries such as aerospace and manufacturing.
Funding
$380K raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Engineers rely on computational fluid dynamics (CFD) simulations to optimize designs, but simulating complex geometries, especially in heat exchanger design for additive manufacturing, is computationally expensive and time-consuming. These lengthy simulation times slow down product development cycles across industries like aerospace, manufacturing, and integrated circuits.
Solution
Vanellus is developing a machine learning-based physics simulation solver that accelerates computational fluid dynamics (CFD) for complex geometries. By integrating machine learning with modern numerical programming, the solver achieves significant efficiency gains, reducing simulation times from hours to minutes. The solver is deployed through a web application and API gateway, enabling engineers to efficiently search for optimal designs and develop better products. The technology allows for native deployment to hardware accelerators like GPUs and TPUs and incorporates adjoint methods for optimization.
Target Audience
The primary target audience includes engineers and researchers in aerospace, manufacturing, and integrated circuits who require efficient and accurate CFD simulations for complex geometries.
Features
- Machine learning-integrated physics solver for computational fluid dynamics (CFD)
- Order of magnitude efficiency gains in simulation speed
- Native deployment to hardware accelerators: GPUs and TPUs
- Adjoint methods for arbitrary optimization targets
- Web application and API gateway deployment
- Differentiable solver, enabling machine learning-driven improvements
- Support for inflow and outflow boundary conditions for simulating channel flows
- Validated against experimental and numerical studies, demonstrating second-order accuracy