NavierWorks

Precision in dynamic environments

Real world scenarios

Most physical systems do not operate in an isolated, well-controlled environment. Finding edge cases and complex bugs often needs real-world feedback. By harnessing traditional industry proof engineering solutions and advanced ML models, it is possible to reduce physical testing by building complex digital twins. Agents not only help set up individual tests and simulations, but also build complex real world simulations where coupled effects produce vastly deeper, invaluable insights. NavierWorld is the playground that provides an accurate training and validation environment where users and agents collaborate to build highly accurate virtual simulations.

Physics informed machine learning

Speed-up simulations

Traditional solvers, while accurate, are hard to configure and often slow on iterations. By using surrogate models and neural operators, it is possible to speed up the exploration of design spaces and reduce the number of candidate solutions where high fidelity and precise simulation are necessary. Setting up complex multiphysics simulations and evolutionary algorithms to Pareto-front optimize can be a true headache; that's where NavierAgent can speedup the process by performing exploratory analyses and perform low resource based exploration to support design and optimization path planning.

Performance (case study)

Engineered for speed

No matter the complexity or the stage of the system design, speed to accurate numbers will always win. NavierWorks is currently being battle tested in race car designs and on-track aerodynamics setup optimizations. Using traditional racecar log data and pressure sensors on racecars, it is possible to digitally clone the car and perform setup space discoveries to quickly find improvements. Using all the data gathered and learned on trackside, engineers can later test geometrical modifications and backtest them as if they were still on track.

Performance (case study)

Control system and algorithm tests

NavierWorld real-time real-world (RTRW) extension allows users to use sensors, controllers, and other physical components as part of the digital twin simulations. Using physical flight computers of drones and rockets, NavierWorld is able to mimic the real-world massiveness and complex physics, closing the feedback loops without firing any engine so engineers can better test and validate their designs.

|

Real world scenarios

Speed-up simulations

Engineered for speed

Control system and algorithm tests

NavierWorks Release Coming This Fall