Learn how to use NVIDIA Modulus to develop physics-based surrogate using physics-informed neural networks (PiNNs). The talk focuses on a specific use-case of multiscale chemical reactors modeled using hierarchical PiNNs structure. Topics include how to design such a complex PiNNs structure mimicking physical reality in digital-world, and how to leverage NVIDIA Modulus as a tool to achieve this. The surrogate model thus developed enables inferencing at ~10^8 times faster than the conventional methods with negligible loss of accuracy.

Speakers:
▫️ Tymofii Nikolaienko, Researcher, Softserve Inc.
▫️ Aniruddha Panda, SciML Researcher, Shell
▫️ Harshil Patel, Scientific Machine Learning Researcher, Shell

Explore more GTC 2024 sessions like this on NVIDIA On-Demand: https://nvda.ws/3U33qo7

Learn more about Modulus: https://nvda.ws/3UvXcx3

Read and subscribe to the NVIDIA Technical Blog: https://nvda.ws/3XHae9F

#GTC24 #NVIDIA #GTC #AI #Modulus #Physics #Simulation #Modeling #ComputationalChemistry #Chemistry #MaterialScience