Using Machine Learning to Optimize the Mixture of a Hydrogen IC Engine

May 29, 2024 9:30 AM – 10:30 AM CDT

Presented by:

Alen Jose, Specialist Engineer
Volvo Group Trucks Technology

Dan Probst, Sr. Principal Engineer
Convergent Science

Mukul Biware, Sr. Research Engineer
Convergent Science

Hydrogen internal combustion engines (H2ICEs) have sparked global interest as a potential solution to reduce carbon emissions from transportation. Research and development efforts are underway around the world to advance H2ICE technology to make it a viable solution for widespread use. State-of-the-art methods such as computational fluid dynamics (CFD) and artificial intelligence (AI) can help accelerate this process, enabling engineers to efficiently optimize the design of H2ICEs. In this webinar, guest speaker Alen Jose from Volvo Group Trucks Technology will present a study that employed CFD, machine learning (ML), and high-performance cloud computing to optimize the injection parameters of an H2ICE to achieve the best mixture preparation. In the study, a base CFD model was established in CONVERGE and validated against experimental data from the Sandia H2ICE. The research team identified factors that affected mixture preparation, and performed a large Design of Experiments (DoE) study using computationally efficient CFD simulations run concurrently in the cloud. An ML model was trained on the CFD data and then used to optimize the mixing-preparation process. Convergent Science engineer Dan Probst will additionally demonstrate a new ML tool integrated in CONVERGE Studio that enables users to conduct their own design optimization studies with CONVERGE.

Register Now!