Combustion Modelling

Supervised learning of fast, compact models for calculations needed to simulate combustion.

DOMAINS | ai-for-science | combustion-modelling

Inspired by a challenge raised by Prof. Jean-Pierre Hickey in Mechanical Engineering I have worked to investigate novel ways to use Deep Neural Network to revolutionize the way combustion simulation calculations are computed and stored. This research is relevant for the design of combustion engines and energy production. Our results are significantly better than existing approaches and could lead to an order of magnitude speed up in some combustion simulation models. This work was presented at the European Conference on Machine Learning conference in September 2019.

Our Papers on Combustion Modelling

  1. ECML
    Compact Representation of a Multi-dimensional Combustion Manifold Using Deep Neural Networks
    Sushrut Bhalla, Matthew Yao, Jean-Pierre Hickey, and Mark Crowley
    In European Conference on Machine Learning (ECML-19). Wurzburg, Germany. 2019.