Biomimicing generative Artificial Intelligence for Design of Heat EXchangers (BAIDHEX)

Purpose and goal

The project Biomimicing Generative AI for Design of Heat Exchangers aims to develop methods to generate and optimize heat exchangers inspired by nature. The methods will focus on non-periodic geometries with a large number of flow elements, which would be too demanding to create with conventional tools. The goal is to create an efficient description of these geometries and a way to generate alternative designs which satisfies functional, mechanical and manufacturing requirements.

Expected effects and result

The heat exchanger designs developed by the project have a potential to be lighter, more compact and give lower pressure drop than conventional ones. Consequently, we see an opportunity to develop these for applications in aerospace vehicles and engines. The geometrical models and methods to connect these to machine learning may have additional applications to other types of designs and engineering tasks. The project will also increase competence for machine learning related to 3D geometries.

Planned approach and implementation

A new PhD student at Chalmers will work with geometrical description, flow analysis and machine learning and senior researcher will set requirements and assist with testing. GKN Aerospace will give input to the preferred geometry and performance for the bioinspired heat exchanger, and how the flow elements should be shaped to achieve for performance and manufacturability. A wider group of researchers and engineers will evaluate alternative designs to give training input to the algorithm.