Computational fluid dynamics for improved components in refrigeration systems

Knut Emil Ringstad, PhD Candidate at NTNU

The use of natural refrigerants in modern refrigeration systems has seen large growth in recent years, boosting high efficiencies without impacting the environment. To facilitate this transition, development of efficient systems and components is necessary.

Today, state-of-the-art computational tools, such as computational fluid dynamics (CFD), can give insight into the local flow behaviors inside these components.

In our work, we focus on two-phase ejectors for CO2 refrigeration systems, a critical component for large scale adoption of CO2 systems, especially in warm climates.

The two-phase flow in these components is highly complex due to phenomena such as super-sonic flow, bubble/droplet interactions, non-equilibrium thermodynamics and two-phase turbulence. Therefore, an extensive review of current knowledge in the field has been conducted and is set to be published in 2020.

Currently, development of a novel model which can achieve accurate results with low computation time is under way. The model under development is based on a Two-Fluid multiphase model, which will give new insights into CO2 ejectors.

Furthermore, an algorithm for automatic meshing of generic ejector geometries has been developed. This algorithm dramatically reduces the development time of novel ejector designs. Based on this, an optimization algorithm using automation of the CFD workflow and machine learning is under development. This will allow for better and faster design processes for ejectors, which will improve efficiency for CO2 based refrigeration systems.

Further work will be in cooperation with Danfoss to investigate next generation CO2 two-phase ejectors.