Supervisor:
Henrik Sørensen

Co-Supervisor:
Jakob Hærvig

Amir Sajjad Bahman

Assessment Committee:
Francesco Iannuzzo(Chair)

Prof. Marco Marengo, MSc PhD, Department of Civil Engineering and Architecture, University of Pavia, via

Timo Pättikangas, MSc PhD, Research Team Leader Process modelling team, VTT Technical Research Centre of Finland Ltd,

Moderator:
Torsten Berning

Abstract:

One of the current trends in modern engineering and design is the push toward miniaturization without compromising functionality and performance. This compact design often results in a significant increase in heat flux, thereby hindering the system's optimal performance, particularly when cooling is not adequately handled. Traditional single-phase cooling systems that rely on forced convection are seemingly reaching their limitations and are no longer capable of meeting the escalating cooling requirements. On the other hand, two-phase cooling technologies, particularly those involving boiling, are promising in managing high heat flux scenarios, thanks to the exceptional heat transfer performance resulting from substantial flow mixing due to bubbling and the latent heat of phase change.

 

Regarding boiling cooling methods, an in-depth understanding of flow and heat transfer processes is necessary during phase-change phenomena to enhance system performance and prevent unexpected off-design circumstances. This emphasizes the significance of research into thermally-driven phase-change processes.

 

This research focuses on two primary objectives: first, studying different methods to enhance the performance of simulation techniques for phase change phenomena; and second, studying approaches for improving the performance of boiling cooling methods.

 

In addressing the first objective, a significant challenge in modeling flows with thermal phase change is accurately pinpointing the gas-liquid interface. The Volume of Fluid (VOF) technique remains the most widely adopted approach for characterizing interfaces in commercial and open-source CFD software. However, the use of VOF can lead to imprecise curvature computation and smeared interface prediction, resulting in non-physical velocities, particularly close to the interface. To recover accuracy in curvature computation for VOF simulation of boiling, a solver that combines VOF with the level-set method for interface depiction is used. VOF is employed to capture the interface due to its mass-conserving nature, while the level-set method is used to compute the curvature and physical properties near the interface. Another alternative way to enhance accuracy is by addressing the smeared face problem through the application of a recent geometric method called isoAdvector.

 

In the initial phase of this Ph.D. project, the coupled level-set and VOF (CLSVOF) and VOF-isoAdvector methods are incorporated into an in-house OpenFOAM thermal phase change solver. A comparative analysis with VOF, the default interface capturing method in OpenFOAM, is subsequently carried out using various thermal phase change benchmark cases to highlight the respective advantages and disadvantages of these methods. The CLSVOF simulation results indicate that this approach yields enhanced accuracy in curvature computation and a higher degree of alignment with analytical/benchmark solutions, albeit at the cost of extended computational time. Moreover, our investigations reveal that, in most thermal phase change scenarios, isoAdvector provides a faster solution than MULES while preserving nearly the same accuracy and convergence rate.

 

In the latter part of this research, a new algorithm for simulating nanofluid boiling is developed and implemented, which addresses the shortcomings of previous CFD models in the literature. This novel approach models the behavior of nanofluid boiling by considering the increasing nanoparticle concentration during boiling, making it a significant step forward compared to existing methods. The new algorithm is numerically implemented and tested using the FVM-based OpenFOAM toolbox, with experimental correlations integrated into new C++ classes and OpenFOAM modules to compute properties such as viscosity, density, specific heat, and surface tension for nanofluids. The results from the study using this improved simulation method indicate that employing nanofluids as a coolant instead of pure fluids enhances heat transfer and accelerates the bubble formation rate. These factors contribute to increased heat transfer performance, further highlighting the potential of nanofluids in optimizing the effectiveness of boiling cooling technologies. The development of this novel algorithm for modeling nanofluid boiling, along with the subsequent simulation results, showcases the benefits of this approach and its potential impact on future advancements in nanofluid boiling simulation methods.