- Convective heat transfer due to rising bubbles interacting with natural convection flow from heated plates - PhD Student S. Senthil KUMAR.

- Efficient parallel Krylov Subspace Solvers for Multifluid flows - PhD Student Bipin KUMAR.

To see a poster on the work of Yann Delaure and his group, presented at the first Sci-Sym group meeting in January 09, please follow the link: POSTER of Yann Delaure and Coworkers 09.

 Convective heat transfer due to rising bubbles interacting with natural convection flow from heated plates .

PhD student: S. Senthil KUMAR, supervised by Yann Delaure

Funded by IRCSET Basic Research Grant 04.

Collaborations with the Fluids anf Heat transfer Laboratory of Trinity College Dublin.

Project Description: Two-phase flows occur in a wide variety of energy systems ranging from boiling solar collectors and nuclear reactors to cooling systems for microelectric devices. In such applications vaporisation of the liquid phase occurs under normal operating conditions and gives rise to bubble nucleation, detachment and convection. The high rate of heat transfer associated with this nucleate pool boiling phenomenon is known to result from the phase change at heated surfaces as well as from convective cooling due to the effect of bubble detachment and motion on liquid agitation. A numerical code based on the VOF multifluid method has been developed to investigate the convective cooling part of this process. Sample computational results are shown below.

Efficient parallel Krylov Subspace Solvers for Multifluid flows

PhD student: Bipin KUMAR, co-supervised by Martin Crane and Yann Delaure.

Funded by DCU Connect Grant.

Project description: The numerical solution of multifluid flow problems present significant difficulties whether they refer to large scale problems such as ocean wave propagation and interaction with marine structures or microscale problems such as rising air bubbles in water. The modelling method adopted by the group is known as the Volume of Fluid method and treats the two fluids as a single mixture whose properties vary as a function of the volume fraction of each phase. The method is extremely versatile and can account for complex interface deformation as well as inter-phase exchanges. The sharp gradients in the fluid and flow properties,which occur at the interface between gas/liquid interfaces, however, mean that ill conditioned matrices must be solved repeatedly at each time step of the flow simulation. Another difficulty from a computational efficiency point of view, arises from the need to refine the underlying grid at the interface if all interface phenomenons are to be accurately modelled. These two issues mean that extremely efficient algebraic solvers must be used to model realistic problems. This project investigatesi nnovative parallel algorithms to achieve efficient solutions on the centre's High Performance Cluster. 

More detailed descriptions of both of these projects can be found on the website of the Energy and Environmental Flow Modelling Group in DCU.