Machine learning for more sustainable energy and transport technologies
Project Leader: Richard Sandberg
Collaborators: General Electric
Primary Contact: Richard Sandberg (email@example.com)
Keywords: computational fluid dynamics; energy efficiency; fluid dynamics; machine learning; turbulence
Disciplines: Mechanical Engineering
One of the greatest challenges to modern computational fluid dynamics is the accurate simulation of turbulent phenomena in complex environments. An example of critical importance to aviation are aero-engine compressors or turbines. In those components, the flows are transonic, highly turbulent and can involve significant levels of heat transfer. There is also a strong interaction of stationary and rotating blades which creates a significantly unsteady and non-uniform flow field. This is important because the unsteadiness and turbulence affects both the aerodynamic efficiency and the heat transfer from the gas to the blades which can shorten the blade life through thermal damage. To make aero-engines ever more efficient, there is a critical need from both industry and academia to further our understanding of heat-transfer and aerodynamic loss mechanisms.
The successful applicant will conduct high-fidelity simulations of turbomachinery flows using a well-established high-performance DNS/LES code. The data resulting from the massively parallel simulations will be used to a) investigate physical mechanisms that cannot be captured by traditional CFD approaches, b) train novel turbulence models with a unique machine-learning capability developed in Melbourne. Further, it is also expected that the computational capabilities of the flow solver are further extended.
The ideal candidate will have experience in performing fluid flow simulations and some knowledge of turbomachinery. High level computer skills, including MPI, OpenMP and OpenACC; and experience with various HPC platforms is highly desirable.
The applicants must have a background in Engineering or a relevant discipline. Applications from women are strongly encouraged.