MSc internship/assignment: "Quantifying uncertainties in predictions of cavitation inception"
Publication dateApr 12, 2021
For MARIN Academy we are looking for a student for the following MSc internship/assignment:
Quantifying uncertainties in predictions of cavitation inception
The inception process is primarily driven by the difference between local pressure and saturated vapour pressure, with secondary effects arising due to gas nuclei concentration and properties in the water, presence of roughness or pits on solid surfaces, and turbulence. The main challenge of this work will be bringing together two methodologies which have previously not been used in tandem – uncertainty quantification and cavitation modelling. In UQ analysis of cavitation inception, we are interested in two types of uncertainties: input and discretization. Input uncertainty is related to the lack of information regarding an input parameter, and how this generates uncertainties in some output parameter of interest. Discretisation uncertainty is related to the fact that discretized forms of fluid flow equations are solved numerically on a computational grid, and it can be estimated with a systematic grid refinement study. Here, we are primarily interested in the combined effect of the two types of uncertainties, and this brings about a challenge to overcome. Carrying out this project will require solid skills in scripting languages (Python and bash), and at least basic knowledge of computational fluid dynamics and working on a Linux supercomputer. Both the CFD and UQ frameworks are readily available and fully-developed, but a rudimentary understanding of how they work will also need to be developed in order to use them effectively. Due to the novel nature of the project, final results should allow for a publication in either an international conference or a journal article.
Figure 1: Cavitation inception is a mechanism driven by gas nuclei entering low pressure regions and undergoing rapid growth. This micro-scale mechanism leads to macro-scale cavitation patterns that may be divided into different topological groups.
Uncertainty quantification analysis of flat plate and model-scale propeller using laminar-turbulent transition model in CFD was carried out previously at MARIN. In this work, experience was gained on how to build a surrogate model in CFD in a computationally efficient manner. A particularly challenging aspect of the work was to design a framework in order to study the combined effect of input and discretization uncertainties. The result of this work was presented in a conference in 2019 (see: https://www.marin.nl/publications/parameter-uncertainty-quantification-applied-to-the-duisburg-propeller-test-case), and journal article preparation is currently underway.
The internship is for the duration of 6 to 12 months. The start date of the internship period can be determined in consultation with the supervisors.
Master student in Naval Architecture, Aerospace Engineering, or other applied science or engineering.
Interest in CFD and hydrodynamics and numerical methods.
Basic programming skills in Python (or alternatively Matlab).
Interest in learning to work with high-performance computers in a Linux environment. Previous experience a plus but not necessary.
Department and supervisor
During the internship you will be connected with the department R&D of MARIN. The supervisor for this internship, Artur Lidtke, is working as Researcher at the R&D Department. Co-supervisors will be Bulent Duz, Douwe Rijpkema, also from the R&D department, and Chiara Wielgosz, a PhD student.