Institute of Fluid Dynamics

Search

en

Research Group Prof. Kleiser (Emeritus)

The group of Prof. Kleiser was concerned with computational fluid dynamics (CFD). The main research interests lay in the elucidation of fundamental flow phenomena by advanced theoretical and numerical methods. Transitional, turbulent and particle-laden multiphase flows, biomedical flows and aeroacoustic noise were of particular interest. New accurate and efficient simulation methods were developed simultaneously for this purpose. A major part of the computational studies involved Direct Numerical Simulations (DNS) and Large Eddy Simulations (LES). The LES approach developed in the group has been transferred into an industrial CFD code.

The DNS approach to turbulent flow computation resolves all relevant space and time scales and permits high-fidelity flow simulations without using any turbulence model assumptions. However, the application of DNS is presently restricted to prototype flows at low and moderate Reynolds numbers due to the wide range of scales present in turbulent flows and the corresponding immense numerical resolution requirements.

In the LES approach, only the large scales are resolved, while the effect of the small fluctuations is accounted for by a subgrid turbulence model. Computationally, LES can be much less expensive than DNS (e.g. of order 1%), enabling also the simulation of turbulent flows at high Reynolds numbers.

In contrast, the conventional approach to turbulent flow modeling and the "workhorse" of industrial computations still is the solution of the Reynolds-averaged Navier-Stokes (RANS) equations, where a full-scale, heuristic turbulence model is required to complete the mathematical problem description.

Thus, deficiencies of currently existing turbulence models remain a key problem of Computational Fluid Dynamics (CFD). DNS high resolution flow data bases can be exploited in this context to perform numerical flow experiments and to validate improved models.

Publications

Complete list

Research Areas and Projects

Transition to turbulence is and remains to be one of the most intriguing problems of fluid dynamics. It is the process that renders a laminar flow to a turbulent flow. Transition processes can be observed in many applications, e.g., the transition in a boundary layer on a wing. Today’s knowledge on the transition process is still too limited to make accurate predictions of transition locations (information that is required to predict the skin friction).

We are studying the transition to turbulence on the example of canonical flows (e.g. channel flow, flat-plate boundary layer, etc.). This allows us to investigate the transition systematically and to reveal underlying mechanisms that govern transition processes in general.

The prediction of aeroacoustic noise is a problem of contemporary interest. In particular, the noise emission of aircraft at take-off and landing remains a hot topic in many public discussions. A lot of the emitted noise is created in the jet wake behind aircraft engines.

This projects aim at the prediction of aeroacoustic jet noise. To this end, we develop and use various numerical methods for computational aeroacoustics, such as DNS, LES and acoustic far-field solvers.

Wall-bounded turbulent flows is a large class of flows that can be found in most engineering applications. The appropriate subgrid-scale modelling under the influence walls is a difficult task which many subgrid-scale models fail.

We are studying a special class of disperse two-phase flows, in which the solid particles are much smaller than the smallest relevant scale of the fluid motion. This allows us to model the particles as infinitely small point forces. In the numerical simulations we solve the fluid equations in an Eulerian framework, whereas the particles are tracked individually along their trajectories.

The study of fluid motion in the inner ear is only one of many applications of fluid dynamics to the field of biology. Here, we are specifically looking at the flow of the endolymphatic fluid in the semicircular canals which host the balance sense (including diseases related to this flow, e.g. various forms of vertigo). In another project, we study the flow in the cochlea which is the sensory organ of our hearing sense.

This interdisciplinary work brings together clinical medical research, fundamental fluid dynamics, applied mathematics and modern numerical methods for particle-laden flows. The results of this work can be directly exploited by our collaboration partners from the University Hospital Zurich.

For all of our numerical investigations the availability of high-fidelity discretization methods, robust simulation tools and advanced post-processing and visualization tools is indispensable. Commercial CFD codes usually are much too inaccurate for our demanding research tasks. Therefore, we also develop highly-accurate and efficient numerical schemes, such as high-order compact upwind-biased finite-difference schemes or spectral-element methods. We implement them in research codes that run efficiently on a variety of computers, from workstations to modern supercomputing architectures.

JavaScript has been disabled in your browser