# Graduate Seminar in Fluid Mechanics

When:
March 2, 2018 @ 4:00 pm – 5:00 pm
2018-03-02T16:00:00-05:00
2018-03-02T17:00:00-05:00
Where:
132 Gilman Hall

4:00 pm Presentation

“Cavitation Inception in Turbulent Shear Layers”

Presented by KARUNA AGARWAL (Adviser: Prof. Katz)

Cavitation in turbulent shear layers initiates along streamwise vortices. This has been argued to be the cause of Reynolds number dependence of the cavitation index. However, no volumetric pressure and flow-field measurements exist to explain this. Experiments to obtain tomographic PIV data downstream of a backward-facing step in the high speed water tunnel facility are planned. To characterize the turbulent boundary layer at the step, 2D PIV images are recorded. High speed images in wall-normal and spanwise planes are recorded to study the cavitating structures and find the conditions at which they first appear.  To better understand cavitation in turbulence, very high speed (5 MHz frame rate) holographic study of injected free stream nuclei will be performed.

4:25 pm Presentation

An Ensemble-Based Algorithm for Characterization of Scalar Sources in Turbulent Environment”

Presented by QI WANG (Adviser: Prof. Zaki)

An algorithm to determine the location and intensity of a scalar source with a parametrized shape is proposed and tested in a canonical turbulent channel flow at $Re_\tau = 180$. The algorithm uses forward simulations of an ensemble of scalar-source distributions, and can be easily applied to scenarios with a growing time horizon. The history of the scalar concentration at the sensor location due to the true source is compared with predictions from the ensemble members in order to determine the parameters of the source. Prediction errors are due to the approximation of the eigenvectors of the impulse-response matrix, or “eigen-sources”. In order to obtain a better approximation of the eigen-sources, a POD projection is used and is demonstrated to enhance the accuracy of the algorithm. The effect of measurement noise on the quality of reconstruction is quantified using the ratio of the standard deviation in the predicted source parameters and in the observation noise.  The results provide a measure of the difficulty of source reconstruction for different relative positioning of sources and sensors.