When: Sep 25 2020 @ 4:00 PM
Where: https://wse.zoom.us/j/93762992307
https://wse.zoom.us/j/93762992307

Department of Mechanical Engineering
VIRTUAL GRADUATE SEMINAR IN FLUID MECHANICS
Join on-line via Zoom: https://wse.zoom.us/j/93762992307
Friday, September 25, 2020 | 4:00 p.m. – 5:00 p.m. (EDT)

“Statistically Constrained Neural Networks for Augmenting LES Wall Modeling”
Presented by YUE HAO
(Advisers: Profs. Tamer Zaki & Charles Meneveau)
The equilibrium wall model in large-eddy simulations (LES) is designed to predict the mean behavior but, applied to instantaneous realizations, it appreciably underpredicts the variance of the stress. We introduce a formalism whereby the equilibrium model provides a prior estimation of the stress, and a statistically constrained neural network (NN) provides a correction to that estimate. The network design is motivated by universal properties of the joint probability density functions of the local LES Reynolds number at the first grid point above the wall and the normalized instantaneous wall stress. Inputs and outputs of the network are normalized, conditioned on the estimate from the equilibrium wall model; and the loss function is designed to ensure the statistics of the corrected stress match the universal trends. Spatially filtered data from the JHTDB channel flow at $Re_{tau}$ =1000 and 5200 are used for training and testing. A priori tests are performed to assess the accuracy of the model relative to filtered wall stress from the database. The NN demonstrate better accuracy than the equilibrium wall model in (i) predicting statistics of the wall stress and (ii) the correlation of instantaneous predictions with the true filtered stress.

“Simulating in Vitro Models of Cardiovascular Fluid-Structure Interaction”
Presented by JAE HO “MIKE” LEE
(Adviser: Prof. Rajat Mittal)
Computer modeling and simulation (CM&S) in biomedical science is a powerful tool that can offer a cost- and time-efficient complement to traditional clinical diagnosis or medical devices testing. For example, CM&S can be used to study intraventricular fluid dynamics of the left ventricle (LV) of the heart to characterize heart function in health and disease. CM&S can also be used to assess the performance of medical devices such as bioprosthetic heart valves (BHVs). The immersed boundary (IB) method can facilitate models such as heart valves with very large structural deformations. My thesis describes work that advances the in vitro experimental validation of computational models of blood flow in the LV of the heart and through bioprosthetic heart valves. These models are based on a hyperelastic finite element extension of the the IB method for FSI. For the case of BHVs, we focus on porcine tissue and bovine pericardial BHVs, which are commonly used in surgical valve replacement. We compare our numerical simulations to experimental data from a customized flow circuit for the LV phantom and a commercial ViVitro pulse duplicator for BHVs. We then use this numerical platform to study the role of BHV geometry in leaflet dynamics and possibly leaflet durability. This approach has many potential applications, especially in developing a high-fidelity computational platform for designing and testing cardiovascular devices.