Raden Sanggar Dewanto and Mohit Katraggada
Location: Stephenson Building, room F16 (first floor)
Time/Date: 3rd May 2012, 16:00 - 17:00
Raden Sanggar Dewanto: "Reliability Prediction of Single-Crystal Silicon MEMS using Dynamic Raman Spectroscopy"
The work proposes an extension and improvement to reliability predictions in single-crystal silicon MEMS by utilizing dynamic Raman spectroscopy to allow the gathering process of Weibull fracture test data to be done directly on devices thereby taking account of actual geometrical tolerances, dynamic load conditions and effects from the microfabrication process. A theoretical model is derived to calculate the probability of fracture of device during resonance and Raman spectroscopy is used to measure dynamically induced strains on microcantilever test structures. Acquiring this fracture data on devices will improve reliability prediction accuracy.
Mohit Katraggada: "Development of Unified Flame Surface Density Based Reaction Rate Models for Turbulent Premixed Combustion"
Designing of new environment-friendly and energy-efficient combustors is not only complex but also expensive in nature. Some of these costs can be averted through the use of high fidelity computer simulations. Reynolds averaged Navier-Stokes (RANS) simulations have been employed by industry in past two decades for design purposes, while large eddy simulation (LES) has only recently found its way into the practising engineer’s tool set. Both RANS simulations and LES face the common problem of having to deal with modelled governing equations.
The aim of this work was to develop models in context of LES and RANS for the filtered/mean reaction rate closure using the Flame Surface Density (FSD). The FSD represents flame surface area per unit volume and the closure of reaction rate translates to the modelling of FSD once the consumption of reactants per unit surface area is evaluated. It is possible to model the FSD using either an algebraic expression or by solving an additional transport equation alongside other governing equations. In the present analysis Direct Numerical Simulations (DNS) data has been used for the purpose of model development. All the necessary length scales and time scales of turbulence are appropriately resolved in DNS and thus can be treated as experimental data with resolution up to the Kolmogorov length scale. Here DNS data was explicitly Reynolds averaged/LES filtered to obtain the unclosed terms related to the FSD transport. The physical insight obtained from DNS was used to develop new FSD based models and validate them against the unclosed terms extracted from DNS data.
Based on this exercise, an algebraic FSD model was proposed for LES, which was shown to predict the FSD accurately for a wide range of combustion parameters. In order to model for the unclosed terms of the FSD transport equation, modelling was first accomplished in context of RANS, which were subsequently extended for LES.
Published: 13th April 2012