Research | FLAME Lab, CDS, IISc

Asynchronous Numerical Schemes

Designing scalable CFD codes on massively parallel computers is a challenge. This is mainly due to the large number of communications between processing elements (PEs) and their synchronization, leading to idling of PEs.

Cavity Stabilized Flames

Flame stabilization in scramjet combustors used in hypersonic flight is a challenge. At the relevant flight conditions that result in short residence times of the order of 0.1 ms, flame stabilization is often achieved using a cavity flame holder that recirculates hot products of combustion including radicals.

Detection of Extreme Events

With the increasing availability of data, machine/deep learning methods are becoming an important tool for modeling, analysis and prediction of several phenomena. In this work, we develop a new anomaly detection method to identify anomalous/extreme events in scientific phenomena, which are often described by multi-scale, multi-variate, smoothly varying data.

Dimensionality Reduction for Reacting Flows using Machine learning

Reacting flows are characterized by a multitude of interacting variables and scales, ranging from molecular-level reactions to macroscopic flow structures involving many chemical species. High-fidelity simulations, such as Direct Numerical Simulations (DNS), capture intricate spatiotemporal variations but are computationally expensive. Dimensionality reduction techniques can alleviate computational costs by learning a low-dimensional manifold and evolving the solution in that space with reasonable accuracy.

Gas Turbine Engine Combustion

Staged gas turbines with two sequential combustion chambers are being developed for power generation for their ability to achieve low emissions within a wide operational range while conserving high thermal efficiency.

Reactive Molecular Dynamics

Chemical kinetic (CK) is a vital part of combustion system modeling. It provides reaction mechanisms and rate constants extracted from theoretical approaches. However, CK models are only valid in a specific range of temperatures and pressures and eventually fail at extreme conditions.

Super-resolution of Turbulent Fields

Deep learning has been extensively used for modeling and analyzing fluid turbulence. One such application is the use of super-resolution (SR) algorithms to reconstruct small-scale structures from large-scale ones for turbulent flows.