Research

Aerospace

Before finance I was an aerospace engineer. I designed supersonic nozzles as an undergraduate, then did computational fluid dynamics at the University of Michigan. The habits carried over: model the system, check it against reality, and distrust any result you can't reproduce.

Projects

Preconditioning the Navier–Stokes equations
2010 – 2011

University of Michigan, guided by Prof. Bram van Leer

Solvers for the Navier–Stokes equations become really slow when the speeds in a flow differ by orders of magnitude, as they do at low Mach numbers or inside viscous boundary layers. Preconditioning rescales the equations so all the waves travel at comparable speeds and the solution converges to steady state much faster. I worked on these methods as a research assistant in van Leer’s group.

CFD analysis of hypersonic re-entry vehicles
2010

University of Michigan, guided by Prof. Iain Boyd and Dr. Alexandre Martin

When a vehicle re-enters the atmosphere at hypersonic speed, the air in front of it gets hot enough to break apart chemically. How much of that heat reaches the surface decides how the thermal protection system has to be designed.

I simulated that flow around a blunt sphere-cylinder body at about 5.5 km/s, roughly Mach 16, to calculate peak heat flux, heat load, temperature and pressure. The work used LeMANS, the hypersonic CFD code developed in Boyd’s Nonequilibrium Gas and Plasma Dynamics Laboratory, with a five-species model of air.

Most of the effort went into the mesh. I built and compared several grids, including ones aligned with the bow shock, and showed that refining the mesh at the shock and in the boundary layer moves the predicted peak heating towards the stagnation point, where theory says it should be.

Aerodynamic parameters for re-entry vehicles
2009

IIT Guwahati, B.Tech project with Ankit Upadhyay, guided by Prof. Vinayak Kulkarni and Prof. Anoop Kumar Dass

We built a toolkit of approximate methods that estimates drag, lift, center of pressure and stagnation point for supersonic and hypersonic flows. It returns an answer in seconds, where a 3-D CFD run takes hours, which makes it useful for optimizing both CFD runs and experiments.

Real-time control on small devices
2008

Tampere University of Technology, Finland, guided by Prof. Jose L. Martinez Lastra and Andrei Lobov

A research internship in the Factory Automation Systems and Technologies (FAST) Laboratory. I designed a service explorer that let resource-constrained devices control a production line in real time, built on Nokia’s Symbian operating system.

Designing a supersonic wind tunnel nozzle
2007

Indian Institute of Science (IISc), Bengaluru, guided by Dr. Shrinivasa Udipi and Dr. Srinivasa Patikallu

I wrote software that designs a wind tunnel nozzle for any requested test-section Mach number. It uses the method of characteristics, which traces Mach waves through the expanding flow and shapes the wall so they cancel, leaving a uniform, shock-free stream in the test section.

I also studied a sliding-barrier design that gets different Mach numbers and flow characteristics out of a single nozzle, instead of building a new nozzle for each test condition.

Published as “Method of Characteristics for a 2-D Supersonic Nozzle Design: A Programming Approach”, AIAA Student Journal, Fall 2007, Vol. 45, No. 3, pp. 1–10, and presented at a conference.

Degrees

University of Michigan
2009 – 2010

M.S.E., Aerospace Engineering

IIT Guwahati
2005 – 2009

B.Tech, Mechanical Engineering