AEM faculty spotlight:
In the event of an emergency, a submarine may reverse its
propellers. While traditional methods can accurately predict the constant,
forward movement of fluid around this propeller, that sudden reversal creates a
mess beyond what many methods can handle. This is where Krishnan Mahesh, a
professor in the Department of Aerospace Engineering and Mechanics, comes in.
Utilizing numerical simulations, Mahesh can predict highly turbulent,
seemingly-random flow in exceedingly complex situations. These methods help the
navy design more efficient propellers, help industry better understand fuel
combustion in engines, and more.
Below, Professor Mahesh explains a bit about numerical simulations and how his simulation methods are impacting the field of fluid dynamics.
Firstly, what are numerical methods?
If you want to predict something, you have to have equations that describe its behavior. So if I were to place a leaf on a river and watch it, I notice it’s basically being carried by the flow. If I create an equation for that, it would be called a governing equation. To compute transport from one point to another, you lay a grid down and then numerically represent the governing equation on this grid. You then have numerical theory which lets you approximate the spatial gradients and integrate values over time, and this constitutes a numerical method.
What is different between your method and traditional methods of calculation?
Generally speaking, what we do is conduct research into turbulent flows with an emphasis on complex physics. Traditionally, high fidelity simulations for these problems have been restricted to simple geometries. We are now able to perform three dimensional, high fidelity simulations with practical applications. These models succeed where traditional models fail.
What are those sorts of problems?
As an example, in modern engines, if you want to reduce Nitrous oxide gases you use dilution jets to cool the combustion products. How do these jets mix with the combustion products? How does this process depend on the shape and distribution of the dilution holes, and how does the velocity of the jets affect this process? To better understand these issues, we simulate and study flows ranging from just a single jet to an entire combustor. In doing so, we both advance the state of the art in simulating the real-life engineering flow, as well as answer fundamental questions about the basic physics. What allows us to do this is the fidelity of our numerical tools. Other problems that we work on at present include: developing strategies to control turbulent mixing, modeling how a laser interacts with turbulent air, simulating how underwater flows cavitate, and shock waves and combustion in high speed flows. Our interest in high-speed flows is driven by scramjets.
What is difficult about flow in a scramjet engine?
If you look at a typical flow, it could be Mach 15 around the vehicle, but inside the scramjet it’s about one-third of that. The flow comes in, shock waves form, those shocks hit fluid near walls and when that happens, you have a mess. Similarly when you inject fuel into a high-speed airstream, the fuel stream breaks down, mixes and burns in a very complicated manner. What we are doing is developing our brand of numerical methods for high speed flows.
Why are you interested in turbulent flow?
I guess I am interested because of its presence everywhere; if you look at waves in the ocean, sandstorms, flow over airplane wings, sound produced by helicopter rotors, flow inside engines, it’s all turbulent flow. Seeing it everywhere and seeing that it is not a problem that’s going to go away anytime soon makes it both challenging and intriguing.
Is there one big question that needs solving in terms of turbulence?
I guess the one big question would be, 'is there a complete solution to the turbulence problem?’ Meaning, can we theoretically predict the instantaneous velocities, pressure, chemical species, temperature, etc. for any turbulent flow? We are so far away from this that the question almost seems philosophical. But what I do see as happening is that simulations are increasingly playing an important role in practical engineering. The question I ask is, “Can we develop simulation tools and flow models that will allow the scientist as well as the designer to develop next-generation devices and next-generation physical understanding?”
Last Modified: Tuesday, 22-Mar-2016 10:46:50 CDT -- this is in International Standard Date and Time Notation