skip to main content
Caltech

GALCIT Colloquium - PhD Defense

Friday, May 17, 2019
3:00pm to 4:00pm
Add to Cal
Guggenheim 133 (Lees-Kubota Lecture Hall)
Using the force: Applications and implications of turbulence forcing terms in direct numerical simulations
Chandru Dhandapani, PhD Candidate, GALCIT, California Institute of Technology,

Most energy requirements of modern life can be fulfilled by renewable energy sources, but it is impossible in the near future to provide an alternative energy source to combustion for airplanes. However, combustion in aviation can be made more sustainable by using alternative jet fuels, which are made from renewable sources like agricultural wastes, solid wastes, oils, and sugars. These alternative fuels can be used in commercial flights only after a long certification process by the Federal Aviation Agency (FAA) and ASTM International, and in over 50 years of fuel research, only five fuels have been certified. This research project aims to speed up this certification process with quicker testing of alternative fuels. Engine testing and even laboratory testing require large amounts of time and fuel. Simulations can make the process much more efficient, but accurately simulating highly turbulent flames in such complex geometries would need large amounts of computational resources. The current research aims at creating an efficient computational framework, that can replicate different engine-like turbulent flow conditions in simple geometries with numerical tractability.

Direct Numerical Simulations (DNS) are performed on triply periodic cubic domains, and the statistically stationary shear turbulence, that is observed in free shear flows, is accurately captured. Turbulent flame simulations require a computational domain that is longer in the flame normal direction and an inflow/outflow in that direction for time-independent statistics. DNS of isotropic and shear turbulence are performed in long cuboidal domain of high aspect ratios, and the long domain slightly reduces the integral length scales, while the other statistics are unaffected. Simulations with inflow/outflow boundary conditions generate non-homogeneity in the turbulence and enable the inclusion of shear convection in the simulation. The shear convection improves the anisotropy results, which match better with experiments and simulations of free shear flows.

DNS of highly turbulent n-heptane air flames are performed under different flow conditions, while accurately resolving both the turbulence and chemistry. Turbulent flames involve two-way coupling between fluid mechanics and combustion. The effects of the flame on the turbulence and the impact of the turbulent flow conditions on the flame behavior are analyzed, namely the effects of turbulence production, strong pressure gradients, and shear convection on the flame. The anisotropy produced in the turbulence due to the different flow conditions and the flame are also compared and contrasted. While the global behavior and flow anisotropy were affected by these conditions, the local chemistry effects were unaffected, and depend only on the laminar flame properties and turbulence intensity. With this improved understanding, one can predict how alternative fuels will burn in jet engines, just from the fundamental knowledge of their chemistry, like chemical composition and structure. These predictions can find better fuels quicker and reduce CO2 emissions from airplanes.

For more information, please contact Wesley Yu by phone at (915) 309-7972 or by email at [email protected].