Date of Award

Fall 2022

Document Type

Thesis - Restricted

Degree Name

Master of Science (MS)


Mechanical Engineering

First Advisor

Dempsey, Adam B.

Second Advisor

Allen, Casey M.

Third Advisor

Roy, Somesh P.


With the projected decline of demand for gasoline in light duty engines and the advent of ethanol as a green fuel, the use of gasoline-ethanol blend fuels in heavy duty applications are being investigated where the primary mode of combustion is mixing controlled combustion. In mixing controlled combustion, a wide range of mixture conditions (equivalence ratio) exist inside the engine cylinder. The sooting tendencies of light fuels at richer conditions are not well understood. The goal of this research is to model the particulate matter emissions for gasoline/ethanol fuel blends, especially at fuel rich conditions. The computational models can then be used with in-cylinder conditions to predict the emissions characteristics of light fuel blends. The ethanol-gasoline blend fuels are modelled in a three-dimensional numerical simulation in a Rapid Compression Machine (RCM) using CONVERGE computational software. Several chemical kinetic mechanisms are used with SAGE chemistry solver and a RANS k-e turbulence model with a geometrically accurate sector model of the RCM including the creviced piston. The creviced piston is used in the experimental setup to reduce boundary layer effects and to maintain a homogeneous core in the reaction cylinder. The reaction mechanisms have been previously validated at engine like conditions for different fuel blends. Computational fluid dynamics simulations are conducted for different gasoline-ethanol fuel blends from E10 (10\% ethanol v/v) to E100. The fuel blend is modelled as a surrogate mixture of toluene, di-iso-butane, iso-octane, n-heptane for gasoline content, and ethanol. The computational results were validated against experimental results from an optical RCM using pressure measurements and laser extinction diagnostics. Different soot models and kinetic mechanisms are investigated to accurately predict the sooting tendencies of fuel blends, especially in richer conditions experienced during mixing-controlled combustion. The experimental combustion characteristics of different blends of fuel as well as soot generation are reasonably well predicted indicating that the use of the computational model can be extended to predict particulate matter emissions in heavy duty engine conditions.


Restricted Access Item

Having trouble?