Mark Subramanian

email: msubram@erc.wisc.edu 
HTTP: 

Research Title:
 

__

Gasoline direct injection is an emerging technology for achieving superior fuel economy in spark-ignition engines. At light load, the fuel is injected late in the cycle, creating a stratified charge in the vicinity of the spark plug that is suitable for stable ignition. This enables overall ultra-lean operation and the elimination of losses from throttling. The challenge is to control the pollutant emissions from GDI engines. High local temperatures and spray impingement are the primary causes of high nitrogen oxides and unburned hydrocarbons in the exhaust. As an added challenge, three-way catalysts for exhaust gas aftertreatment perform poorly amidst the lean conditions created by stratified operation. The current project will combine multidimensional modeling with a global search technique, known as a Genetic Algorithm. The goal is to optimize the operating parameters of the engine at a part-load condition. The operating parameters that will be subject to permutation include injection timing, spark timing, cone angle for the fuel spray, and duration of injection. Split injections will also be investigated by varying the time between injections and the percent of mass in each pulse. The merit function will be configured to select designs that minimize hydrocarbon and NOx emissions, while maximizing fuel efficiency and power. Also included in this function are provisions to keep the total wall heat transfer to a minimum.