Project title: EGR/Spark Timing control for internal combustion
engines via Extremum seeking
Xiaojian Yang
Abstract
Currently, most of the production EGR control systems are mainly open-loop. EGR
setting for any engine operating condition is first pre-determined off-line
through extensive engine calibrations based on emission and combustion
stability considerations and then implemented in real time as feedforward maps scheduled by engine operating conditions.
The lack of feedback correction mechanism necessitates allowing enough margins
at each operating point for guaranteed combustion stability at the expense of
comprising the full potential of EGR. Furthermore, Spark Timing (ST) is another
key variable affecting the combustion stability. As EGR is increased, spark
timing also needs to be advanced to allow extra time for combustion to grow and
achieve the Maximum Brake Torque (MBT) spark timing. Therefore, collaborative
use of EGR and spark timing will be beneficial for effective EGR management.
In the project, a closed loop EGR/ST control algorithm will be
proposed to address the stability/performance trade-off. In-cylinder-pressure
signals will be used as feedback signals to observe the effects of EGR on
combustion performance, and Coefficient of Variance (COV) will also be measured
by the in-cylinder-pressure. The lower value of COV the higher MBT the engine
can reach, so for a specified COV, as shown in the figure below, the optimal
state is always located at the tip of maximum EGR. An extremum
seeking algorithm will be employed to approach this optimal state for changing
engine work conditions.