Abstract

When high levels of cooled external exhaust gas recirculation (EGR) are used to increase engine efficiency, the laminar flame speed during the flame initiation period is reduced and the combustion duration is elongated. The reduction in flame propagation speed together with the increase in cycle-to-cycle variability makes the combustion process prone to failure of flame initiation (misfire) or slow burning rates. A highly visited operating speed/load point is used to investigate the feedback control of spark advance (SA) and EGR valve. A decentralized proportional–integral (PI) controller and a centralized linear quadratic Gaussian (LQG) controller are designed to maintain a desired combustion initiation and duration that indirectly ensures a proper flame kernel initiation and flame propagation. A simple control-oriented combustion model is derived from system identification to perform simulation and linear analysis of the closed-loop (CL) system. Experimental validation of the controllers shows that the LQG performs better in transients and produces the least amount of cycle-to-cycle variability in CL operation. Analysis of the linearized system in frequency domain shows that a multivariable architecture is required to handle the input–output coupling efficiently. Sensitivity analysis of PI and LQG controllers under gain variability is reported to guide the tuning process, which also influences the cycle-to-cycle variability. The two-input two-output controllers designed in this paper are compared with a single-input single-output (SISO) controller designed to only adjust SA for the maximum brake torque. When the EGR valve is adjusted based on a lookup table, uncertainty in the EGR-rate will directly affect the combustion duration for which the SISO controller alone cannot handle. In conclusion, a coordinated spark and EGR controller based on the measured combustion features is deemed to be required for combustion control under diluted conditions.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call