Better fuel economy, reduced exhaust emissions and better drivability strongly depend on precise control of air-fuel ratio (AFR) during both steady and transient engine operations.
A discrete, nonlinear fuel-injected SI engine model was developed and used for the design of AFR control algorithms. The engine model includes intake manifold air dynamics, fuel wall-wetting dynamics, and cycle delays inherent in the four-stroke engine processes. The sampling period is synchronous with crank angle (“event-based”) as opposed to the conventional time synchronous sampling scheme (“time-based”). The model was validated with test data over a wide range of engine operating conditions.
The exhaust O2 sensor can only provide a delayed and lagged AFR signal to the controller. This inherent delay in the measurement will slow down the system response if conventional feedback control design is used. Modern estimation theory propagates the engine model as an embedded observer to obtain an instantaneous estimate of the AFR and allows high bandwidth closed-loop control of AFR.
Precise transient AFR control requires precise knowledge of the air mass inducted which depends on the motion of the throttle with respect to the intake event. A drive-by-wire throttle was instrumented to be synchronous with the intake event. With the sampling period corresponding to every intake stroke, cycle-to-cycle AFR control could be achieved.
A laboratory port fuel-injected single cylinder CFR engine with drive-by-wire throttle was used to demonstrate the model-based AFR control structure. Air-fuel ratio followed commanded stoichiometric value within 0.5% rms during throttle transients at various operating points.