A/F ratio estimation and control based on induced engine roughness

2 versions of an engine air-to-fuel (A/F) ratio estimator were described and evaluated through transient engine testing. A four-parameter model was derived from steady-state identification and a ten-parameter model was derived from transient identification. Inputs to both estimators include a fundamental metric that relates the instantaneous change in engine speed to prescribed fuel injector pulse width modulation. Additional inputs include many of the standard engine state and control variables. The performance of the ten-parameter A/F model was found to be superior to the four-parameter model under cold engine testing. A closed-loop PID A/F ratio control system was described in which the A/F ratio estimator provided state feedback. Under limited testing conditions the ten-parameter estimator was shown to reduce open-loop A/F ratio excursions. This work demonstrates the feasibility of designing an A/F ratio estimator based on the response of engine speed to fuel pulse width modulation and incorporating that estimator in a closed loop A/F ratio feedback control system. This work opens up the possibility of including this type of A/F estimator into an A/F observer and of investigating alternate A/F estimator structures, such as those obtained from artificial neural networks.