Transcripción del resumen del autor. The effects of the different air/fuel ratio (AFR) controller on lean limit and emission characteristics of a compressed natural gas (CNG) engine have been studied. The CARIMA model for the CNG engine was established, and a generalized predictive control (GPC)-based adaptive AFR controller was built by Matlab/Simulink. A simulation model of the engine was established by GT-Power. The experimental investigation was conducted on a six-cylinder multipoint injection lean-burn turbocharged spark-ignition (SI) CNG engine. The results of the simulation and experiment show that, in comparison to the proportional-integral-derivative (PID) algorithm, the GPC algorithm has not only better dynamic behavior but also better adaptability to the disturbances of the system, especially at transient conditions. Using the GPC controller, the lean misfire is improved because of the smaller overshoot of the excess air ratio and the operating range of the engine is extended. As a result, the emissions decrease slightly.

6