Two air path observers for turbocharged SI engines with VCT

Nowadays, downsizing is a major way to reduce fuel consumption and pollutant emissions of spark ignition (SI) engines. In downsized engines, new air path management systems such as turbocharging or variable camshaft timing (VCT) are included, and an efficient control of the air actuators is required for engine torque control. Two non-linear estimators are proposed to estimate non-measured variables of the air path. The first one is an in-cylinder air mass observer that combines feedforward neural static models and a linear parameter varying (LPV) polytopic observer. The second one is a neural estimator of the burned gas and scavenged air masses. Test bench results on a turbocharged SI engine with VCT show the real time applicability and good performance of the proposed estimators. Finally, a strategy for developing the engine supervisor is presented.