| Downsizing and turbocharging is an effective method for improving fuel economy, while meeting customer expectations for vehicle performance in a spark ignition engine. The turbocharger, however, introduces a physical feedback in the system that produces complex system interactions. This complicates development of engine control strategies, particularly those that regulate air to fuel ratio for emissions control and deliver driver demanded torque. Simple and efficient estimates of exhaust manifold pressure and turbocharger speed are helpful in meeting these control objectives.In this thesis, we explore model-based methods that produce computationally efficient estimates useful for on-line implementation. The foundation of the work is a nonlinear model of the system we develop for air path control development and analysis. Singular perturbation concepts are applied to a simplified version of this model to develop observers for exhaust manifold pressure and turbocharger speed that can be efficiently implemented in software. In addition, equilibrium analysis is used to develop static estimates that, when combined with lead compensation, produce computationally efficient algorithms that significantly reduce calibration effort. This approach provides a practical estimation strategy for real-time control. |
