Nonlinear Model-Based Multivariable Control for Air & Charging System of Diesel Engine with Short and Long Route EGR Valves

The objective of this study is to investigate a nonlinear model-based multivariable (MIMO, Multi Input Multi Output) technique to decouple actuators interaction and to reduce the calibration effort, while increasing control performances, above all in transient conditions, and robustness with respect to model uncertainties and system parameter variations. The presented control technique is based on the development of a nonlinear dynamical physical model of the diesel air and charging system. Feedback Linearization control is then applied to decouple actuators’ interactions and compensate for nonlinearities. A new set of virtual inputs are defined inverting the system differential equations. Relation among the new virtual inputs and the outputs is purely linear and decoupled, meaning that each virtual input affects linearly only one output. Moreover, a linear control block is added to guarantee transient and steady state performances and closed loop robustness. The proposed control approach has been validated through small diesel engine dyno and vehicle activities. Transient test bench maneuvers show that the control is able to coordinate the actuators in order to fulfill the targets and to guarantee similar performances in different operating points. In addition the robustness to environmental changes has been demonstrated by vehicle tests at different ambient conditions.