Offset-free model predictive control of diesel engine by combined design of disturbance model and observer

The diesel engine is a typical multi-input multi-output system with strong couplings, actuator constraints, and fast dynamics. The objective is to operate the engine to meet driver's speed demand and reduce exhaust emissions of nitrogen oxides (NOx) during transient process. Interactions between the actuators and nonlinear behavior of the system make the problem difficult to handle using classical control design methods. Instead, we propose an offsetfree model predictive control system. It is based on the combined design of integrating disturbance model and state observer. The dynamic observer is designed by solving an H-infinity control problem aimed at minimizing the effect of unmeasured disturbances and model mismatch on the output prediction error. The proposed approach has a low computation requirement and is suitable for implementation in the engine control unit on board. The application of this system in continuous working process by using dSPACE MicroAutoBox verified its feasibility and effectiveness for achieving precise reference tracking of engine speed and reducing emissions.