Model Based Nonlinear Predictive Control of IC Engine in Unsteady Operation Mode

摘要

Increasing demands for vehicle emission production limits with increasing requirements on the vehicle dynamic performance call for new, more complex engine control algorithms. Present ECUs use feedforward control algorithms with a large number of lookup tables that are based on working points and tuned in steady states with transient corrections used for engine unsteady operation. Feedback control is usually only used for certain quantities that are easy to measure. This paper is focused on predictive control based on a real-time predictive model of the engine, and thus allows using signals in the feedback loop that are not actually measured, just estimated by the predictive model. The predictive controller minimizes a criterion of optimality involving controlled variables, with respect to given limits of relevant quantities.