Engine High-Level Nonlinear Model Predictive Torque Control

In this article, a new nonlinear optimal methodology regarding transient torque control of spark-ignition (SI) engines is presented. The internal combustion engine is a highly nonlinear system, and transients are of great importance especially for hybrid vehicles as the engine is frequently started and loaded up to demanded torques. To address bandwidth requirements, purely feedback controllers are avoided, and a semi-feed-forward controller is incorporated into the high-level layer of the hierarchical structure. The controller is derived based on offline nonlinear model predictive control (MPC), minimizing the fuel consumption for optimal transition to demanded torques. The methodology can be extended to the whole desired engine torques.