Nonlinear model predictive controller design based on learning model for turbocharged gasoline engine of passenger vehicle

In this paper, a neural-network-based nonlinear model predictive control (NMPC) scheme is investigated to realize coordinated control over the throttle and wastegate of a turbocharged gasoline engine of a passenger vehicle. First, due to the presence of MAPs and the complex structure of the turbocharged engine, establishing a mechanism model for controller design is very complicated. Benefiting from a large amount of experimental data, a predictive model is learned by a neural network to predict the future dynamics of the engine air-path system, and the accuracy of this model is verified. Second, to address the system constraints and coupling, a nonlinear model predictive controller is proposed to track the desired intake manifold pressure and boost pressure for meeting the engine torque demand. Third, quantum-behaved particle swarm optimization (QPSO) is applied for optimization of the NMPC objective function to obtain a more accurate solution. Finally, the performance of the control system is tested using the commercial simulation software AMESim.