Nonlinear model predictive control of SI engine using discrete time identification model

摘要

Model predictive control(MPC) is a potential control technique for nonlinear systems. It can also control MIMO system and treat constraints easily. However MPC has a disadvantage which is a big computational cost. MPC is successfully applied in the chemical industry, where the sampling period is sufficiently large, e.g., several tens of second or longer. However, MPC is not suitable for mechanical systems controlled with a sampling period in the order of milliseconds. So, in this paper, a feedback compensation method for discrete nonlinear systems is proposed to overcome the problem. To verify the effectiveness of the proposed method, a discrete time identified SI engine model is used for a simulation, and it is shown that computational time of the proposed method is faster than a conventional method, i.e. C/GMRES.