抄録
An approach to Nonlinear Model Predictive Control(NMPC) is presented on the basis of the process models described by nonlinear differential equations in time-domain. Since NMPC is formulated in discrete-time space, the continuous process models must be converted to discrete-time equivalents by some approximation such as Taylor series expansion. In this paper, the continuous process models are converted to discrete-time equivalents using orthogonal collocation in order to reduce the truncation error. Simulation study shows that the performance of proposed control system is applicable to start-up of batch chemical process which shows nonlinear behaviour.
