Time-Delay System Identification Using Genetic Algorithm - Part Two: FOPDT/SOPDT Model Approximation

Abstract The First-Order-Plus-Dead-Time (FOPDT) or Second-Order-Plus-Dead-Time (SOPDT) model approximation to a complicated process system can be carried out through either a kind of model reduction approach or a kind of system identification approach. This paper investigates this model approximation problem through an identification approach using the real coded Genetic Algorithm (GA). The desired FOPDT/SOPDT model is directly identified based on the measured system's input and output data. In order to evaluate the quality and performance of this GA-based approach, the proposed method is compared with two typical model reduction methods, namely Skogestad's rules and Sung et al method. The obtained results exhibit a very promising capability of GA in handling the data-driven time-delay system approximation.