Identification of Synchronous Generator Parameters Using Continuous ARX Model and Least Square Estimation

In this paper, least square estimation (LSE)-based synchronous generator model identification is carried out. The synchronous generator model including electromechanical dynamics and primary frequency control is first expressed by continuous-time AutoRegressive with eXogenous (ARX) models. The high-order derivatives in a continuous ARX model are approximated by expressions in terms of discrete-time data. The proposed approximation method (central difference approximation) is applied to second and third order derivatives. The continuous time model is now converted to a discrete time model. Generator parameters are then identified by LSE. We proposed two estimation structures to apply LSE. Benchmark model is simulated with and without input noise. The simulation data is used for estimation. It is found that one estimation structure is better than the other in terms of noise handling.