Identification of the Physical Parameters of a Noisy Process in Discrete and Continuous-Time Domains

Abstract In recent years, techniques from the areas of system identification and parameter estimation have found extensive use in the modeling and control of power systems. From an applications view point, an important consideration arises for implementation with parallel versus serial data acquisition procedures. Factors such as cost, model accuracy, parameter value convergence, and so on dictate the necessity of careful investigations of such issues. This paper addresses the problem of the identification of physical parameters in noisy processes from both discrete and continuous time models. Simulation results are given for generalized least squares estimation, for both cases of parallel and serial data acquisition, and the effects of phase lag introduced are compared and analyzed. Finally, results are given for parameter estimation exercises with a single phase transformer