IDENTIFICATION OF CONTINUOUS-TIME, LINEAR, AND NONLINEAR MODELS OF ANELECTROMECHANICAL ACTUATOR

A pragmatic approach to the identification of type 1, continuous-time, linear, and nonlinear models of dynamic systems is presented, This ty pe of model is common for electromechanical actuators, and the approac h is presented using the identification of models for a particular ele ctromechanical actuator. The experimental setup, model identification, parameter estimation, and model verification are discussed, The appro ach utilizes state-variable filters and least-squares in the parameter estimation procedure, The rationale for the estimation of parameters of a continuous-time model is given by contrasting it with the estimat ion of parameters of a discrete-time model, which is far more common a nd simple. Rationale is also given for the use of the simple least-squ ares algorithm rather than more complicated methods such as instrument al-variable methods, which may be more applicable in theory. The appli cation of the approach to the identification of both linear and nonlin ear models of the electromechanical actuator resulted in acceptable mo dels, as determined through verification procedures using several sets of experimental data in simulation, Although it would be impossible t o completely validate the approach for the nonlinear model, the result s for the specific case to which it was applied are encouraging.