Nonlinear Parameter Identification Method And Its Application In Servo Drive System

With the rapid development of science and technology, the servo system has madea breakthrough progress in recent years. They can be widely used in different kinds offields, such as industrial control, aerospace, laser processing, household appliances.Because there exists nonlinear phenomena in practical application, containing dead zone,saturation, friction and interval. It is the key to choose a simple and effectivemathematical model to accurately describe servo system. To bring the nonlinearparameter identification into servo system will be a good solution to the problem thatlinear model can not solve, and therefore it has far-reaching significance for theidentification of nonlinear system.The structure of Hammeistein model is common model in nonlinear systems. Thispaper focuses on the parameters identification of nonlinear Hammeistein model, andapplies the model to servo system modeling. Initially, the least squares identificationalgorithm, stochastic gradient algorithm and iterative algorithm are derived for theequations error model and the output error model, according to key variables separationprinciple and auxiliary model ideas. Maeawhile, the effectiveness of various algorithmsfor the two models is compared and analyzed through simulation examples. Then thevoltage equation and torque equation of the servo system are elaborated in details,meanwhile, the linear model of servo system can be established, its input is load angularvelocity and output is control voltage. Owing to the nonlinear factors of friction anddead zone, the model of servo system contains nonlinear and linear element, its basicstructure is characterized by nonlinear and linear part in series, and the Hammersteinmodel of servo system is established and discretized. Analyzing the nonlinear systemidentification method by contrast, Finally, the least squares method and the forgettingfactor stochastic gradient algorithm are performed to estimate the dynamics modelparameters of servo system, furthermore, the identification results of linear model andHammerstein model are compared in the paper.The results show that the estimation error of servo system described byHammerstein model is lower than the linear model, that is to say, it has higher accuracyfor the identification of nonlinear model than the identification of linear model, it iseffective for Hammerstein model to describe the servo system, it not only improves the accuracy of the model but also accurately describes the dynamic behavior of the servosystem.