Method and algorithms of nonlinear dynamic processes identification

Purpose: Increasing accuracy of dynamic models of the complex nonlinear processes for solution of the tasks for these processes control. Methodology: Structural-parametric identification of nonlinear dynamic processes including the identification of the model structure based on selection by non-shift criterion as well as self-reactance identification of the optimum structure model by the regularity criterion through the whole sampling of experimental data. Findings: The algorithms of global and local optimization of nonlinear dynamic process models realizing the procedure of structural-parametric identification by their structural and self-reactance optimization were developed, which allows getting the models of extra accuracy. Originality: The method of identification of nonlinear dynamic processes consisting of procedures for estimation of the state and characteristics of the process, as well as their structural-parametric identification was offered. It allows, unlike the known of methods, to fulfil the identification of these processes in the batch mode by structural-parametric and in real-time mode by self-reactance optimization of their models. Practical value: The results of the research can be used while developing algorithms for controlling complex nonlinear processes based on their complex estimation and identification. © Komiienko V.l., Matsiuk S.M., Udovyk I.M., Alekseiev O.M., 2016.