ИСПОЛЬЗОВАНИЕ ЧЕТЫРЁХПОЛЮСНОГО ПРЕДСТАВЛЕНИЯ ФИЛЬТРА ПУАНКАРЕ-СТЕКЛОВА ДЛЯ ПОЛУНАТУРНОГО МОДЕЛИРОВАНИЕ НЕЛИНЕЙНЫХ СИСТЕМ

Abstract

The article shows the possibility of using the Poincare-Steklov filter to ensure the stability ofharware in the loop (HIL) simulation of nonlinear systems.HIL simulation involves splitting theinitial system into parts, with one part being modeled numerically on a computer, and the secondpart is represented by a real physical object. The parts of the system exchange data with eachother through a hardware-software interface, which can be implemented in different ways andshould ensure stability, as well as convergence of the results of HIL simulation to the results ofmodeling the original system. The variants of constructing software and hardware interfaces ITM,TLM, TFA, PCD, DIM, GCS and the Poincare-Steklov filter are described in the relevant literaturesources. The article shows how the original nonlinear system was divided into parts using thePoincare-Steklov filter, which, accordingly, led to the splitting into parts of the system of equationsdescribing the behavior of the original system. Next, it was shown how the values of the stabilizingparameters of the Poincare-Steklov filter were calculated and the systems of equations of the systemdivided into parts were corrected in accordance with the obtained values. At the next stage,the article presents the results of numerical modeling of the initial and partitioned system inMATLAB. When modeling in parts, the parts of the system exchanged data with each other at eachstep of the simulation only once with a delay of h. This method of numerical modeling of a systemdivided into parts is as close as possible to the processes occurring during semi-natural modelingof systems. A comparison of the obtained simulation results of the initial and the system dividedinto parts allowed us to conclude that the Poincare-Steklov filter, with the correct choice of thevalues of the stabilizing parameters, allows for the stability and convergence of the results of seminaturalmodeling of both linear and nonlinear systems