USING OF ARTIFICIAL NEURAL NETWORK FOR THE IDENTIFICATION OF DYNAMIC PROPERTIES OF HYDRAULIC-PNEUMATIC SYSTEM

Abstract

The aim of the paper is to demonstrate using of artificial neural networks for the solution of practical problems of the identification of the complex non-linear systems' dynamic behavior. The mathematical model of the hydraulic-pneumatic system was investigated in order to build an alternative of this model, namely in the form of the artificial neural network (ANN). The model presents generally nonlinear multi-dimensional system with two inputs and two outputs. Input variables are flows through controlled pumps and output variables are water levels in the bottom tanks of the system. Both inputs and outputs of the system are represented as unified voltage signals. Solution of the problem consisted in the description of selected single dependences between particular input and output variables by means of ANN. For the problem solution Neural Network Toolbox was used a toolbox of the computing system MATLAB/SIMULINK.