Original scientific paper

IDENTIFICATION AND SIMULATION MODELS OF OPERATING SYSTEMS BASED ON ARTIFICIAL NEURAL NETWORKS

Marko Valčić
Julije Skenderović

Abstract

The identification and simulation results of dynamic and static operating systems significantly depend upon the quality and choice of input parameters. The paper presents a generalised identification and simulation model of an operating system dependant on different classes of parameters based on a generalised regressive neural network (GRNN). In addition the iterative procedure model is proposed here which, in virtue of the probability neuron network (PNN), makes it possible to effect efficiency assessment of the results developed as GRNN network responses. Both models have been tested on system parameters for the control and regulation of steam turbine installations utilising for the purpose the software package MATLAB 7.0.1.

Keywords

artificial neural networks; identification; simulation; classification; , steam turbines

Hrčak ID:

3945

URI

https://hrcak.srce.hr/3945

Publication date:

1.12.2005.

Article data in other languages: croatian

Visits: 2.179 *