HEATING OF TRANSFORMER - PROCESS IDENTIFICATION BY USING ARX MODEL

Description

Abstract (Serbian)

By simulating the appropriate power distribution network operating points the analysis of their exploitation
capabilities is significantly improved. Beside the methodology used in the analyses, the greatest impact to the
quality of results obtained has the accuracy of the model itself, or in other words it is necessary to respect the
real driving conditions of each element being modeled.
Since extreme operating conditions such as minimum and maximum network load are usually modeled, when
calculating power flow and voltage distributon across the network, it is important to take into account the
ambient temperature influence and the operating temperature of the network elements. For transmission lines,
ambient temperature and load heating are usually taken into account by calculating electrical resistance for each
type of conductor. Due to the complexity of the construction and non-linearity of the processes that take place
during the heating and cooling of the transformer windings, it is necessary to define the appropriate practical
thermal model for power transformers.
Since the most of the power transformer stations in the distribution network are integrated into the some
distributed remote measurement and control system, electrical values sampled after fifteen minute interval as
well as the ambient temperature, transformer windings and oil temperature are readily available. Using these data
and developed software tools to identify the process, a simple practical thermal models of power transformers
can be identifed, without performing a detailed power transformer nonlinear thermal analysis.
For the purpose of analysis whose results are presented in this paper, the ARX mathematical model and software
tool for identification of the processes of the software package MATLAB are used.