Узбекистан академия наук республики узбекистан

USING ARTIFICIAL NEURAL NETWORK TO CALCULATE STEADY STATE 
CONDITIONS 
 
Khudayarov M.B., Normamatov N.N. 
Tashkent State Technical University named after Islam Karimov 
 
Energy strategy is extremely important for developing countries. As the economy of these 
countries grow rapidly, their energy consumptions increase substantially. The power systems 
steady-state problems are described by a system of nonlinear equations, and for their solution are 
widely used iterative techniques such as the Newton-Raphson and others. Recently, techniques 
based on the use of genetic algorithms, the theory of fuzzy sets, artificial neural networks have 
been applied to solve this problem. In this article feedforward neural networks are used for 
calculating the steady-state regimes. The modeling results were obtained with the results of 
calculations using the Newton-Raphson method.
 
I.
INTRODUCTION. 
Electricity consumption prediction has been considered an effective measure that helps the 
power grid designers and planners build robust, adaptive, efficient, and economic smart grids.
It 
is aimed at modeling electricity consumption under different constraints along with environmental 
factors and the rules. A pre-estimated and calculated electricity demand can be obtained based on 
the history data including dates, economic, climate and so on. 
Among the factors that provoke difficulties in achieving this goal, the inherent variability 
of the load and the fast growth of the demand are foremost, followed by requirements of clean 
environment, weather, quality fuels, accelerated aging of the plants and fast changes in technology 
[1, 2]. 
Recently, promising Artificial Neural Networks (ANN) approaches have been developed 
to solve problems in power plants and power systems tuning of controllers, process identification, 
sensor validation, monitoring and fault diagnosis, in power plants, and security assessment, load 
identification, load modeling, forecasting and fault diagnosis, in power systems. 
The power systems steady-state calculations are a very urgent task of the electric power 
industry both at the design stage and during its operation. The accuracy and reliability of the results 
obtained ensure the correct functioning of the power systems [3, 4]. 
That is why the search for the most optimal methods for calculating steady-state regimes 
is an important task. From this point of view, a promising direction is associated with the use of 
modern intelligent data processing tools such as genetic algorithms, fuzzy sets, artificial neural 
networks and other. 
Given the noted in this article feedforward neural networks is used for calculating steady-
state regimes.