Optimal Selection of Filter-Compensation Devices Based on a Genetic Algorithm in Low-Voltage Distribution Power Networks

Abstract

Relevance: at present, the level of harmonic distortion in 0.4 kV low-voltage distribution networks is increasing
due to the growing share of nonlinear loads, including electronic power supplies of LED lighting devices,
switched-mode power supplies, frequency converters, and other electronic consumers. As a result, voltage
nonsinusoidality increases, the current in the neutral conductor rises, additional power losses occur in transform
ers and cable lines, and electric power quality indicators deteriorate. The use of conventional compensation de
vices does not always ensure effective performance and, in some cases, may lead to resonance phenomena.
Therefore, the optimal selection of filter-compensation devices for reducing harmonic distortion in 0.4 kV net
works is an urgent scientific and practical task.
Aim: to determine the optimal type, rating, and connection point of filter-compensation devices for reducing
harmonic distortion in 0.4 kV distribution networks, decreasing the total harmonic distortion of voltage, reducing
active power losses, and improving electric power quality indicators.
Methods: in the study, a mathematical model of a 0.4 kV radial distribution network was developed, taking into
account the harmonic composition of nonlinear loads. A genetic algorithm was used as the main method for solv
ing the problem of selecting filter-compensation devices. In this process, the device type, rating, and connection
point were optimized based on the objective function. The objective function included the total harmonic distor
tion of voltage, the share of individual harmonics, active power losses, and the current in the neutral conductor.
Results: the problem of reducing harmonic distortion in 0.4 kV distribution networks was analyzed using an
8-bus radial test system as an example. The results of assessing the influence of harmonic distortions caused by
nonlinear loads on the network operating parameters are presented. The possibility of improving electric power
quality indicators through the selection of filter-compensation devices based on a genetic algorithm was estab
lished.