Impact of Photovoltaic Systems Placement, Sizing on Power Quality in Distribution Network

Abstract

The paper presents a Decision Tree Algorithm for impact of photovoltaic systems placement and sizing in order to improve power quality in distribution networks. The proposed approach is based primarily on ID3 and J48 Decision Tree Algorithms. With these algorithms, for different sizes of photovoltaic systems, can be defined the optimal location as well as minimal power losses of the distribution network. The simulations were carried out with real data obtained from the Kosovo Distribution Network and visualized with WEKA Toolbox. The obtained results compared with Genetic Algorithm MATLAB toolbox and DIGSILENT/Power factory software, prove that the Decision Tree Algorithm works well with an excellent and fast accuracy. The results from the application of the proposed method showed reduced power losses and optimal location in the distribution network confirmed method's validity. This approach can be used by engineers, electric utilities and distribution network operators for a quick decision with more efficient integration of new photovoltaic systems in the current distribution networks.

Keywords

• distributed power generation
• genetic algorithms
• photovoltaic systems
• power quality
• power systems