Optimal integration of distributed generation (DG) resources in unbalanced distribution system considering uncertainty modelling

Abstract

The advancements in distributed generation (DG) technologies and growing concern for environmental friendly sources of energy necessitate an accurate analysis of distribution system with DG sources. The photovoltaic (PV) source is one of the most promising DG types that can be used for power generation at the distribution level because of its abundance. The efficient operation of the distribution system after the interconnection of DG sources is highly dependent on its integration to the distribution network. The uncertainty associated with the generation from the DG sources should also be considered while planning the integration. This paper presents the optimal sizing of the PV sources in the unbalanced distribution network by Reinforcement Learning, which is an efficient strategy for handling the stochastic data in practical situations. The uncertainty associated with the power output from the PV source is included in the power flow as a variable with multiple states. Here, the beta probability density function is used to model the randomness of the PV source. The seasonal variation in the power loss reduction obtained shows the effectiveness of the uncertainty model. The proposed algorithm is validated and tested for the Institute of Electrical and Electronics Engineers (IEEE) 13-bus and 37-bus distribution feeders, which shows its suitability for implementation in a real system. Copyright © 2016 John Wiley & Sons, Ltd.