Application of Machine Learning for Optimal Placement of Distributed Generation

Inappropriate placement of distributed generation (DG) can cause bus voltage profile deterioration and augmentation of active power loss due to high R/X ratio of the distribution lines. In this paper, a machine learning-based scheme is proposed for optimal positioning of DG units in the bus of a distribution network to lessen the active power losses and enhance the bus voltage profile in a noteworthy manner. Dynamic load condition of the network is prognosticated from a few foregoing years load characteristics using k- Nearest Neighbours (KNN) regression model of machine learning approach. The IEEE-14 test bus system is used to train with forecasted dynamic loads to recognize the consequent changes in bus voltage profile and line losses, from which optimal bus location of inserted DG is determined.