Optimal Placement of Renewable Energy Sources Distributed Generation in an Unbalanced Network for Modern Grid Operations

Optimal placement of distributed generation (DG) is important in modern distribution network planning. Most DG placement models in the literature are based on normal state power loss and voltage profile enhancement rather than probabilistic security indices. In this paper, a decision tree (DT) classification approach for multiple renewable energy source DGs (RES-DGs) placement in an unbalanced distribution network using probabilistic security indices is proposed. The indices considered in this research are the feeders' risk index (RI) and power loss (PL). The optimal feeders for RES-DG placement are determined through the feeders' RI and PL while the optimal nodes are determined using the nodal hosting capacity (HC) and the critical voltage (
$V_{cr}$
) indices. The proposed technique was tested on a modified IEEE 33-node radial distribution network. After the optimal placement of the RES-DGs, the worst case of the average probabilistic voltage, steady-state power loss, and risk index obtained for the base case is improved from 0.93 pu, 144 kW, and 5.17 to 0.96 pu, 22 kW and 0.042, respectively. The results show that the proposed technique is effective to allocate multiple RES-DG for long-term distribution network planning and expansion, to increase the RES-DG penetration while minimizing the voltage risk.