Clustering based optimal sizing and placement of PV-DG using neural network

This paper presents clustering based optimum sizing and placement of PV Distributed Generation system on 20 kV network. The optimization objective is to minimize the total power losses on the entire network, but the treatment is conducted in smaller area, called cluster. In the system study a bus 30 of IEEE 20 kV distributionnetwork was divided into 3 clusters. A cluster consists of bus with similar voltage and losses performances. Neural Network was used to determine the member of a cluster. Bus with voltage 0.999 p.u to 1 p.u and network losses 29.2 kW was defined as the 1st group. The 2nd and 3rd group represents worse conditions. Then, an optimum size of PV system was injected to the optimum location in the worse group using NN. Simulation result shows that losses fell by 34.4% and voltage profile rises to 0.993 p.u to 1 p.u.