Optimal Location Identification of Solar PV Systems in Distributed Generators Based on Prediction of Load Flow and Factor Using Rule Based Deep Learning Algorithm

Optimal sizing and location identification for the installation of Solar Photovoltaic (SPV) sources in distributed generators (DG) is a challenging task. DGs supports the power grid and avoids the power loss due to increase in demand of electric power. In this paper, sizing and location of SPV are obtained based on microclimatic data, because DGs power generation output varies for different climatic condition. Traditional methods consider only electrical data such as load active and reactive powers for the Optimal sizing and location identification of SPV. This paper proposes Rule Based (RB) input data selection from microclimatic data such as temperature, wind speed, humidity, and total cloud cover, and electrical data such as load active and reactive power for the prediction of Optimal sizing and location identification of SPV. The selection of data varies based on agricultural area or residential area or commercial area. The rule based selected data are fed to the Long-Short Term Memory (LSTM) layer and Particle Swarm Optimization (PSO) algorithm for Optimal sizing and location identification of SPV. The proposed RB-PSO-LSTM method is applied in IEEE-32 bus and Indian Utility 62-bus systems. Objective function of RB-PSO-LSTM method is to minimize grid dependence, reduce power loss, maintain Voltage Stability Index and reduces the annual operating cost. Performance evaluation of the proposed RB-PSO-LSTM method is compared with existing optimization techniques. The minimal grid dependency is attained through proposed RB-PSO-LSTM method and reduces the running cost of about 85% of the Power grids.