Optimal Power Flow with Renewable Energy Sources Using Improved Neural Network Algorithm

An optimal schedule is necessary for the efficient and economical operation of modern grids that integrate different power generation sources. Therefore, this study proposes an improved neural network algorithm (INNA) for optimal power flow (OPF) solution with penetration of stochastic wind and solar photovoltaic (SPV) energies. The OPF aims to minimize the total generation cost of conventional generators, windfarms, and SPV plant. Furthermore, emission costs are combined in the objective function of total cost in a specific case. To model uncertainties of renewable energy sources (RESs), lognormal and Weibull probability distribution functions are applied. Accordingly, the generation cost of RESs is divided into direct cost, penalty cost for underestimation, and reserve cost for overestimation. The developed INNA is tested on the IEEE 30 bus network, and its performance is compared with other approaches. Total generation costs obtained by INNA are 781.9496 $/h and 810.2765 $/h for Case 1 and Case 2, respectively. The findings show that INNA achieved a very promising solution for the OPF problem under study.