A neural net fitting-based approach for optimal power flow

Optimal power flow (OPF) is one of the most important management actions to achieve economical, safe and reliable operation of power systems. Require both high accuracy and speed of calculation so that appropriate corrective control actions can be implemented in real time. Classical methods such as interior-point method, sequential programming methods, as well as metaheuristic optimization methods used today require a lot of time to achieve satisfactory accuracy of OPF results. With the development of artificial intelligence and neural networks, it is possible to drastically reduce the OPF calculation time. This paper proposes the application of Neural Net Fitting (NNF) based approach for OPF. The basic idea of this work is to use a combination of metaheuristic optimization methods (MOA) and the NNF technique to realize a kind of hybrid method-network, which would solve the OPF problem in a very efficient and fast way, in a way suitable for real-time application. The proposed approach is tested on the standard IEEE 30-bus test system. The simulation results show that the application of the proposed NNF technique yields the same accuracy OPF results as the application of the classical metaheuristic methods PSOGSA or GA with the hybrid fmincon interior-point function in MATLAB^®, but with significantly less computation time.