Abstract
The performance of a power system can be measured and evaluated by its power flow analysis. Along with the penetration of renewable energies such as wind and solar, the power flow problem has become a complex optimization problem. In addition to this, constraint handling is another challenging task of this problem. The main critical problem of this dynamic power system having such variable energy sources is the intermittency of these VESs and complexity of constraint handling for a real-time optimal power flow (RT-OPF) problem. Therefore, optimal scheduling of generation sources with constraint satisfaction is the main goal of this study. Hence, a renewable energy forecasting–based, day-ahead dynamic optimal power flow (DA-DOPF) is presented in this paper with the forecasting of solar and wind patterns by using artificial neural networks. Moreover, contribution factors are calculated using triangular fuzzy membership function (T-FMF) in the sub-interval time slots. Furthermore, the superiority of feasible (SF) solution constraint handling approach is used to avoid the constraint violation of inequality constraints of optimal power flow. The IEEE 30-bus transmission network has been amended to integrate a solar photovoltaic and wind farm in different buses. In this approach, the computing program is based on MATPOWER which is a tool of MATLAB for load flow analysis which uses the Newton–Raphson technique because of its rapid convergence. Meteorological information has been gathered during the time frame January 1, 2015, to December 31, 2017, from Danyore Weather Station (DWS) at Hunza, Pakistan. A Levenberg–Marquardt calculation–based artificial neural network model is utilized to foresee the breeze speed and sunlight-based irradiance in light of its versatile nature. Finally, the results are discussed analytically to select the best generation schedule and control variable values.
Highlights
Optimal power flow (OPF) is an efficient and legitimately organized tool to control the settings of decision variables of the power system
The accuracy of the artificial neural networks (ANNs) models is analyzed based on the maximum prediction error (MPE) and root mean square error (RMSE)
Comparison of different algorithms suggests that the LM algorithm performs better than others on the input data; it is adopted for the training phase of the network
Summary
Optimal power flow (OPF) is an efficient and legitimately organized tool to control the settings of decision variables of the power system. A system operator cannot know the actual real-time (RT) conditions for a specified day-ahead (DA) schedule To solve this day-ahead dynamic optimal power flow (DA-DOPF) problem, an RTOPF problem is integrated with the ANN model to predict the wind speed and solar irradiance. The novelty of the proposed DA-DOPF approach is a two-stage optimization approach consisting of an LM algorithm–based ANN model and particle swarm optimization–based economic load dispatch along with OPF problem constraints with an effective constraint handling technique The former precisely generates the available power of VESs with respect to the wind speed and solar irradiance, and the latter provides the best-fit generation schedule of the generation sources along with the satisfaction of power system constraints. The main contribution of the present study is listed
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call