DeepOPF: A Deep Neural Network Approach for Security-Constrained DC Optimal Power Flow

We develop DeepOPF as a Deep Neural Network (DNN) approach for solving security-constrained direct current optimal power flow (SC-DCOPF) problems, which are critical reliable and cost-effective system operation. is inspired by the observation that SC-DCOPF problems given network equivalent to depicting high-dimensional mapping from load inputs generation phase angle outputs. first train DNN learn predict generations inputs. then directly reconstruct angles loads using equations. Such predict-and-reconstruct reduces dimension of learn, subsequently cutting down size amount training data needed. further derive condition tuning according desired approximation accuracy load-generation mapping. post-processing procedure based on $\ell _1$ -projection ensure feasibility obtained solution, can be independent interest. Simulation results IEEE test cases show generates feasible solutions with less than 0.2% optimality loss, while speeding up computation time two orders magnitude compared state-of-the-art solver.