eduzhai > Applied Sciences > Engineering >

High-Fidelity Machine Learning Approximations of Large-Scale Optimal Power Flow

  • Save

... pages left unread,continue reading

Document pages: 9 pages

Abstract: The AC Optimal Power Flow (AC-OPF) is a key building block in many powersystem applications. It determines generator setpoints at minimal cost thatmeet the power demands while satisfying the underlying physical and operationalconstraints. It is non-convex and NP-hard, and computationally challenging forlarge-scale power systems. Motivated by the increased stochasticity ingeneration schedules and increasing penetration of renewable sources, thispaper explores a deep learning approach to deliver highly efficient andaccurate approximations to the AC-OPF. In particular, the paper proposes anintegration of deep neural networks and Lagrangian duality to capture thephysical and operational constraints. The resulting model, called OPF-DNN, isevaluated on real case studies from the French transmission system, with up to3,400 buses and 4,500 lines. Computational results show that OPF-DNN produceshighly accurate AC-OPF approximations whose costs are within 0.01 ofoptimality. OPF-DNN generates, in milliseconds, solutions that capture theproblem constraints with high fidelity.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×