Learning Optimal Solutions for Extremely Fast AC Optimal Power Flow

  title={Learning Optimal Solutions for Extremely Fast AC Optimal Power Flow},
  author={Ahmed S. Zamzam and Kyri Baker},
  journal={2020 IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm)},
  • Ahmed S. ZamzamK. Baker
  • Published 27 September 2019
  • Computer Science
  • 2020 IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm)
We develop, in this paper, a machine learning approach to optimize the real-time operation of electric power grids. In particular, we learn feasible solutions to the AC optimal power flow (OPF) problem with negligible optimality gaps. The AC OPF problem aims at identifying optimal operational conditions of the power grids that minimize power losses and/or generation costs. Due to the computational challenges with solving this nonconvex problem, many efforts have focused on linearizing or… 

Figures and Tables from this paper

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

This paper proposes an integration of deep neural networks and Lagrangian duality to capture the physical and operational constraints of the AC Optimal Power Flow and produces highly accurate approximations whose costs are within 0.01% of optimality.

Learning to Solve AC Optimal Power Flow by Differentiating through Holomorphic Embeddings

This approach constitutes the first learning-based approach that successfully respects the full non-linear AC-OPF equations and reports a 12x increase in speed and a 40% increase in robustness compared to a traditional solver.

A learning-augmented approach for AC optimal power flow

Constraint-guided Deep Neural Network for solving Optimal Power Flow

Emulating AC OPF solvers for Obtaining Sub-second Feasible, Near-Optimal Solutions

A neural network is trained to emulate an iterative solver in order to cheaply and approximately iterate towards the optimum, and it is shown that the proposed method can solve “difficult” AC OPF solutions that cause DC-warm started algorithms to diverge.

Learning to Solve the AC Optimal Power Flow via a Lagrangian Approach

This work uses deep neural networks to learn the dual variables of the ACOPF problem and proposes a Lagrangian-based approach that can reach more globally optimal solutions with significant computational speedup even when the training data consists of mostly suboptimal solutions.

A Convex Neural Network Solver for DCOPF With Generalization Guarantees

This work proposes a novel algorithm for solving DCOPF that guarantees the generalization performance, and significantly outperforms other machine learning methods.

Machine Learning-Aided Security Constrained Optimal Power Flow

A learning augmented optimization approach is developed to solve the security-constrained optimal power flow (SCOPF) problem, using a multi-input multi-output random forest model to first solve network voltage magnitudes and angles of buses.

Learning to Optimize Power Distribution Grids using Sensitivity-Informed Deep Neural Networks

Numerical tests showcase that sensitivity-informed deep learning can enhance prediction accuracy in terms of mean square error (MSE) by 2-3 orders of magnitude at minimal computational overhead.

Model Information-Aided Deep Learning for Corrective AC Power flow

A novel model information-aided deep learning approach to approximate the corrective AC optimal power flow solution with proper model information with better performances in both training time and accuracy is presented.



Online optimization in closed loop on the power flow manifold

This work designs an adaptive feedback controller that steers the system in real time to the optimal operating point without explicitly solving an AC OPF problem and proposes a discrete-time projected gradient descent scheme on the power flow manifold (PFM).

Learning Warm-Start Points For Ac Optimal Power Flow

  • K. Baker
  • Computer Science
    2019 IEEE 29th International Workshop on Machine Learning for Signal Processing (MLSP)
  • 2019
A multi-target approach is utilized to learn approximate voltage and generation solutions to ACOPF problems directly by only using network loads, without the knowledge of other network parameters or the system topology.

Learning for DC-OPF: Classifying active sets using neural nets

This paper proposes the use of classification algorithms to learn the mapping between the uncertainty realization and the active set of constraints at optimality, thus further enhancing the computational efficiency of the real-time prediction of the optimal power flow.

Convexification of optimal power flow problem

  • J. LavaeiS. Low
  • Mathematics
    2010 48th Annual Allerton Conference on Communication, Control, and Computing (Allerton)
  • 2010
The optimal power flow (OPF) problem is nonconvex and generally hard to solve. We provide a sufficient condition under which the OPF problem is equivalent to a convex problem and therefore is

Machine Learning for AC Optimal Power Flow

This work presents two formulations of ACOPF as a machine learning problem: 1) an end-to-end prediction task where the optimal generator settings are predicted, and 2) a constraint predictiontask where the set of active constraints in the optimal solution are predicted.

Scalable Optimization Methods for Distribution Networks With High PV Integration

A suite of algorithms to determine the active- and reactive-power setpoints for photovoltaic inverters in distribution networks by leveraging a linear approximation of the algebraic power-flow equations and simplification from QCQP to a linearly constrained quadratic program is provided under certain conditions.

The Optimal Power Flow Operator: Theory and Computation

This work formalizes this operator theoretic approach to treating an OPF problem as an operator which maps user demand to generated power and allow the network parameters to take values in some admissible set, and defines and characterize a restricted parameter sets under which the mapping has a singleton output, independent binding constraints, and is differentiable.

Fast power system analysis via implicit linearization of the power flow manifold

  • S. BolognaniF. Dörfler
  • Computer Science, Engineering
    2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)
  • 2015
This paper derives the best linear approximant of such a relation around a generic solution of the power flow equations as an implicit algebraic relation between nodal voltages and nodal power injections.

Distributed Optimal Power Flow Using ADMM

  • T. Erseghe
  • Engineering
    IEEE Transactions on Power Systems
  • 2014
A fully distributed and robust algorithm for OPF is proposed which does not require any form of central coordination and is based upon the alternating direction multiplier method (ADMM).

Learning for Convex Optimization

This work considers the problem of using the information available through this solution process to directly learn the optimal solution as a function of the input parameters, thus reducing the need of solving computationally expensive large-scale parametric programs in real time.