Learning to Solve the AC-OPF Using Sensitivity-Informed Deep Neural Networks
Learning to Solve the AC-OPF Using Sensitivity-Informed Deep Neural Networks
复制标题
DOI：
10.1109/tpwrs.2021.3127189
复制
发表时间：
2021-03
影响因子：
6.6
通讯作者：
M. Singh;V. Kekatos;G. Giannakis
中科院分区：
文献类型：
--
作者：
M. Singh;V. Kekatos;G. Giannakis
为了将延迟关键型电力系统应用中的计算负担从实时转移到离线，最近的研究提出了使用深度神经网络（DNN）来预测出现负载需求时的交流最优潮流（AC-OPF）的解决方案的想法。由于网络拓扑可能会发生变化，因此必须以样本有效的方式训练该 DNN。为了提高数据效率，这项工作利用了 OPF 数据不是简单的训练标签，而是构成参数优化问题的解决方案这一事实。因此，我们主张训练敏感度通知的 DNN (SI-DNN)，不仅匹配 OPF 优化器，还匹配它们相对于 OPF 参数（负载）的偏导数。结果表明，在温和条件下，所需的雅可比矩阵确实存在，并且可以很容易地从相关的原始/对偶解中计算出来。所提出的 SI-DNN 与多种 OPF 求解器兼容，包括非凸二次约束二次规划 (QCQP)、半定规划 (SDP) 松弛和 MATPOWER；而 SI-DNN 可以无缝集成到其他学习到 OPF 方案中。对三个基准电力系统的数值测试证实了 SI-DNN 相对于传统训练的 DNN 预测的 OPF 解决方案的高级泛化和约束满足能力，特别是在低数据设置中。
To shift the computational burden from real-time to offline in delay-critical power systems applications, recent works entertain the idea of using a deep neural network (DNN) to predict the solutions of the AC optimal power flow (AC-OPF) once presented load demands. As network topologies may change, training this DNN in a sample-efficient manner becomes a necessity. To improve data efficiency, this work utilizes the fact OPF data are not simple training labels, but constitute the solutions of a parametric optimization problem. We thus advocate training a sensitivity-informed DNN (SI-DNN) to match not only the OPF optimizers, but also their partial derivatives with respect to the OPF parameters (loads). It is shown that the required Jacobian matrices do exist under mild conditions, and can be readily computed from the related primal/dual solutions. The proposed SI-DNN is compatible with a broad range of OPF solvers, including a non-convex quadratically constrained quadratic program (QCQP), its semidefinite program (SDP) relaxation, and MATPOWER; while SI-DNN can be seamlessly integrated in other learning-to-OPF schemes. Numerical tests on three benchmark power systems corroborate the advanced generalization and constraint satisfaction capabilities for the OPF solutions predicted by an SI-DNN over a conventionally trained DNN, especially in low-data setups.