Dispatch-Aware Deep Neural Network for Optimal Transmission Switching

TL;DR

A dispatch-aware deep neural network (DA-DNN) is proposed that accelerates DC-OTS without relying on pre-solved labels, eliminating costly OTS label generation that becomes impractical at scale.

Abstract

Optimal transmission switching (OTS) improves optimal power flow (OPF) by selectively opening transmission lines, but its mixed-integer formulation increases computational complexity, especially on large grids. To address this, we propose a dispatch-aware deep neural network (DA-DNN) that accelerates DC-OTS without relying on pre-solved labels, eliminating costly OTS label generation that becomes impractical at scale. DA-DNN predicts line states and passes them through an embedded differentiable DC-OPF layer, using the resulting generation cost as the loss function so that physical network constraints are enforced throughout training and inference. To stabilize training, we adopt a customized weight and bias initialization that keeps the embedded DC-OPF feasible from the first epoch. To improve inference robustness, we incorporate a binary regularization term that reduces ambiguity in the relaxed line-status outputs prior to thresholding. Once trained, DA-DNN produces a feasible topology and dispatch pair with highly predictable computation time comparable to a single DC-OPF solve, while conventional MIP solvers can become intractable. Moreover, the embedded OPF layer enables DA-DNN to generalize to untrained system configurations, such as changes in line flow limits, and to support post-contingency corrective operation. As a result, the proposed method captures the economic advantages of OTS while maintaining scalability and generalization ability.

Figures (6)

• Figure 1: Training process of the proposed DA-DNN for optimal transmission switching.
• Figure 2: Time varying MIP gap in solving 300 bus system (MIP-M).
• Figure 3: Distribution of predicted relaxed line-status $\hat{\mathbf{z}}$ for $\alpha = 0$ (blue, wider bars) and $\alpha = 10$ (red, narrower bars) in 300 bus system. Bars show the empirical ratio per bin: 0.0,...,0.9 denote intervals $[0.0,0.1)$,...,$[0.8, 0.9)$,$[0.9,\sigma(9))$ and $1.0$ is $\hat{z}_l\in[\sigma(9), 1)$.
• Figure 4: Distribution of the predicted relaxed line status values from untrained DA-DNN for 300 bus system with different weight and bias initialization (He: black, wider bars; Proposed: red, narrower bars). Bars show the empirical ratio per bin: 0.0,...,0.9 denote intervals $[0.0,0.1)$,...,$[0.8, 0.9)$,$[0.9,\sigma(9))$ and $1.0$ is $\hat{z}_l\in[\sigma(9), 1)$.
• Figure 5: Training curve in each test system.