Digital Twin-Empowered Voltage Control for Power Systems

TL;DR

Voltage control in distribution grids faces strong variability from loads and DERs, challenging traditional DT approaches due to high computational and sampling demands. The paper introduces GC-DT, a digital twin framework that combines a Gumbel-top sampling-based policy improvement with a consistency loss across latent states, integrated into a three-model DT (transformation, dynamic, prediction). Empirical results on IEEE 123-bus, 34-bus, and 13-bus systems show GC-DT achieving higher control rewards with fewer simulations and faster sampling, outperforming the state-of-the-art DT method. This approach offers a scalable, real-time-ready solution for voltage management in complex power systems.

Abstract

Emerging digital twin technology has the potential to revolutionize voltage control in power systems. However, the state-of-the-art digital twin method suffers from low computational and sampling efficiency, which hinders its applications. To address this issue, we propose a Gumbel-Consistency Digital Twin (GC-DT) method that enhances voltage control with improved computational and sampling efficiency. First, the proposed method incorporates a Gumbel-based strategy improvement that leverages the Gumbel-top trick to enhance non-repetitive sampling actions and reduce the reliance on Monte Carlo Tree Search simulations, thereby improving computational efficiency. Second, a consistency loss function aligns predicted hidden states with actual hidden states in the latent space, which increases both prediction accuracy and sampling efficiency. Experiments on IEEE 123-bus, 34-bus, and 13-bus systems demonstrate that the proposed GC-DT outperforms the state-of-the-art DT method in both computational and sampling efficiency.

Figures (1)

• Figure 1: Comparisons of the GC-DT and the original DT with $n=16$ in IEEE system. (a) IEEE 123-bus. (b) IEEE 34-bus. c) IEEE 13-bus.