Learning Sequential Distribution System Restoration via Graph-Reinforcement Learning

Zhao, Tianqiao; Wang, Jianhui

doi:10.1109/tpwrs.2021.3102870

Cited by 40 publications

(25 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Besides, Ref. [19] examined a graph-reinforcement learning technique to link the power system topology with a graph convolutional network, which captures the complex mechanism of network restoration in power networks and understands the mutual interactions among controllable devices.…”

Section: B Literature Reviewmentioning

confidence: 99%

A Parallel Fast-Track Service Restoration Strategy Relying on Sectionalized Interdependent Power-Gas Distribution Systems

Gazijahani

Salehi

Shafie‐khah

2023

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

In the distribution networks, catastrophic events especially those caused by natural disasters can result in extensive damage that ordinarily needs a wide range of components to be repaired for keeping the lights on. Since the recovery of system is not technically feasible before making compulsory repairs, the predictive scheduling of available repair crews and black start resources not only minimizes the customer downtime but also speeds up the restoration process. To do so, this paper proposes a novel three-stage buildup restoration planning strategy to combine and coordinate repair crew dispatch problem for the interdependent power and natural gas systems with the primary objective of resiliency enhancement. In the proposed model, the system is sectionalized into autonomous subsystems (i.e., microgrid) with multiple energy resources, and then concurrently restored in parallel considering cold load pick-up conditions. Besides, topology refurbishment and intentional microgrid islanding along with energy storages are applied as remedial actions to further improve the resilience of interdependent systems while unpredicted uncertainties are addressed through stochastic/IGDT method. The theoretical and practical implications of the proposed framework push the research frontier of distribution restoration schemes, while its flexibility and generality support application to various extreme weather incidents.

show abstract

Section: B Literature Reviewmentioning

confidence: 99%

A Parallel Fast-Track Service Restoration Strategy Relying on Sectionalized Interdependent Power-Gas Distribution Systems

Gazijahani

Salehi

Shafie‐khah

2023

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

show abstract

“…Unfortunately, this approach still suffers from poor efficiency and presents convergence issues [1]. Very recently, [19,20] considered the graph correlation of voltage phasors in their DRL design. However, the authors ignored the temporal correlation of their time series.…”

Section: B Related Workmentioning

confidence: 99%

“…However, the authors ignored the temporal correlation of their time series. Note that [19,20] require the full system state.…”

Section: B Related Workmentioning

confidence: 99%

Deep Reinforcement Learning with Graph ConvNets for Distribution Network Voltage Control

Wu¹,

Carreño²,

Scaglione³

et al. 2022

Preprint

View full text Add to dashboard Cite

This paper proposes a model-free Volt-VAR control (VVC) algorithm via the spatio-temporal graph ConvNet-based deep reinforcement learning (STGCN-DRL) framework, whose goal is to control smart inverters in an unbalanced distribution system. We first identify the graph shift operator (GSO) based on the power flow equations. Then, we develop a spatiotemporal graph ConvNet (STGCN), testing both recurrent graph ConvNets (RGCN) and convolutional graph ConvNets (CGCN) architectures, aimed at capturing the spatiotemporal correlation of voltage phasors. The STGCN layer performs the feature extraction task for the policy function and the value function of the reinforcement learning architecture, and then we utilize the proximal policy optimization (PPO) to search the action spaces for an optimum policy function and to approximate an optimum value function. We further utilize the low-pass property of voltage graph signal to introduce an GCN architecture for the the policy whose input is a decimated state vector, i.e. a partial observation. Case studies on the unbalanced 123-bus systems validate the excellent performance of the proposed method in mitigating instabilities and maintaining nodal voltage profiles within a desirable range.

show abstract

“…Alternatively, researchers have also tackled the VVC problem by formulating it as multi-agent reinforcement learning (MARL) problem and proposing a novel efficient and resilient MARL algorithm [Gao et al, 2021]. Additionally, a more recent and closely related work in terms of methodology by [Zhao and Wang, 2021] also proposed to combine RL with graph neural networks for power system restoration via a multi-agent formulation. Inspired by these related works, we apply the idea of graph-based RL to solve the volt-var control problem using a graph representation as a complementary extension to the works mentioned above.…”

Section: Related Workmentioning

confidence: 99%

A Graph Policy Network Approach for Volt-Var Control in Power Distribution Systems

Lee¹,

Sarkar²,

Wang³

2021

Preprint

View full text Add to dashboard Cite

Volt-var control (VVC) is the problem of operating power distribution systems within healthy regimes by controlling actuators in power systems. Existing works have mostly adopted the conventional routine of representing the power systems (a graph with tree topology) as vectors to train deep reinforcement learning (RL) policies. We propose a framework that combines RL with graph neural networks and study the benefits and limitations of graph-based policy in the VVC setting. Our results show that graph-based policies converge to the same rewards asymptotically however at a slower rate when compared to vector representation counterpart. We conduct further analysis on the impact of both observations and actions: on the observation end, we examine the robustness of graph-based policy on two typical data acquisition errors in power systems, namely sensor communication failure and measurement misalignment. On the action end, we show that actuators have various impacts on the system, thus using a graph representation induced by power systems topology may not be the optimal choice. In the end, we conduct a case study to demonstrate that the choice of readout function architecture and graph augmentation can further improve training performance and robustness.

show abstract

Learning Sequential Distribution System Restoration via Graph-Reinforcement Learning

Cited by 40 publications

References 31 publications

A Parallel Fast-Track Service Restoration Strategy Relying on Sectionalized Interdependent Power-Gas Distribution Systems

A Parallel Fast-Track Service Restoration Strategy Relying on Sectionalized Interdependent Power-Gas Distribution Systems

Deep Reinforcement Learning with Graph ConvNets for Distribution Network Voltage Control

A Graph Policy Network Approach for Volt-Var Control in Power Distribution Systems

Contact Info

Product

Resources

About