An artificial intelligence based method for evaluating power grid node importance using network embedding and support vector regression

Wang, Hui‐Fang; Zhang, Chenyu; Dongyang, Lin; Ben-teng, HE

doi:10.1631/fitee.1800146

Cited by 12 publications

(4 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Machine learning modeling relies on a sufficient and complete sample set that reflects the mapping relationship between the features and objective, as the short-circuit current changes with the operation modes and fault conditions. To obtain a sufficient sample set, the method used in [19] is adopted to generate different operation modes and fault conditions.…”

Section: Sample Set Establishmentmentioning

confidence: 99%

A Novel Machine Learning-Based Short-Circuit Current Prediction Method for Active Distribution Networks

et al. 2019

View full text Add to dashboard Cite

The traditional mechanism models used in short-circuit current calculations have shortcomings in terms of accuracy and speed for distribution systems with inverter-interfaced distributed generators (IIDGs). Faced with this issue, this paper proposes a novel data-driven short-circuit current prediction method for active distribution systems. This method can be used to accurately predict the short-circuit current flowing through a specified measurement point when a fault occurs at any position in the distribution network. By analyzing the features related to the short-circuit current in active distribution networks, feature combination is introduced to reflect the short-circuit current. Specifically, the short-circuit current where IIDGs are not connected into the system is treated as the key feature. The accuracy and efficiency of the proposed method are verified using the IEEE 34-node test system. The requirement of the sample sizes for distribution systems of different scale is further analyzed by using the additional IEEE 13-node and 69-node test systems. The applicability of the proposed method in large-scale distribution network with high penetration of IIDGs is verified as well.

show abstract

Section: Sample Set Establishmentmentioning

confidence: 99%

A Novel Machine Learning-Based Short-Circuit Current Prediction Method for Active Distribution Networks

et al. 2019

View full text Add to dashboard Cite

show abstract

“…In recent years, node importance identification has been useful in many applications across various domains. It allows us to identify the most influential users in social networks, to enable the better monitoring of public opinions and the forecasting of major events [9]; to locate crucial intersections or nodes in transportation networks, to guide the optimization of bus network services and improve the capacity of the public transport system [10]; and, in the realm of power networks, to prioritize the monitoring and maintenance of critical nodes, thereby ensuring the reliability and stability of the power grid [11]. Additionally, it is possible to explore the functionality and disease mechanisms of biological systems by recognizing genes or proteins that play pivotal roles in processes such as protein interaction networks and gene-regulatory networks [12,13].…”

Section: Introductionmentioning

confidence: 99%

Entropy-Based Node Importance Identification Method for Public Transportation Infrastructure Coupled Networks: A Case Study of Chengdu

Zeng,

Sun,

Zhang

2024

Entropy

View full text Add to dashboard Cite

Public transportation infrastructure is a typical, complex, coupled network that is usually composed of connected bus lines and subway networks. This study proposes an entropy-based node importance identification method for this type of coupled network that is helpful for the integrated planning of urban public transport and traffic flows, as well as enhancing network information dissemination and maintaining network resilience. The proposed method develops a systematic entropy-based metric based on five centrality metrics, namely the degree centrality (DC), betweenness centrality (BC), closeness centrality (CC), eigenvector centrality (EC), and clustering coefficient (CCO). It then identifies the most important nodes in the coupled networks by considering the information entropy of the nodes and their neighboring ones. To evaluate the performance of the proposed method, a bus–subway coupled network in Chengdu, containing 10,652 nodes and 15,476 edges, is employed as a case study. Four network resilience assessment metrics, namely the maximum connectivity coefficient (MCC), network efficiency (NE), susceptibility (S), and natural connectivity (NC), were used to conduct group experiments. The experimental results demonstrate the following: (1) the multi-functional fitting analysis improves the analytical accuracy by 30% as compared to fitting with power law functions only; (2) for both CC and CCO, the improved metric’s performance in important node identification is greatly improved, and it demonstrates good network resilience.

show abstract

“…At this moment, the principal component analysis (PCA) is usually employed in the feature extraction of nodes in power grids [17,18]. Since the PCA are suitable for the linear system, the improved PCA, namely kernel PCA [19], Nsytrom PCA [20], recurrence quantification analysis [21], are proposed to deal with the nonlinearity of the power grid [22][23][24]. Whereas nonlinear relationship between different nodes hasn't been described well using the improved PCA, and the nonlinear feature extraction method for nodes in DN need to be further investigated.…”

Section: Introductionmentioning

confidence: 99%

An Improved AP Clustering Algorithm Based Critical Nodes Identification for Distribution Network With High PV Penetration

Lin

et al. 2022

IEEE Access

View full text Add to dashboard Cite

The power flow of the distributed network (DN) with high penetration of distributed PV changes significantly, such as bidirectional power flow and larger deviation of volage magnitude, under the variation of PV power output, and the number and the position of critical nodes would change accordingly. Hence, a data-driven nodes clustering and critical node identification method is proposed in this paper. Since the relationship between different nodes are nonlinear, the autoencoder method is firstly applied to obtain the unified features of the nodes connected with different number of branches. Using the unified features, the Frechet distance is calculated to demonstrate the similarity between any two nodes. The improved affinity propagation (AP) clustering algorithm is also proposed to classify the nodes into different clusters, and the number the clusters and the critical nodes in different clusters could be obtained automatically. The electrical distance is considered in the improved AP clustering algorithm, so that nodes in different branches with similar features are avoided to be classified into one cluster. The simulations are carried out on the IEEE 123bus system and an actual DN system, the effectiveness and efficiency of proposed critical node identification method are verified.

show abstract

An artificial intelligence based method for evaluating power grid node importance using network embedding and support vector regression

Cited by 12 publications

References 25 publications

A Novel Machine Learning-Based Short-Circuit Current Prediction Method for Active Distribution Networks

A Novel Machine Learning-Based Short-Circuit Current Prediction Method for Active Distribution Networks

Entropy-Based Node Importance Identification Method for Public Transportation Infrastructure Coupled Networks: A Case Study of Chengdu

An Improved AP Clustering Algorithm Based Critical Nodes Identification for Distribution Network With High PV Penetration

Contact Info

Product

Resources

About