The cognitive smart grid (SG) communication paradigm aims to mitigate quality of service (QoS) issues in obsolete communication architecture associated with the conventional electrical grid. This paradigm entails the integration of advanced information and communication technologies (ICTs) into power grids, enabling a two-way flow of information. However, due to the exponential increase in wireless applications and services, also driven by the deployment of the Internet of Things (IoT) smart devices, SG communication systems are expected to handle large volumes of data. As a result, the operation of SG networks is confronted with the major challenge of managing and processing data in a reliable and secure manner. The existing works in the literature proposed architectures with the objective to mitigate the underlying QoS issues such as latency, bandwidth, data congestion, energy efficiency, etc. In addition, a variety of communication technologies have been analyzed for their capacity to support stringent QoS requirements for diverse SGs environments. This notwithstanding, a standard architecture designed to mitigate the aforementioned issues for SG networks remains a work-in-progress. The main objective of this paper is to investigate the emerging technologies such as cognitive radio networks (CRNs) as part of the Fifth-Generation (5G) mobile technology for reliable communication in SG networks. Furthermore, a hybrid architecture based on the combination of fog computing and cloud computing is proposed. In this architecture, real-time latency-sensitive information is given high priority, with fog edge based servers deployed in close proximity to home area networks (HANs) for preprocessing and analyzing of information collected from smart IoT devices. In comparison to the recent works in the literature, which are mainly based on CRNs and 5G separately, the proposed architecture in this paper incorporates the combination of CRNs and 5G for reliable and efficient communication in SG networks.

The elbow technique is a well-known method for estimating the number of clusters required as a starting parameter in the K-means algorithm and certain other unsupervised machine-learning algorithms. However, due to the graphical output nature of the method, human assessment is necessary to determine the location of the elbow and, consequently, the number of data clusters. This article presents a simple method for estimating the elbow point, thus, enabling the K-means algorithm to be readily automated. First, the elbow-based graph is normalized using the graph’s minimum and maximum values along the ordinate and abscissa coordinates. Then, the distance between each point on the graph to the minimum (i.e., the origin) and maximum reference points, and the “heel” of the graph are calculated. The estimated elbow location is, thus, the point that maximizes the ratio of these distances, which corresponds to an approximate number of clusters in the dataset. We demonstrate that the strategy is effective, stable, and adaptable over different types of datasets characterized by small and large clusters, different cluster shapes, high dimensionality, and unbalanced distributions. We provide the clustering community with a description of the method and present comparative results against other well-known methods in the prior state of the art.

No abstract

The rapid development of new information and communication technologies (ICTs) and the deployment of advanced Internet of Things (IoT)-based devices has led to the study and implementation of edge computing technologies in smart grid (SG) systems. In addition, substantial work has been expended in the literature to incorporate artificial intelligence (AI) techniques into edge computing, resulting in the promising concept of edge intelligence (EI). Consequently, in this article, we provide an overview of the current state-of-the-art in terms of EI-based SG adoption from a range of angles, including architectures, computation offloading, and cybersecurity concerns. The basic objectives of this article are fourfold. To begin, we discuss EI and SGs separately. Then we highlight contemporary concepts closely related to edge computing, fundamental characteristics, and essential enabling technologies from an EI perspective. Additionally, we discuss how the use of AI has aided in optimizing the performance of edge computing. We have emphasized the important enabling technologies and applications of SGs from the perspective of EI-based SGs. Second, we explore both general edge computing and architectures based on EI from the perspective of SGs. Thirdly, two basic questions about computation offloading are discussed: what is computation offloading and why do we need it? Additionally, we divided the primary articles into two categories based on the number of users included in the model, either a single user or a multiple user instance. Finally, we review the cybersecurity threats with edge computing and the methods used to mitigate them in SGs. Therefore, this survey comes to the conclusion that most of the viable architectures for EI in smart grids often consist of three layers: device, edge, and cloud. In addition, it is crucial that computation offloading techniques must be framed as optimization problems and addressed effectively in order to increase system performance. This article typically intends to serve as a primer for emerging and interested scholars concerned with the study of EI in SGs.

Data aggregation may be considered as the technique through which streams of data gathered from Smart Meters (SMs) can be processed and transmitted to a Utility Control Center (UCC) in a reliable and cost-efficient manner without compromising the Quality of Service (QoS) requirements. In a typical Smart Grid (SG) paradigm, the UCC is usually located far away from the consumers (SMs), which has led to a degradation in network performance. Although the data aggregation technique has been recognized as a favorable solution to optimize the network performance of the SG, the underlying issue to date is to determine the optimal locations for the Data Aggregation Points (DAPs), where network coverage and full connectivity for all SMs deployed within the network are achieved. In addition, the main concern of the aggregation technique is to minimize transmission and computational costs. In this sense, the number of DAPs deployed should be as minimal as possible while satisfying the QoS requirements of the SG. This paper presents a Neighborhood Area Network (NAN) placement scheme based on the unsupervised K-means clustering algorithm with silhouette index method to determine the efficient number of DAPs required under different SM densities and find the best locations for the deployment of DAPs. Poisson Point Process (PPP) has been deployed to model the locations of the SMs. The simulation results presented in this paper indicate that the NAN placement scheme based on the ageless unsupervised K-means clustering algorithm not only improves the accuracy in determining the number of DAPs required and their locations but may also improve the network performance significantly in terms of network coverage and full connectivity.