No abstract

Summary Transmission and distribution systems for electricity have undergone a technological revolution in terms of operation and management using computer networks, automation, remote sensing, and information and communication technologies to improve the performance of digital electronic meters. This work describes the integration of a wireless sensor networks (WSNs)–based communication system with an electrical energy‐measurement structure, to verify the feasibility of large‐scale installation of intelligent electronic meters in low‐voltage consumer units. The study is based on simulations, using Castalia, considering 2 scenarios, the first in a flat network and the second in a hierarchical network of WSNs to analyze the feasibility of sending messages from intelligent electronic meters to the concessionaires through a ZigBee network.In addition, the time requirements of the IEC 61850 standard for sending and receiving manufacturing message specifications and generic object‐oriented substation event type messages are verified. This work demonstrated the technical feasibility of using WSNs for different node densities by region and evaluated the location of the sink node, and adequate infrastructures for WSNs were found. This extends time checks for both vertical (usually for supervision) and horizontal (used for protection) messages. The proposed model has great potential to use a WSN infrastructure and to evaluate if this infrastructure allows data transmission of the protocols used in smart grids, mainly verifying the requirements of transmission times required by each application.

In the past decades, fuzzy logic has played an essential role in many research areas. Alongside, graph-based pattern recognition has shown to be of great importance due to its flexibility in partitioning the feature space using the background from graph theory. Some years ago, a new framework for both supervised, semi-supervised, and unsupervised learning named Optimum-Path Forest (OPF) was proposed with competitive results in several applications, besides comprising a low computational burden. In this paper, we propose the Fuzzy Optimum-Path Forest, an improved version of the standard OPF classifier that learns the samples' membership in an unsupervised fashion, which are further incorporated during supervised training. Such information is used to identify the most relevant training samples, thus improving the classification step. Experiments conducted over twelve public datasets highlight the robustness of the proposed approach, which behaves similarly to standard OPF in worst-case scenarios.