Keywords: Artificial Intelligence, Machine Learning, Virtualization, Simulation Techniques and Augmented Reality
Abstract: Durante décadas a inclusão na educação de pessoas surdas e deficientes auditivas tem sido um problema recorrente sem solução efetiva. Partindo do crescimento na área da inteligência artificial, este artigo apresenta um estudo na comparação de dois sistemas de reconhecimento de voz para o desenvolvimento de uma aplicação de legendagem de aulas para acessibilidade de alunos surdos. Os testes realizados foram baseados nos parâmetros mínimos para a melhor utilização dessas legendas em tempo real.

Keywords: Virtualization, Simulation Techniques and Augmented Reality, Automation and Process Control, Cyber Physical Systems, Digital Twins and Knowledge Systems
Abstract: Este trabalho teve por objetivo criar um ambiente virtual sob demanda em nuvem para simulação de plantas industriais de pequeno porte utilizando Node-RED como ferramenta e verificar sua viabilidade para utilização no aprendizado acadêmico. Foi desenvolvida uma biblioteca de componentes (nodes) juntamente com supervisório Scada-LTS que possibilitou simular e monitorar uma planta de controle de nível. Simulações foram executadas e foi verificada sua usabilidade como laboratório remoto. Foi evidenciado que é perfeitamente possível criar um ambiente em nuvem sob demanda utilizando o Node-RED para simulação e utilizar como plataforma didática para aprendizado de controle de processo bem como constatar seus benefícios com flexibilidade, redução de custo, segurança, acessibilidade.

Keywords: Virtualization, Simulation Techniques and Augmented Reality, Autonomous Vehicles and Drones, Electrical Machines and Drives
Abstract: This work aims to develop and identify mathematical models capable of describing the movement conditions of a semi-autonomous wheelchair based on wheel slip. An incremental kinematic model is used to determine the spatial position of the vehicle. Euler-Lagrange equations are used for calculating the vehicle's inertia forces. The ground contact forces, which characterize wheel slip, are obtained from a proposed empirical model that relates tire slip with the normal force on the ground and the coefficient of friction of the tire-ground contact.

Keywords: Electrical Vehicle and Energy Storage, Smart Grids, Deep Learning and Machine Learning
Abstract: The rapid adoption of electric vehicles (EVs) necessitates advanced and efficient charging management systems. This paper introduces an AI-based charging system that emphasizes the importance of load management, equitable energy distribution, and the detection of charging anomalies. Designed to provide users with accurate cost forecasts and proactive anomaly detection, our proposed system aims to enhance the EV charging experience. Implemented on the Azure cloud platform, the system's core functionalities have been proven effective. Through a case study conducted in Espírito Santo, Brazil, we demonstrate the capability of our system to improve the EV charging infrastructure. While the potential for integration with smart grid data exists, the primary focus of this work is on the standalone capabilities of the charging system, laying the groundwork for future smart grid integrations.

Keywords: Electrical Machines and Drives, Electrical Vehicle and Energy Storage, Robotics and Mechatronics
Abstract: This work presents a testing framework for evaluating the performance of linear electromagnetic semi-active and active suspension systems working under standard road profiles or periodic references. The framework consists of the definition of an experimental setup, test scenarios, the validation of the dynamic response of the platform, and performance assessment based on the position and acceleration of the sprung mass. This paper provides a reliable and reproducible methodology for evaluating the performance of linear electromagnetic semi-active and active suspension systems under different operation conditions. The attained results may be used to optimize the performance of linear electromagnetic actuators applied to suspension systems and develop new control strategies.

Keywords: Automation and Process Control, Industry Applications, Diagnosis, Prognosis and System Identification
Abstract: The possibility of comparing, storing and analyzing values obtained by pressure sensors is of great interest for research related to the oil and gas industry, since the accuracy of the collected values is of fundamental importance in tasks such as process control, production, detection and correction. of leaks. This work proposes the development of a system that allows identifying measurement errors in pressure transducers through the collection of analog signals, enabling the experience through a user interface. The complete solution integrates hardware and software elements where the signals obtained by two pairs of different pressure sensors are conditioned through amplifier circuits, sent to the computer through serial communication, and presented in a graphic interface. It allows to compute the average pressure per time, to storage collected data, to analyze quality of sensors and to study the margin of error.

Keywords: Power and Eneergy Systems, Renewable Energy
Abstract: No mercado livre brasileiro, quando o agente é classificado como autoprodutor de energia, ele precisa analisar suas condições de demanda e potência instalada para verificar se, de fato, é vantajoso se cadastrar na Câmara de Comercialização de Energia Elétrica para vender seus excedentes. Neste artigo, além de apresentar uma contextualização do mercado livre e da autoprodução de energia no Brasil, foi feita uma breve análise de viabilidade de injeção de potência na rede para um autoprodutor localizado na cidade de Sentinela do Sul, Rio Grande do Sul, e por fim, foi apresentada uma metodologia para controle de geração.

Keywords: Power Electronics, Power Quality, Diagnosis, Prognosis and System Identification
Abstract: This article describes a brief application of the Conservative Powers Theory, especially the unbiased integral voltage and reactive energy, applied to estimate the equivalent impedance values in single-phase systems for fundamental frequency voltages with linear or nonlinear loads, considering low voltage distortion percentage. It contains detailed equations, methodology, and a simulation results table that validates the method's applicability.

Keywords: Automation and Process Control, Diagnosis, Prognosis and System Identification, Industry Applications
Abstract: Para obter o melhor controle possível de um sistema, é necessário ter informações precisas sobre seus estados atuais. É comum que os sensores utilizados forneçam medições ruidosas e imprecisas. Para superar esse problema, técnicas surgiram ao longo do tempo com o objetivo de melhorar a resposta do sistema e otimizar os resultados. Este trabalho apresenta a fusão do sensor de vazão com o sensor de nível já presente na planta didática do Laboratório de Controle do CEFET-MG em Araxá, e a implementação do filtro de Kalman estendido para redução de ruído na variável de processo. Isso permite que as técnicas de controle aplicadas durante as aulas sejam mais eficientes. Os resultados demonstram que após a filtragem do sinal os valores dos níveis apresentaram uma variância menor do que antes da realização deste trabalho.

Keywords: Robotics and Mechatronics
Abstract: This article provides a brief overview of robot teleoperation, addressing recent technological advances to enhance operator immersion. Furthermore, we review the most relevant and recent research on teleoperation applications from the past three years. Overall, this article is a concise resource for understanding the fundamental concepts and recent advancements in robot teleoperation.

Keywords: Personalized Products, Self Configuration & Self Diagnosis, Virtualization, Simulation Techniques and Augmented Reality
Abstract: O homem vem aprendendo com a natureza há milhares de anos, buscando entender suas estratégias para aplicar no seu cotidiano. Através do conhecimento de 3.8 bilhões de anos de inovação natural, esse conhecimento no século XX se transformou em ciência, sendo denominada biomimética. A área da biomimética busca aplicar um método natural na solução de um problema humano, mimetizando a forma e a função natural. A biomimética é uma área de pesquisa interdisciplinar que navega em diversos campos, como engenharia, arquitetura, design entre tantos outros, sendo uma poderosa ferramenta de inovação. Esta revisão da literatura teve como objetivo aprofundar o conhecimento no campo da biomimética e foi possível mapear os principais conceitos da ciência, autores, áreas de aplicação e exemplos de sucesso. Além disso, discutir sua diferença frente a outros conceitos como design bioinspirado, biomimetics, biomimicry, biônico e bioreplicação. Além disso, foi discutido também sua relação com o meio ambiente. E por fim, essa revisão bibliográfica identificou uma grande oportunidade para pesquisas futuras no entendimento dos conceitos dos termos biomimicry e biomimetics e disciplinas correlatas, assim como a inclusão da relação com o meio ambiente nas aplicações bioinspiradas.

Keywords: Industry 4.0, Internet of Things, Blockchain Technology
Abstract: Esse projeto apresenta a criação de uma arquitetura de comunicação interoperável de dados entre dispositivos IoT industriais de diferentes organizações e sistemas, utilizando o protocolo MQTT da internet das coisas, oferecendo uma integração confiável e transparente através da tecnologia blockchain. Possibilitando uma integração entre diferentes empresas e instituições acessando os dados dos dispositivos IoT de maneira hierárquica, confiável e interoperável, sem comprometer a escabilidade do sistema, visando um aumento lucrativo e tecnológico.

Keywords: Power and Eneergy Systems, Industry 4.0, Smart Grids
Abstract: The study emphasizes the importance of a solid protection system in an electrical power system to ensure reliable supply. Reclosers in the distribution network must comply with regulations and be well coordinated. The researchers used accurate data from the regulatory agency, analyzed with QGIS for topology, and OpenDSS for simulations and equipment adjustments. Consistent results were obtained for settings, where these tools have the potential for measurement, monitoring, and operation in digital substations.

Keywords: Renewable Energy, Power and Eneergy Systems, Optimization Heuristics and Methods
Abstract: In this paper, we show a simple mathematical modeling and its comparison with real measurements for the estimation of solar radiation in places without a measuring instrument. For this prediction we a full year data from three adjacent weather stations located around the point to predict solar radiation. From these real data, we propose a model to estimate radiation and present the results obtained for the model versus the control site for predictions over a full day.

The results obtained show a very good approximation to the data obtained by the proposed mathematical model and the instruments to measure radiation for the test point. The mean square error (MSE) was used as a figure of merit to measure the performance of the estimates.

Keywords: Machine Learning, Cybernetic Systems, Industry Applications
Abstract: In recent years, biometric authentication methods have gained significant traction as a highly secure means of encrypting information. Fingerprint identification has emerged as a popular choice owing to its distinctiveness and inalterability. Nevertheless, most research concentrates on the direct storage of fingerprint templates and depends on fingerprint characteristics for the purpose of identification. Extracting keys from fingerprint images and encrypting information with that key is still an open question. The present study aims to extract a distinctive key, derived from the characteristics of fingerprint patterns, from multiple fingerprint images belonging to a single user. The utilization of error correction methodologies, particularly the Reed-Solomon Code, is anticipated to yield consistent outcomes in terms of the extracted keys, regardless of the specific set of images employed. Thus, we can determine that the fingerprints are from the same individual. The accuracy of a method is evaluated by computing the False Negative Rate and False Positive Rate. The proposed scheme exhibits enhanced efficacy in safeguarding users' fingerprint information through the implementation of measures that effectively prevent the inadvertent or intentional disclosure of such sensitive data.

Keywords: Life Support Systems & Techniques
Abstract: Este artigo discute diferentes tipos de sensores, como sistemas de plataforma e sistemas in-shoe, e descreve os sensores capacitivos, resistivos, piezoelétricos e piezoresistivos. A aplicação desses sensores pode ajudar no monitoramento da pressão plantar e da tensão de cisalhamento, proporcionando um gerenciamento mais eficaz das úlceras no pé em pacientes diabéticos.

Keywords: Internet of Things, Smart Grids
Abstract: A Internet das Coisas dá vida a um novo conjunto de serviços que podem ser implantados na operadora de rede móvel e oferecidos ao usuário final, tais como touch internet, realidade aumentada, realidade virtual, redes médicas, previsão do tempo, casas e cidades inteligentes (smart homes e smart cities), entre outros. Este artigo trata da possibilidade de implantação dos serviços do conceito “Smart City”. Com base na experiência de outras cidades, verifica-se a relevância da implementação da iluminação pública inteligente (IPi), que permite ao município economizar recursos orçamentários, gastos com iluminação pública na cidade e otimizar os processos de manutenção. A implantação deste conceito por meio de tecnologia sem fio propiciará um incremento no número de serviços disponíveis à comunidade, sendo possível agregar funcionalidades no futuro com um investimento mínimo. Este artigo apresenta as tecnologias sem fio de longo alcance e baixo consumo de energia mais promissoras avaliando-as sob a ótica de determinados parâmetros, como complexidade e custo. O trabalho também apresenta a proposta de um sistema de IPi e resume algumas experiências de laboratório envolvendo a amostragem de dados de tensão e corrente por meio de circuitos integrados dedicados, tendo em vista sua pertinência para o dispositivo ora em desenvolvimento.

Keywords: Diagnosis, Prognosis and System Identification, Automation and Process Control, Industry Applications
Abstract: This work presents a proposal for implementing fault-tolerant distributed supervisory control for Discrete Event Systems. The supervisor localization approach is adopted to obtain the control structure and the obtained supervisors are implemented in different Function Blocks, executed in a Programmable Logic Controller. The fault tolerance was obtained through the use of reconfiguration block between the controller and the plant. The proposal is applied in a case study developed on a virtual plant of the Factory IO software.

Keywords: Deep Learning and Machine Learning, Artificial Intelligence
Abstract: Um dos principais segmentos para a criação de Cidades Inteligentes é o suporte a Sistemas de Transporte Inteligentes (do inglês ITS - Intelligent Transportation System). Para isso, a análise do fluxo de tráfego é fundamental quando o objetivo é aumentar a infraestrutura, reduzir acidentes, custos e emissões de poluentes, focando na saúde e segurança populacional. Devido à ampla disponibilidade de câmeras de monitoramento na maioria das cidades, sistemas baseados em visão computacional têm se tornado tendência, com grande potencial para fornecer informações através do processamento da imagem. Neste trabalho foi proposta uma abordagem de extração de dados através da detecção, rastreamento de objetos e estimativa de velocidade por matriz de projeção de uma seção na via. Isso permite criar os diagramas essenciais de fluxo de tráfego, constituídos pela velocidade, fluxo e densidade dos veículos. O funcionamento do sistema foi comprovado por meio da validação dos dados extraídos conforme os modelos de fluxo de tráfego mais conhecidos da área, mostrando que o framework proposto funciona satisfatoriamente.

Keywords: Artificial Intelligence, Deep Learning and Machine Learning
Abstract: Colorectal cancer (CRC) can be detected analysing the presence of polyps, which are abnormal tissue growth with several shapes and sizes presented in colonoscopy images. Detectig polyps can help an early diagnostic for CRC, and contribute with the patient treatment, since all CRC cases originate from polyps. In this paper we propose a Transformerbased classifier to detect polyps in colonoscopy images. The database addressed here is the CP-CHILD. Our approach presented promising results with an accuracy of 97.86% and with average precision values of 97.51% on the test set, as well as a recall of 91.87% and a F1 score of 94.56%.

Keywords: Deep Learning and Machine Learning, Micro & Nano Manufacturing, Industry 4.0
Abstract: Real-time control of SACE gas film stability is crucial, as it significantly impacts micromachining repeatability and quality in this technology. Gas film stability and discharging activity are interconnected, and monitoring real-time parameters like mean discharge current and energy, which serve as indicators of gas film stability, is the first step in this effort. An intelligent algorithm deployed on a dSPACE platform uses LSTM for online discharge activity monitoring, identifying discharges and calculating indicators. Maintaining a short enough sampling time for prompt discharge detection presents overrun errors. Therefore, a real-time multitasking environment with a 1.6e-5 seconds sample time is executed. A more complex LSTM enhances detection accuracy but extends execution time, potentially resulting in more unprocessed data loss. The research examines the real-time model with various algorithm feed batch sizes and LSTM complexities, particularly the number of hidden units. An example of a 2-hidden-unit LSTM demonstrates promising 90.45% accuracy, processing data every 264 milliseconds with a 131-millisecond batch (approximately 0.5 processing ratio), indicating superior performance. In the future, exploring LSTM hyperparameter optimization and real-time model parameter tuning is recommended to enhance accuracy and processing ratio.

Keywords: Machine Learning, Deep Learning and Machine Learning, Artificial Intelligence
Abstract: Graph Neural Networks have played a key role in several machine learning applications in recent years. They have proven effective in dealing with structured data, capturing complex information, and allowing more sophisticated modeling in various domains. As such, these networks have contributed significantly to advancing and improving several areas in the machine learning field. This paper proposes a new approach for gesture classification using graphs, consisting of three main blocks: a Graph Constructor, a Feature Extractor, and a Gesture Classifier. A skeleton graph is constructed for each video frame in the first block. Feature extraction is performed using a Graph Neural Network, where the feature vectors of the nodes are modified based on information from the joints themselves and their neighboring joints. These vectors are then concatenated to form an embedding that represents the frame. The final representation of the video is obtained by concatenating these embeddings to form a pseudoimage. This pseudoimage is used as input to a classifier based on convolutional and dense layers. Also, we propose a new approach for the GraphConv layer, which attributes different weights for the neighboring node based on the distance from a reference node in the graph. Through experiments performed with the Chalearn data set, we achieved an accuracy of 91.84%. Even using a Graph Neural Network with only one layer and a reduced number of parameters to be trained, the results proved to be competitive compared to works using more complex architectures. This indicates that the approach proposed in this work has a promising and efficient performance in gesture classification.

Keywords: Industry Data Science Applications, Automation and Process Control, Machine Learning
Abstract: In this work, the problem of monitoring the performance of control loops and retuning is addressed. Poor performance of control loops reduces product quality, generates losses and can affect safety in industrial processes. There is a great interest in the industry in the search for methods that are simple to use and interpret. Cascaded control loops are widely used because they mitigate the effect of disturbances and process variations on the controlled variables. The fast dynamics of the inner loop, whose reference is generated by the controller of the outer loop, represents a challenge when one wants to investigate its performance during operation. The methods proposed here are completely data-based, requiring the user only to provide operating data and to approve the performance of the control loops to be monitored, which must always be done in some way. Both monitoring and retuning are validated by statistical tests, reducing the effect of uncertainties introduced by noise that is always present in industrial processes. The data used in all steps are generated by applying small disturbances added to the external loop control signal, reducing the deviations of the controlled variables from their references. The methodology is applied to a pilot plant with industrial equipment, and controlled by a distributed control system. A flow loop acts as an inner loop for a level control loop. The obtained results show that the methodology allows successfully attacking the proposed problem, being additionally replicable in industrial environments in general.

Keywords: Deep Learning and Machine Learning, Industry Applications
Abstract: Due to the complexity and high financial costs involved in production processes, the steel industry can benefit from applications of intelligent systems, capable of performing automated activities. This research paper addresses a description of the process of creating a data-driven computational system for the purpose of developing a computational thermal model of a real steel plate reheating furnace. Sufficiently accurate computational models can be used in conjunction with combustion control optimization techniques, such as model-based predictive control (MPC), or even a Digital Twin of the combustion system of a plate reheating furnace. The tool can be used in predictive failure diagnosis, fundamental for the maintenance and operation teams responsible for asset management. For this development, Recurrent Artificial Neural Networks have been widely applied, validating the existence of series that have temporal links between their samples, a typical case of monitoring industrial process variables. To meet the proposed objective, the performance of models based on recurrent neural networks of the Long Short Term-Memory (LSTM), Gated Recurrent Unit (GRU) and Temporal Convolutional Network (TCN) type was analyzed. The results were evaluated under different prediction horizons, since such techniques demand models capable of accurate predictions that are several steps ahead, premised on prediction capability.

Keywords: Artificial Intelligence, Deep Learning and Machine Learning, Power and Eneergy Systems
Abstract: Neste trabalho é apresentado um estudo preliminar sobre a utilização da Rede YOLOv5 para inspeção de isoladores utilizados em torres de transmissão de energia elétrica de 345 kV. O objetivo geral do trabalho é a classificação dos isoladores com base na detecção do nível de corrosão nos pinos dos isoladores de vidro. O algoritmo desenvolvido procura complementar trabalhos do estado da arte que utilizam visão computacional e drones para a detecção de isoladores quebrados ou faltantes. A detecção do estado dos pinos isoladores pela Rede YOLOv5 mostrou-se promissora, apresentando uma acurácia superior a 75% em testes iniciais com uma base de dados de apenas 784 imagens. Esse resultado supera a classificacão humana devido às dificuldades encontradas durante a inspeção manual. A classificação dos isoladores neste trabalho foi dividida em três categorias: os que estão em bons estados e nãoo precisam de monitoramento anual, os que necessitam de monitoramento anual e os isoladores que precisam ser substituídos o quanto antes.

Keywords: Power and Eneergy Systems, Machine Learning
Abstract: Este artigo apresenta uma extensa análise comparativa entre dois diferentes métodos de regressão baseados em árvore para o problema da estimação da margem de carga de potência ativa. As investigações realizadas neste trabalho permitiram concluir que o algoritmo extremely randomized trees é superior ao random forest, este último largamente utilizado na literatura técnica. Além disso, os estudos apresentados comprovaram a eficácia dos métodos para a estimação em tempo-real da margem de carga em situações de operação normal do sistema, contingências do tipo N-1 e N-2 e considerando ruído nas medidas provenientes de unidades de medição fasorial. Ademais, é apresentada uma validação estatística para os modelos a fim de comprovar a robustez e estabilidade na previsão da margem de carga. As simulações numéricas das técnicas estudadas foram desenvolvidas considerando o sistema de teste IEEE 14 barras.

Keywords: Power and Eneergy Systems
Abstract: Este trabalho apresenta uma verificação do impacto da modelagem do acoplamento mútuo de sequência zero, adotada para o sistema teste, no desempenho de um algoritmo de proteção de distância adequado para linhas de transmissão longas paralelas equilibradas conectadas a barramentos independentes em ambos os terminais. Considerando um modelo de linha de transmissão hexafásica, curtos-circuitos monofásicos-terra com valores de resistência de falta desprezíveis foram simulados no domínio das fases por meio do software MATLAB. Mostra-se que um erro relativo percentual médio de 0,37% — com máximo de 0,76% — foi encontrado para os módulos das impedâncias aparentes calculadas utilizando o algoritmo de proteção avaliado.

Keywords: Power and Eneergy Systems, Optimization Heuristics and Methods
Abstract: Este artigo propõe a definição dos ajustes de dispositivos de controle volt-var para operação diária otimizada nos sistemas de distribuição de energia elétrica. Sob a ótica da distribuidora, propõe-se um modelo de otimização mono-objetivo: (i) minimização dos custos, ou (ii) maximização da receita líquida. Ao considerar a presença de geração distribuída (GD), bancos de capacitores (BCs) e regulador de tensão (RT) na rede, foi proposto ajustar o controle volt-var, resolvendo o problema por meio de dois métodos de otimização heurística: Hill-Climbing e Simulated Annealing. Os resultados obtidos demonstraram a eficácia de um controle volt-var otimizado no controle dos níveis de tensão e, consequentemente, na melhoria da operação de uma rede de distribuição, apresentando até 16% de diminuição nos custos diários relacionados às perdas de energia na distribuição e a eliminação do custo associado ao pagamento de compensações. A partir da modelagem cronológica da carga e GD, foi possível avaliar a influência da sazonalidade nos controles ótimos.

Keywords: Micro & Nano Manufacturing, Renewable Energy, Industry Applications
Abstract: This work consists of presenting a brief review of the evolution of metallic connection technologies in busbars applied to the front and rear of solar cells, thus demonstrating their fundamentals, materials and essential manufacturing processes that impact on improving the energy efficiency of solar modules.

In this brief review, one can follow the technical limitations that led to the evolution of metallic connection technologies in busbars over time and their impacts on the photovoltaic market. Contributing to a broad understanding of the resistive components of solar cells that consists of the series resistance parameter.

Keywords: Power and Eneergy Systems
Abstract: Este trabalho apresenta uma metodologia de gestão contínua de contratos de energia desenvolvida e implementada em uma empresa de saneamento brasileira. Detalham-se os indicadores adotados e suas metodologias de cálculo, abrangendo ultrapassagem de demanda, excesso de demanda contratada, adequação de modalidade tarifária, baixo fator de potência, operação em horário de ponta, fator de carga e tarifa média. Expõem-se estudos de caso de otimização de demanda contratada e de seleção de modalidade tarifária mais econômica. Os resultados alcançados evidenciam a assertividade e a viabilidade do projeto executado e das rotinas implementadas, as quais permitem a maximização da economia à medida que as rotinas operacionais das instalações forem otimizadas.

Keywords: Power Quality, Power and Eneergy Systems, Smart Grids
Abstract: Este artigo demonstra que a integração da Teoria de Potência Conservativa com a estratégia Generalized Current-Based Control, a qual é adotada na coordenação de conversores de potência em microrredes despacháveis, provê a compensação distribuída de correntes indesejadas, mesmo perante tensões senoidais desequilibradas. Assim, um balanceamento de condutâncias pode ser obtido quando correntes reativas, de desbalanço e harmônicas são concomitantemente mitigadas. Portanto, possibilita-se que a microrrede seja interpretada pela rede principal como uma carga resistiva balanceada.

Keywords: Power Quality, Power and Eneergy Systems, Smart Grids
Abstract: A utilização de banco de capacitores em sistemas elétricos é amplamente difundida e consolidada. Estes são equipamentos que podem prover diversos benefícios a rede elétrica, atuando na regulação de tensão, compensação de reativos e fator de potência, e, consequentemente, contribuindo para o aumento da eficiência do sistema como um todo. Em contrapartida, as manobras de chaveamento de banco de capacitores causam transitórios eletromagnéticos de alta frequência que podem impactar e causar distúrbios e eventuais danos a equipamentos. Neste contexto, o estudo de coordenação de isolamento em subestações se faz necessário para determinar e analisar os efeitos das sobretensões, sobrecorrentes, além das frequências resultantes dos chaveamentos dos bancos de capacitores. De forma a promover esta avaliação, este artigo utiliza como mote principal, um circuito elétrico de banco de capacitores em paralelo (back-to-back) em uma subestação com uma possível futura ampliação, simulando as resultantes desta solução, através da utilização do software ATPDraw 7.2.

Keywords: Automation and Process Control
Abstract: This paper proposes the analysis of robustness and performance of Generalized Predictive Controller (GPC) applied in systems widely used in the industry, the so-called Benchmarks plants. Nine plants with very popular characteristics in industrial processes are presented, multiple equal poles, of fourth order, with zero in the right half-plane, transport delay and lag, transport delay and double lag, with fast and slow poles, of conditional stability, with oscillatory dynamics and with unstable poles. The objective is to analyze the efficiency of the GPC in ensuring robust performance, stability, reference tracking and disturbance rejection in such plants, whether with simple or complex dynamics, in addition to presenting the tuning parameters used for this purpose.

Keywords: Automation and Process Control, Robotics and Mechatronics, Cyber Physical Systems, Digital Twins and Knowledge Systems
Abstract: In the present paper the mathematical model of a wafer manufacturing process is presented using Discrete Event Systems. The novelty of the proposed mathematical description is the parametric number of vacuum chambers and robotic manipulators, as well as that three vacuum chambers and the respective buffers are served by each robotic manipulator. The last robotic manipulator serves a parametric number of champers depending on the total number of vacuum chambers modulo 3. The desired behavior of the process is expressed in the form of a set of regular languages. The realization of each regular language in the form of a supervisor automaton is developed. The satisfactory performance of the controlled automaton is derived. The present result can be considered as the first attempt towards the study of wafer manufacturing processes, where each robotic manipulator serves a parametric number of vacuum chambers and buffers.

Keywords: Automation and Process Control, Industry Applications, Machine Learning
Abstract: In this work, a comparison between supervised learning techniques is provided for predictive maintenance in twin screw air compressors. Significant data are selected, acquired and stored in an Industry 4.0 context. Different operating conditions of the process are considered. Subsequently to data collection, data analysis and preprocessing phases are performed in order to prepare tailored datasets to be entered into supervised learning classifiers for predictive maintenance. Four classes associated to the required time priority for maintenance are defined, concerning the state of degradation of the oil used by the compressor. In order to compare different supervised learning techniques, also correlation matrix and Principal Component Analysis are exploited. Accuracy, specificity and sensitivity are evaluated together with the confusion matrix. Significant results are obtained which prove that predictive maintenance policies can be applied to twin screw compressors instead of the widely adopted periodic and corrective maintenance policies.

Keywords: Automation and Process Control, Industry Applications, Robotics and Mechatronics
Abstract: One of the most important topics in the automotive industry is the design of a suspension system that ensures better ride comfort and handling control. This paper presents design conditions for a H∞ controller that mitigates perturbations in dynamic response, and considering a time delay in the actuator. The nonlinear suspension system will be modeled as a Takagi-Sugeno (T-S) fuzzy system, and a Lyapunov-Krasovskii functional that depends on the membership functions will be considered, in order to obtain less conservatism in the design conditions. The initial results present the design conditions of a parallel distributed compensation (PDC) controller. Therefore, the sprung and unsprung masses will be considered as uncertain nonlinear premise variables, which are not available for feedback, and a switched controller is proposed to deal with this problem. Simulation results are presented, demonstrating the effectiveness and improvement of the proposed strategy.

Keywords: Automation and Process Control
Abstract: This study presents an approach for secure communication based on the underactuated synchronization of a 4D-hyperchaotic system recently introduced in the literature. The presence of disturbances is considered in all states of the master and slave systems for a more robust synchronizer. The Lyapunov theory is used to demonstrate the effectiveness of the proposed system, ensuring that the synchronization error is bounded and achieves convergence in a finite time. Also, this work includes computer simulations based on Matlab/Simulink to validate the proposed system, including the master-slave synchronization and the implementation of a secure communication scheme.

Keywords: Power and Eneergy Systems, Smart Grids
Abstract: Inter-area oscillation modes resulting from the interconnection of large electrical systems cause operation problems and even blackouts if minimum damping rate requirements are not met. Recent and promising research shows that Wide-Area Damping Controllers (WADC) are capable of improving the damping rates of inter-area modes because they use data from remote locations of the system. However, the operation of the WADC can be compromised if multiple communication failures occur on the WADC communication channels because this contingency causes changes in control actions. This work proposes a procedure for the desired design of a WADC that meets the robustness requirements to multiple communication failures that may occur. Some bio-inspired algorithms were used, evaluated and discussed in the proposed procedure. The effectiveness of the method was evaluated and discussed on the IEEE 68-bus test system. The results show that it is possible to design a WADC up to a limit of possible permanent communication losses. Furthermore, different bio-inspired algorithms present different performances in solving the optimization problem.

Keywords: Virtualization, Simulation Techniques and Augmented Reality, Automation and Process Control, Optimization Heuristics and Methods
Abstract: Ferramentas de simulação tem sido utilizadas na engenharia há anos, sendo cada vez mais frequente seu uso no desenvolvimento de projetos, planejamento de operações, verificação de modelos, entre outros. Naturalmente, a simulação de modelos cada vez mais complexos acarreta um elevado tempo computacional, que muitas vezes inviabiliza a simulação desses sistemas. Nesse contexto, a utilização do processamento paralelo configura uma alternativa que reduz o tempo computacional gasto. Assim, neste trabalho foi avaliado o processamento paralelo em uma simulação que utiliza o modelo de um motor de indução em conjunto com um Filtro de Kalman, um controlador PID e um controlador preditivo. Esse modelo foi construído sobre a plataforma Matlab/Simulink e foi utilizada a função textit{parsim} para implementação do processamento paralelo, executada em uma máquina com processador de 20 núcleos e 64GB de memória RAM. Os resultados mostram que o aumento do número de núcleos computacionais melhora o tempo de processamento até o limite que a memória RAM consegue armazenar as informações. Utilizando memória virtual em um SSD de 256GB dedicado, o tempo de simulação atingiu um valor mínimo de 96 minutos utilizando 8 núcleos computacionais, vindo a aumentar posteriormente com a utilização de um número maior de núcleos.

Keywords: Cyber Physical Systems, Digital Twins and Knowledge Systems, Industry 4.0, Industry Applications
Abstract: As technology in industry advances and competition for maintenance efficiency intensifies, the industry is looking for more resources to ensure the reliability of its processes.Thus, this paper proposes a tool to optimize the Monitoring and Analysis System (from Portuguese, Sistema de Monitorização e Análise de Motores - SMAM), a tool that helps predictive maintenance of three-phase electric motors through the use of Digital Twin (DT). Therefore, the current of the three motor phases are monitored, as well as temperature and voltage. These measurements are collected by the ESP32 microcontroller and sent to the Firebase database. This current data is fed into the motor simulation performed in Finite Element Method Magnetics (FEMM) software. This simulation run on a virtual machine and the simulation results are stored in the database and can be viewed on a dashboard that shows the progress of each simulation. In this way, the user can check the magnetic flux and the internal temperature of the motor in real time for more precise monitoring of its operation.

Keywords: Internet of Things, Industry 4.0, Smart Grids
Abstract: A solution for the integration of Smart Grid communication protocols is proposed through an IEC 61850 - IIoT (Industrial Internet of Things) test bench. The IIoT is one of the main implementations for the fourth industrial revolution, known as Industry 4.0. However, it is known that the incorporation of new technologies in the electricity sector, for monitoring systems and actuation by digital commands, are slower due to the requirements of efficiency and security in data transmission. Some of the most relevant IoT standards are: Constrained Application Protocol (CoAP) and Message Queue Telemetry Transport (MQTT). The implemented test bench allows the management of information through IEC 61850 services, with MQTT/COAP – IIoT messages, in an SF6 gas monitoring application. Likewise, it provides interoperability evaluations of isolated systems that use non-interoperable communication protocols in the substation and, therefore, are not capable of integrating monitoring signals and data into the SCADA system. Communication performance test results for both protocols, MQTT y CoAP, in the implemented IIoT platform and integration performance comparisons for digital substations are presented, showing the efficiency of the proposal.

Keywords: Virtualization, Simulation Techniques and Augmented Reality, Industry Applications, Diagnosis, Prognosis and System Identification
Abstract: An adequate plant layout expedites the production process, optimizing time and resources to improve the industrial system's performance. This work presents the plant layout of a blackberry pulp production process using the systematic layout planning (SLP) methodology, which allows for analysis and evaluation and renders all the elements involved in the productive process. This qualitative method allows for locating the internal components of the plant-based on different factors. Tools used include material flow analysis through the from/to chart; the Guerchet methodology makes it possible to determine the total area needed; the space relationship diagram allows fixing the proximity between the elements; finally, three alternatives are established and evaluated for the distribution. According to the scores obtained, the alternative improvement is selected, and the plant layout is virtualized using three different 3D design software that allows rendering of the plant layout.

Keywords: Virtualization, Simulation Techniques and Augmented Reality, Diagnosis, Prognosis and System Identification
Abstract: Estima-se que a demência afete gravemente milhões de pessoas ao nível mundial, sendo emergente a investigação na prevenção, tratamento e promoção da qualidade de vida. A terapêutica não farmacológica, nomeadamente através da utilização de ambientes sensoriais suportados por tecnologia, são potencialmente relevantes na abordagem à pessoa com demência. Estas novas abordagens devem estimular a personalização, com base na interpretação dos estados neuroemocionais, através da eletroencefalografia. O objetivo do presente estudo consistiu na caracterização dos estados neuroemocionais durante a projeção de cenários personalizados para guiar o exercício psicomotor em pessoas com demência. Foram recrutadas 7 pessoas com quadro demencial sem quadro de fadiga severa ou condição respiratória e cardíaca diagnosticadas; sem problemas severos de visão; sem dificuldades comportamentais na aceitação do equipamento eletroencefalográfico. Os cenários testados foram desenvolvidos e personalizados de acordo com um portefólio de reminiscência previamente recolhido pela equipa. Os participantes foram avaliados 2 minutos antes e logo após a sessão de teste através do equipamento Emotiv EPOC+ para recolha do eletroencefalograma. Não ocorreram mudanças significativas nos valores recolhidos. A correlação entre magnitude das diferenças na resposta de stress e do foco foi significativa e positiva (r= 0.89, p= 0.01). O envolvimento, o foco e o stress aumentaram em 5 dos 7 participantes. Concluindo, a eletroencefalografia permitiu aferir as diferenças nas respostas neuroemocionais na personalização da terapia em pessoas com demência.

Keywords: Virtualization, Simulation Techniques and Augmented Reality, Industry 4.0, Internet of Things
Abstract: Sensor data is essential for process production, but direct interaction with equipment can lead to health risks, loss of time, and disruption. This article describes the creation of a framework based on Augmented Reality technology that allows users to view the real world through the camera while also seeing digital visual elements superimposed on the scene. An application using coupled tanks was developed, with visual resources that include the real-time display of the values corresponding to the volume of each tank. There is also a dashboard that presents information relevant to the plant control process, such as sensor and actuator outputs. The application efficiently creates a communication bridge between mobile and industrial plants, allowing operators to access equipment's relevant information quickly and safely without interrupting production.

Keywords: Additive Manufaturing, Industry Applications, Industry 4.0
Abstract: The topology optimisation method has gained significant attention in recent decades due to the extensive development and implementation of additive manufacturing, an advanced technology applied to fabricate complex geometries and structures. By following the topology optimisation methodology, the existing geometry can be effectively optimised by minimising or maximising objective functions, such as stiffness, volume, or weight reduction. This paper provides an overview of the topology optimisation algorithm and compares the capabilities of computer-aided software designed to conduct topology optimisation procedures. Four different software are analysed using case studies from various industries. The case study models are categorised based on important parameters for the topology optimisation and evaluated in terms of availability, optimisation method, objective function, and other factors.

Keywords: Additive Manufaturing, Diagnosis, Prognosis and System Identification, Industry Applications
Abstract: The additive manufacturing process commonly referred to as the "3D printing process", requires effective monitoring of the initial layer printing to ensure successful part fabrication. It is possible to achieve this by using acoustic sensors such as the electret microphone, which are widely used and cost-effective measuring devices renowned for their precision. Mounting the sensor onto the extruder of the printer has been demonstrated to be effective in capturing signals that accurately represent the process. Nevertheless, further research is necessary to explore new methods of processing and analyzing these signals to fully realize the microphone’s potential. The Counts method is a trustworthy technique for measuring the activity of a signal above a predetermined threshold. The present study aims to examine the electret microphone's response to the manufacturing of a monolayer square part while evaluating the efficacy of the Counts method in capturing the key features of the sensor's signals, thereby enabling the monitoring of the 3D printing process. Two specific defects, filament running out during printing and extruder nozzle clogging during printing, were generated from the normal pattern. Results confirm that mounting the electret microphone to the printer's extruder generates less noisy signals than alternative placements, consistent with previous research. Moreover, using the Counts method and selecting an appropriate frequency band allows the sensor to differentiate between conditions; clogging reduces the signals' amplitude, while filament lack increases it. In conclusion, the electret microphone is an efficient tool for monitoring 3D printing when mounted to the printer’s extruder, and the Counts method effectively captures the signal’s key features.

Keywords: Diagnosis, Prognosis and System Identification, Big Data in Industry Applications, Machine Learning
Abstract: Predictive maintenance plays an important role in ensuring the profitability of companies that rely on commercial vehicles as a core durable asset in the transportation industry. However, developing accurate and reliable predictive models for such equipment requires specific data analysis methodologies. This work proposes a data-driven methodology for predictive maintenance of commercial vehicles, comprising three main steps adapted from generic data mining methods: Problem Understanding, Data Understanding, and Data Preparation. We used a case study that uses operational data and maintenance history of turbochargers to demonstrate the feasibility of the proposed methodology that is based on the collection and analysis of real industry data for detailed analysis of relevant variables and prepare the dataset for construction of a predictive model. In this way, it is possible to identify trends and patterns that allow for the prediction of possible failures in vehicles before they occur, increasing their Uptime reducing maintenance costs.

Keywords: Artificial Intelligence, Deep Learning and Machine Learning
Abstract: The purpose of this work is to apply some natural language processing (NLP) techniques over a set of statements typical of programming contests, so as to provide a concrete basis to the future development of a recommendation system. In addition to the possibility of generating co-occurrence networks, which presented an accurate representation of the semantic similarity between the statements, it was also possible to perform an analysis of the spatial distribution of the challenges in a bi-dimensional space, which made it possible to indentify clusters based on additionally available attributes such as category, difficulty level and correct answers ratio. Ultimately, it was observed that the term-document vectorization method presented better results when generating co-occurrence networks, while the TF-IDF performed better when searching for cluster formations.

Keywords: Industry Data Science Applications, Machine Learning, Smart Grids
Abstract: In the era of big data analytics, customer segmentation requires new approaches to analyze consumption behavior and develop targeted services. In this study, we explored the K-medoids clusterization method with various similarity metrics and clustering scores, combined with PCA, t-SNE, and UMAP dimensionality reduction techniques to investigate typical and peculiar patterns in electricity consumption from a smart grid system in Brazil. By utilizing traditional and novel features and conducting features interpretation and pattern shift analysis, we demonstrate the effectiveness of this framework in supporting decision-making processes for electric power companies and promoting sustainable energy management through customer segmentation.

Keywords: Machine Learning, Power and Eneergy Systems, Artificial Intelligence
Abstract: Modern restructured power systems are identified by the establishment of competitive markets. After the 90s, many countries have been adopted offer-based short-term (day-ahead and intraday) markets to enforce competitiveness among agents. In these markets, generators can improve their profits by strategic bidding during energy auctions. This paper aims to study generators' bidding strategies using reinforcement learning techniques (Q-learning) to evaluate factors that impact operational profit. As the main result, we discuss how the generator's physical location can be a determinant of its operational profit and, consequently, investment decisions in power systems.

Keywords: Artificial Intelligence, Automation and Process Control, Industry Applications
Abstract: Mining companies are concerned about air pollution caused by dust emissions, as suspended dust particles can be a nuisance for workers and nearby populations. Fog dust suppression devices, also known as fog cannons, are one of the strategies in the modern mining industry to control dust emissions and reduce particulate emissions and dispersion. However, an internal analysis revealed that the cannon's default settings or manual operation have induced resource wastefulness and inefficiency. Based on artificial intelligence and real-time environmental sensor data, a Catboost Decision Tree Regressor prediction model was utilized to estimate the activation time and duration of the cannon. The results demonstrated that fog cannon performance rose by 80% in terms of true diffuse emission, resulting in a 36% reduction in dust emission and a 20% decrease in water use.

Keywords: Machine Learning, Industry Data Science Applications, Deep Learning and Machine Learning
Abstract: This work contributes with a new approach for tuning hyperparameters of machine learning models, based on sequences of optimization studies based on an initial range of hyperparameters. Through the proposed methodology, each sequence of studies allows the delimitation of an optimal range of hyperparameters to be inserted and evaluated by a Bayesian optimization framework, Optuna, in search of better performance metrics for the model used. The technique developed in this work was applied for short-term electrical energy prediction, with 15-minute and 1-hour data, using energy consumption data from a steel industry. We used ensemble and decision tree learning models as predictors, including Random Forest Regressor, Support Vector Regressor and Cubist Regressor, which have already been used in the literature to predict energy consumption using the same database. In an unprecedented way, we used the XGBoost model as a predictor of energy consumption in the proposed context. The results obtained from each model surpassed the performance metrics previously obtained in the literature for the same prediction scenarios, even without the use of specific feature selection techniques or pre-processing. To predict the 15-minute and 1-hour energy consumption, we obtained a Root Mean Square Error of 0.175 kWh and 1.341 kWh for the test set, respectively, using the Cubist Regressor model.

Keywords: Artificial Intelligence, Deep Learning and Machine Learning, Industry 4.0
Abstract: While artificial intelligence adoption increases in several industries and sectors, Occupational Health and Safety (OHS) still, resist adopting disruptive AI innovations in Brazil. This paper aims to present a comparative investigation between the willingness to adopt cognitive systems (CS) and low expert systems (ES) for OHS professionals in the Brazilian industry for risk assessment management to understand potential adoption behavior. Design/Methodology/Approach - The study adopts a survey based on the UTAUT (Unified Theory of Acceptance and Use of Technology) framework, with 88 OHS professionals from several states of Brazil, users of a platform for OHS risk assessment. Users were presented with two vignettes, one testing willingness to adopt a Low Expert System for OHS and the other testing willingness to adopt a Cognitive System for OHS. Results were analyzed through Structural Equation Modeling in SmartPLS software and statistical test comparison of the two scenarios. Findings and Implications - Evidence that the UTAUT score for ES is statistically significantly higher than CS, In this context, this sample of Industry Professionals has less intention to adopt AI Technologies for OHS despite discussing the theme. Several factors explain this behavior, from the perception of a lack of control, legislation, and other issues that constitute barriers to AI adoption in OHS. Results support the need to advance into Human-Centered AI approaches to design AI-based solutions for the OHS Industry and reinforce UTAUT as an essential theoretical framework for supporting AI studies. Limitations - The study is limited to companies that use the Gáutica SaaS OHS platform for risk assessment management in machinery operations and pressure vessel inspection

Keywords: Renewable Energy, Power and Eneergy Systems
Abstract: Photovoltaic solar energy has expanded a lot in recent years, especially distributed generation. The Instituto Federal do Norte de Minas Gerais - Campus Montes Claros is located in a region suitable for the installation of photovoltaic plants, with a 15 kWp plant and another 45 kWp plant. Installed in 2018 and 2019, respectively, these plants did not undergo maintenance or cleaning of their panels. This study aims to determine the impact of soiling in photovoltaic panels on the performance of the studied systems, based on the comparison of power plant generation data and performance indicators. A simulation of the systems is also carried out in order to verify the expected generated energy. For simulation and performance calculation, the components of global radiation on the inclined plane are calculated from global radiation data on the horizon. It is verified that after cleaning the photovoltaic panels there is an increase in the generation of photovoltaic plants.

Keywords: Renewable Energy, Artificial Intelligence, Machine Learning
Abstract: Considerando o caráter intermitente e a crescente inserção da geração solar fotovoltaica (FV) na matriz elétrica mundial nos últimos anos, é imperioso o desenvolvimento de modelos de previsão de geração cada vez mais precisos, de modo a permitir um melhor planejamento da operação da planta FV e do sistema elétrico como um todo. As redes neurais artificiais têm se tornado muito populares por apresentarem resultados promissores, devido a assertividade na previsão da geração FV e desempenho robusto do modelo. A principal contribuição deste trabalho está na implementação e comparação de modelos de previsão de geração FV de uma usina de 164 MWp, utilizando redes do tipo Focused Time Delay Neural Network (FTDNN). Backpropagation, Adam, otimização por enxame de partículas (do inglês, PSO), PSO Caótico (do inglês, CPSO) e PSO com fator de envelhecimento e enfraquecimento (do inglês, PSO-AWF) foram testados no treinamento da rede, enquanto o último algoritmo também foi usado na otimização dos parâmetros da arquitetura da rede FTDNN. Para fins de comparação de desempenho, foram usados os modelos de referência regressão por perceptron multicamadas, regressão linear, regressão por árvore de decisão e persistência. Com base em diferentes métricas estatísticas de desempenho, o modelo FTDNN com ajuste manual de parâmetros e técnica de treinamento PSO-AWF obteve o melhor resultado, com Raiz do Erro Médio Quadrático 18,354 MW, Erro Absoluto Médio 13,784 MW, Coeficiente de Correlação de Pearson 80,042%, Raiz do Erro Médio Quadrático Normalizada 14,155% e Erro Absoluto Médio Normalizado 10,631%.

Keywords: Power Quality, Power and Eneergy Systems, Renewable Energy
Abstract: Solar photovoltaic generation has been considerably increasing in distribution networks. However, the inherent variability of their output power poses challenges to operators to preserve the power quality. Smart inverters feature such as Volt-Var control can regulate voltage level by absorbing and/or injecting reactive power and has been integrated into grid codes. As it is expected to be a part of networks in the future, studies of its technical consequences have been demanded. This paper evaluates the effectiveness of Volt-Var control on voltage regulation in a residential feeder with high photovoltaic penetration. A quasi-static probabilistic power flow is performed using Monte Carlo Simulation to consider irradiance and load demand uncertainties, which are probabilistically modeled based on historical data. The results show that Volt-Var control can effectively mitigate overvoltage problems caused by reverse power flow, but at the cost of increasing technical losses.

Keywords: Power Quality, Machine Learning, Diagnosis, Prognosis and System Identification
Abstract: Harmonic analysis using the Discrete Fourier Transform is a relevant topic in electric power quality. However, such analysis may be impaired by the presence of inter-harmonics, which is ever-growing due to the insertion of power electronic converters in the grid, and causes spectral leakage; despite the efforts of the scientific community in proposing solutions for the mitigation of this problem, there is still no consensus. This work proposes a solution based on Multilayer Perceptron Artificial Neural Networks for inter-harmonic diagnosis. The solution can properly identify and estimate the frequency of inter-harmonics, and the results for synthetic signals indicate the effectiveness of the proposed technique, with a hundred percent identification rate and a less than 1.3 percent relative error.

Keywords: Power and Eneergy Systems
Abstract: This paper investigates the lightning performance of three soil models considering their electrical parameters dependent on the frequency and water content. These soil models are Smith-Longmire, Scott, and Messier models. First, the harmonic grounding impedance of a 30-m rod buried in stratified soil is analyzed using the full-wave electromagnetic software Altair FEKO for a frequency range of 100 Hz to 10MHz. Then, their impact on the ground potential rise (GPR) waveforms for different lightning currents is assessed. Finally, a backflashover analysis is carried out for a transmission line on stratified soil.Results indicated a notable variation in the harmonic impedance above 1 MHz. This variation has a pronounced impact on the GPR waveforms for different currents, especially in the peak values when the three soil models are used compared with the frequency-constant soil model.Finally, transient analysis for the transmission line demonstrates that depending on the lightning current waveform and soil model, a backflashover occurs and may lead to an outage in the power system.

Keywords: Power and Eneergy Systems, Renewable Energy, Power Quality
Abstract: Em estudos de ilhamento, uma prática recorrente é considerar somente uma abertura espontânea do religador criando a ilha, por se tratar da situação mais conservadora, embora seja uma ocorrência muito rara. Como consequência, os resultados destes estudos podem gerar conclusões demasiadamente conservadoras. Neste trabalho uma série de cenários foi definida, com curtos-circuitos monofásicos e trifásicos, com e sem impedância de falta e em locais diferentes da rede sendo os responsáveis pela abertura do religador e criação da ilha. Este trabalho tem como objetivo avaliar o impacto do curto-circuito como evento causador do ilhamento de geradores distribuídos. Portanto, foi possível notar como o bloqueio por subtensão nas funções de frequência pode evitar a atuação incorreta das funções de proteção de frequência dos geradores distribuídos para cenários em que a rede está sob falta e, além disso, o trabalho traz a função de sobretensão de neutro para detecção de faltas mesmo em casos com maior impedância de falta.

Keywords: Electrical Machines and Drives, Industry Applications
Abstract: The investigation for reducing power consumption of the three-phase induction motors is a promising solution to handle the increasing energy demand in the electric sector. However, modern processes require high-precision performance, which is challenging from an energy perspective. With this scenario, this study investigates the efficiency of a three-phase induction motor by employing predictive control strategies. The objective of this study is to explore the performance of the motor at different load and speed-torque operating points while analyzing the effects of different magnetic flux references. Two predictive control strategies were simulated using the mathematical model of the induction motor to enable data collection and subsequent analysis of system behavior. The results show that it is possible to achieve better energy utilization while maintaining dynamic performance.

Keywords: Diagnosis, Prognosis and System Identification, Electrical Machines and Drives, Industry Applications
Abstract: Three-phase induction motors(TIMs) are the main source of mechanical energy in industries. Due to the advent of voltage inverters and advances in power electronics, these devices can be operated over a wide range of speeds. However, TIMs are often exposed to failures, which require different types of maintenance. When an industrial process is interrupted by a motor failure, there can be significant financial losses. Therefore, researchers seek to develop sensor-based approaches to detect mechanical and electrical faults to avoid losses and damages. Different methods have been proposed. However, the acoustic emission (AE) method attacks attention due to new affordable sensors and easy application. Also, one of the most common electrical faults in TIMs is unbalanced voltages. It causes torque losses, overcurrents, and excessive vibration. Currently, there are already several motors being monitored by AE equipment. However, unbalanced voltages are not considered by any of them, which can lead to false diagnoses. Therefore, this work focuses on developing a new method able to detect, identify the affected phase, and classify the magnitude of unbalanced voltages in TIMs. Additionally, a new signal processing technique was proposed, which led to a new index to mitigate the high acoustic noise found in industrial plants. Several experiments were carried out, and the results showed the efficiency of the novel technique in performing the detection, phase identification, and magnitude classification of unbalanced voltages in induction motors.

Keywords: Electrical Machines and Drives, Electrical Vehicle and Energy Storage, Diagnosis, Prognosis and System Identification
Abstract: Normas técnicas são utilizadas internacionalmente para definir e padronizar a obtenção dos parâmetros de modelos de máquinas elétricas. Estes modelos são fundamentais para a integração em projeto de sistemas eletromecânicos complexos, como no caso de máquinas para veículos elétricos. O uso de ímãs em máquinas elétricas está associado ao seu alto desempenho, tanto em máquinas de correntes contínua com escovas como sem escova, bem como em máquinas síncronas. Em veículos elétricos, as máquinas de ímãs permanentes tem ganhado um espaço crescente devido à sua alta densidade de potência. Entretanto, não há norma para identificar os parâmetros dos modelos dessas máquinas de forma padronizada. Este artigo apresenta um método baseado no comportamento em diversas condições de funcionamento utilizando algoritmo genético para identificar os parâmetros de um modelo clássico de máquinas síncronas, considerando aspectos característicos dos ensaios. Os resultados obtidos indicam que é importante considerar o efeito da corrente e da rotação nos parâmetros obtidos, particularmente devido à sensibilidade dos ímãs à temperatura da máquina e no nível de saturação localizado.

Keywords: Electrical Machines and Drives, Electrical Vehicle and Energy Storage
Abstract: In many applications where torque is required in limited physical spaces, a permanent magnet synchronous machine (PMSM) is usually chosen because of its high power density characteristic. One example of this is in electrical propulsion, where high torque and speeds are needed. However, since PMSMs present difficulties in operating above their nominal speed, flux weakening methods are used to decrease the overall magnetic linkage flux, thus allowing the motor to accelerate. The most common method is called space vector flux weakening, and it can result in permanent magnets (PMs) demagnetization. In this paper, the equivalent circuit of a PMSM is explored, leading to the voltage equations used to explain the concept behind the space vector flux weakening method. Finite Element Method (FEM) simulations are executed to compare the demagnetization risk of two permanent magnet synchronous motors used in electrical vehicles (EVs) and hybrid electrical vehicles (HEVs), where flux weakening methods are applicable and relevant. Finally, torque analyses after demagnetization are also presented to understand performance impacts.

Keywords: Electrical Machines and Drives, Diagnosis, Prognosis and System Identification, Machine Learning
Abstract: Bearing failure detection in induction motors is one of the most important processes in the industry. When bearing wear reaches a critical level, the machine collapses, and the production process must be stopped. Since bearings are the most susceptible to failure, several works have proposed a solution for their early diagnosis. However, most of these methods employ signal-processing techniques to extract relevant information from the acquired data. This requires a significant computational effort that can make hardware implementation impractical. In this paper, we propose to extract features directly from vibration time series to overcome this problem. The developed method evaluates the relevance of the temporal measures and iteratively removes those that have a lower impact on the classification. Three techniques are compared to estimate the feature importance, including logistic regression (LR), mutual information (MI), and variance inflation factor. Classifications using decision trees, support vector machines, multilayer perceptron, and k-Nearest Neighbors demonstrate that a smaller set of attributes can provide a more accurate classification than the original set of attributes. In addition, this work compares vibration signals acquired by accelerometers and an industrial vibration sensor. The experiments show that only seven temporal attributes from vibration signals are sufficient to achieve a success rate of about 93% in detecting incipient bearing wear, considering different load and voltage unbalance levels.

Keywords: Industry 4.0, Industry Applications, Industry Data Science Applications
Abstract: The objective of this project is a study and a practical application of a monitoring industrial system, the purpose is to do the monitoring using a Supervisory System, of an industrial process that use two industrial communication standards. The supervisory system used is Elipse E3, and the industrial communication standards were PROFINET and IO-Link.

Keywords: Internet of Things
Abstract: This work discusses data segmentation and compression strategies for sending waveforms through class A LoRaWAN (Long Range Wide Area Network) sensor networks. It proposes an efficient data segmentation algorithm that evaluates the missing data segments only at the end and then retransmit the missing packets. Furthermore, it presents a data compression algorithm implementation based on “lossless entropy compression” (LEC) principle coded in C language, suitable to be executed in microcontroller systems with a permissive license and bindings for use in the Python language. Moreover, it was also conducted practical tests with an inertial sensor to prove the effectiveness of this proposal.

Keywords: Internet of Things, Life Support Systems & Techniques, Cyber Physical Systems, Digital Twins and Knowledge Systems
Abstract: Autistic Spectrum Disorder (ASD) involves several disturbances that affect speech skills and social interaction. People with this syndrome have difficulty transmitting their wants and needs daily, as they delay speech development and difficulty forming sentences. Furthermore, they face other obstacles such as aversion to eye contact, difficulty concentrating, talking about their feelings, or understanding the feelings of others, and resistance to physical contact. Because of these characteristics, children can be socially isolated. Thus, monitoring the treatment of these children is essential. Therapists estimate that 30 to 40 hours of weekly treatment activities are needed. Such activities can be done with the therapist's support, with the teachers at school, and in the family environment. In this paper, we present the AuToy, a system that aims to connect parents and therapists to help in the daily lives of children who have Autistic Spectrum Disorder. Using the Design Science Research method, we develop a device that is an intelligent toy to assist in the performance and feedback of ABA – Applied Behavior Analysis treatment exercises. Also, we conducted experiments on controlled environments, performing unit, integration, and full systems tests to validate the system. We discuss the consumption of two types of developed devices. One was made with RFID readers, and the other with a resistor card reader. By developing intelligent toys, we hope to help promote a child's social interaction with parents and therapists, which is also a significant part of the treatment. Also, it creates a better interaction between technology and neurodevelopment.

Keywords: Virtualization, Simulation Techniques and Augmented Reality, Industry 4.0, Additive Manufaturing
Abstract: With the advent of the Internet of Things and Cyber Physical Systems in recent years, simulations has changing from models of limited lifetime to more dynamic approaches of constant use in which data flows between physical and virtual systems in real or near-real time. It has opening the doors to a concept called Digital Twin that seeks not only narrows the communication gap between these worlds but also allowing faster decision making in the real world based on data processed in its virtual counterpart. This paper proposes the prototype of a Digital Shadow (first step to achieve a Digital Twin) applied in a smart shelf, with the goal to capture its inventory level in nearreal time and produce useful data to be processed and analyzed in the virtual world. The paper have shown great results in tying together both worlds, specially by using popular hardware and software platforms such as Arduino and FlexSim, respectively.

Keywords: Virtualization, Simulation Techniques and Augmented Reality, Power and Eneergy Systems, Optimization Heuristics and Methods
Abstract: Due to increased high penetration of renewable energy resources in power systems, the concept of Virtual Power Plant (VPP) has been proposed to aggregate Distributed Energy Resources (DER) to act like a single power plant. In this context, the purpose of this work is the development of a method to aggregate DER units, which use the interface with the electrical network by converters, representing them by equivalent VPPs. In addition to this representation, it is necessary to estimate the parameters of the equivalent VPP model. The method used for this purpose is based on the Gray-box method and a Differential Evolution (DE) algorithm, which allows the use of measurement data that can be performed at the boundary between the distribution system and the system under study. The model is validated using the MATLAB/SIMULINK software to simulate a test network and obtain parameters of the equivalent model.

Keywords: Virtualization, Simulation Techniques and Augmented Reality
Abstract: The phenomenon of magnetohydrodynamics (MHD) is the study of the behavior that a conducting fluid will exhibit when it is flowing in the presence of an external magnetic field. The MHD has several applications in the industrial sector, as well as for studies of various types of materials such as ferrofluids. Thus, this work aims to analyze how the MHD phenomenon affects both the mechanical and electromagnetic behavior, using computational simulations by the finite element method and the finite volume method. As well as develop a system for measuring the values of the magnetic field using Hall effect sensors connected to a Esp32.

Keywords: Life Support Systems & Techniques, Virtualization, Simulation Techniques and Augmented Reality, Industry 4.0
Abstract: This paper aims to present initial results from the analysis of biosignals to demonstrate the effectiveness of essential oils and the utilization of virtual reality (VR) for the relief of stress and anxiety symptoms. The methodology, with protocols endorsed by the ethics board, consists in capturing neurological signals via electroencephalography (EEG) and electrodermal activity (EDA) signals from a volunteer diagnosed with anxiety. The EEG and EDA signals are collected using Emotiv Epoc X and Biosignalsplux equipment, respectively. The results indicate the effectiveness of the essential oils in conjunction with VR, reinforcing the scientific basis for this complementary treatment approach. In addition, this paper presents a brief narrative review on the subject, as well as, presents proposals for future investigations.

Keywords: Artificial Intelligence, Deep Learning and Machine Learning, Machine Learning
Abstract: In this review, we delve into the advancements and applications of monocular depth prediction, focusing on the intricacies of the protocol experiments, dataset transfer, and fine-tuning. We highlight the incorporation of gradient matching loss and the significance of employing a depth inverse representation. Furthermore, our study underscores the expansive utility of such predictive models, not only in surveillance within smart city paradigms and public security but also in diverse domains such as agribusiness, education, and biomedical engineering. Leveraging datasets like MIX 5, MIX 6, and ADE20K, we elucidate the performance enhancement achieved by Transformer-based Neural Networks in dense prediction tasks. Through these findings, the potential for Transformer-based models in practical applications, respecting data protection norms like LGPD, becomes evident. The collaborative efforts of academic and domestic partners in realizing this review are also acknowledged.

Keywords: Artificial Intelligence, Deep Learning and Machine Learning, Machine Learning
Abstract: Colorectal cancer (CRC) is a leading cause of cancer-related death worldwide, with a strong correlation to the abnormal growth of tissue known as polyps. These polyps exhibit various shapes and sizes, such as flat, elevated, or pedunculated. This paper presents a deep learning approach for polyp segmentation in colonoscopy images, based on the U-net architecture and incorporating pre-processing and post- processing techniques. The approach aims to assist healthcare professionals in the accurate and early detection and diagnosis of CRC. The experiments achieved values of 0.962 for Precision and 0.948 for Recall, demonstrating the potential of the proposed model as a valuable tool for enhancing patient outcomes, reducing public health expenses, and improving overall colonoscopy quality

Keywords: Optimization Heuristics and Methods, Power and Eneergy Systems
Abstract: Muitos pesquisadores vêm empregando algoritmos meta-heurísticas para resolver problemas complexos de engenharia há décadas. A grande maioria dos pesquisadores proclama que a técnica utilizada é eficiente e superior às outras empregadas. Este artigo apresenta uma investigação comparativa de duas meta-heurísticas, a SOS e a PSO, aplicadas ao problema de RRD de energia elétrica, que é um problema de PNLIM de grande complexidade. A SOS é uma meta-heurística recente, enquanto a PSO é uma meta-heurística tradicional, ambas com estratégias distintas de busca pelo ótimo. As duas meta-heurísticas foram avaliadas sob as mesmas considerações, buscando destacar os pontos positivos e negativos de cada uma. Os testes foram realizados em dois sistemas diferentes, um com 33 nós e outro maior, com 110 nós, ambos amplamente estudados na literatura especializada. Tanto a SOS coma a SOS apresentaram uma porcentagem alta de sucesso, no entanto, em geral, a PSO teve um desempenho superior em relação a SOS.

Keywords: Artificial Intelligence, Industry 4.0
Abstract: A detecção e rastreamento no trânsito para o controle semafórico é uma das chaves para o desenvolvimento das cidades inteligentes, desafogando o trânsito, realizando um uso mais racional dos recursos e aumentando a qualidade de vida da população em geral. Um contexto pouco explorado no estudo de detecção e rastreamento de elementos no trânsito é o uso de vídeos omnidirecionais (com 360° de campo visual). Esse trabalho apresenta uma base de dados que pode ser facilmente utilizada para o treinamento de algoritmos de detecção e rastreamento, em imagens de 360°, retratando todo o entorno com apenas uma câmera e economizando recursos computacionais e banda de transmissão. Por fim, também ́e apresentado um benchmark inicial, fornecendo, assim, uma base para comparações futuras. Para isso, utilizou-se a rede YOLOv7 para a detecção e a rede DEEPSORT para o rastreamento dos objetos no trânsito.

Keywords: Industry Data Science Applications, Deep Learning and Machine Learning, Smart Grids
Abstract: Electrical load forecasting is crucial for the efficient planning of electricity companies. Load forecasting traditionally refers to forecasting the expected electricity demand at aggregated levels. More recently, with the rise of high granularity data, load forecasting has shifted towards a disaggregated approach. However, forecasting disaggregated electrical loads such as residential or commercial buildings is a challenge due to high variability in load curves influenced directed by individuals. This study proposes using general regression neural networks to forecast disaggregated electrical loads from an electric power distribution system in São Paulo for four consumption profiles: Residential, Commercial, Municipal Public Power and Public Water Service. In general, the Mean Absolute Percentage Errors (MAPE) obtained in the forecast of the Commercial, Municipal Public Power, and Public Water Services profiles using the GRNN were consistent with the results found in the literature. These MAPEs were, on average, 12,97%, 14,39% and 9,11%, respectively. When using the modified GRNN, lower or equal MAPEs were obtained with a lower training and validation time. The MAPEs obtained in the forecast of the residential profile with the GRNN, on the other hand, averaged 144.13% and demonstrated that the forecasting system was not able to generalize the behavior for these loads.

Keywords: Artificial Intelligence, Power and Eneergy Systems, Machine Learning
Abstract: This work presents the application of an artificial neural network to estimate the total real power losses of an electrical power system. The network used is the Multilayer Perceptron composed of 3 neurons in the input layer (loading factor, real and reactive power generated in the slack bus), 10 neurons in the intermediate layer and 1 neuron in the output layer, representing the total real power losses. The training used is the backpropagation, which uses the desired output (target) to adjust the weights. From the results (IEEE systems of 14 and 30 buses), the network performed well, with mean squared error around 10-4, and R-value at 0.99. For validation and testing, with 20% of samples, the network proved to be efficient, with MSE also around 10-4. In this context, the network was able to estimate the total real power losses in function of loading, showing that the obtained output was very close to the one desired.

Keywords: Deep Learning and Machine Learning, Industry Applications, Artificial Intelligence
Abstract: A importância do transporte ferroviário na economia e na logística mundiais destaca a manutenção preditiva como forma de garantir a disponibilidade dos equipamentos e a segurança da operação. No caso do material rodante, essa manutenção pode ser feita por meio de um monitoramento contínuo e automático, utilizando equipamentos específicos chamados de waysides. Assim, este artigo propõe um sistema modular de inspeção automática de rodeiros, usando técnicas de visão computacional, como detecção de objetos e segmentação semântica, em imagens capturadas pelos waysides. O pipeline construído detecta os objetos de interesse e envia cada componente para um módulo específico responsável pela sua inspeção. São inspecionadas a espessura da bandagem, a presença ou ausência de pads e a altura das molas. Ao final da passagem de cada trem, é gerado um relatório para a visualização dos resultados. Os resultados preliminares são promissores e algumas melhorias e extensões do pipeline são previstas como trabalhos futuros.

Keywords: Power and Eneergy Systems
Abstract: Lightning Induced Voltages (LIVs) are one of the main causes of distribution lines shutdown. It is known that soil resistivity plays an important role in the LIVs on overhead lines. Furthermore, in recent years, the effects of water content and soil porosity are receiving attention as it influences soil resistivity and therefore LIVs in overhead lines. Therefore this paper investigates the influence of the porosity level proposed by Archie’s formula on the lightning performance of distribution lines. For this, the extended formula proposed by Paulino, which determines the peak value of the LIV combined with the Monte Carlo Method were used to estimate the distribution line lightning flashover rate. In this way, simulations were performed using a computational algorithm with different values of water content and porosity and the flashover rate was calculated based on these values. Results have demonstrated that the volumetric water content and soil porosity have a significant impact on the calculation of the performance of distribution lines. Increasing the water content and decreasing the porosity level produce more conductive soils that notably impact the performance responses.

Keywords: Power and Eneergy Systems, Power Quality, Machine Learning
Abstract: The Power Quality analysis in the Electrical Power System is very important in order to promote a stable system. Thus, this paper introduces a neural network-based method to classify different types of disturbances sample by sample. In the classification process, we used Power Quality parameter estimation techniques to extract the features to be used in the model. The model was generated and validated, obtaining satisfactory results.

Keywords: Renewable Energy, Deep Learning and Machine Learning, Artificial Intelligence
Abstract: With the aim of reducing the monitoring time of results from twelve to six months, a methodology using LSTM (Long-Short Term Memory) neural networks for solar irradiation prediction is proposed in this paper. This is important for measurement and verification campaigns in energy efficiency projects, particularly in solar generation projects that depend on solar irradiation. The evaluation of the method in photovoltaic system projects showed satisfactory results, identifying an alternative solution for measurement and verification in energy efficiency projects.

Keywords: Renewable Energy
Abstract: In wind power generation, the wind farm availability index is directly related to failures in wind turbines and is one of the fundamental monitoring indices for the entrepreneur, the National Grid Operator (ONS), and the Brazilian Electric System Regulatory Agency (ANEEL). Financial penalties are applicable to energy generating companies that fail to comply with availability and dispatch of energy indexes defined for the project. The objective of this study is to correlate the most common failures of a group of wind turbines with data obtained from the average wind speed, available in the Supervisory Control and Data Acquisition System (SCADA). A new index is proposed based on the standard deviation of the wind speed mean, instead of the wind measurement from the anemometer tower of the wind farm. The fact that this tower may be far from some wind turbines within the same wind farm can cause a large difference between the turbulence measured at the anemometer tower and the turbulence that the wind turbine is actually experiencing. This study uses the 10-minute average speed of each wind turbines’ anemometer as a method to estimate wind standard deviation. To validate the method, the pitch control failure rate of the Senandes and Vento Aragano wind farms was measured between October 2017 and February 2019. From these data, it was possible to verify a high correlation between the probability of failure and the proposed index.

Keywords: Smart Grids, Power Electronics, Power and Eneergy Systems
Abstract: Microgrids are an increasingly widespread approach to grid configuration that allow for more flexible integration of renewable energy into the power grid. For a microgrid to be able to operate in islanded mode, it must contain a minimum amount of grid-forming generators. While traditional synchronous machines are naturally grid-forming, most Distributed Energy Resources (DERs) are inverter-based and, commonly, grid-following. This situation presents a challenge for distribution systems, which are expected to become increasingly reliant on distributed and renewable sources.

Self-adaptive droop is a control technique for inverter-based systems (especially electric Energy Storage systems) which allows an energy source to smoothly transition between grid-connected and islanded operation modes. Much like traditional droop, this control algorithm has non-linear elements.

This paper presents a small-signal model for a self-adaptive droop controlled inverter. Based on this small-signal model, a stability study is conducted in order to better select the droop control gains (Kip and Kiq) in accordance to the desired dynamic response.

Keywords: Smart Grids, Power and Eneergy Systems, Large Scale and Network Control
Abstract: A performance adequada do sistema elétrico de potência (SEP) depende de diversos parâmetros e requisitos, principalmente aqueles atrelados à segurança e estabilidade de tensão da rede. Nesse contexto, a Margem de Estabilidade de Tensão configura-se como a métrica principal para análise, proteção, controle e segurança de tensão. Para realizar as análises é necessário investigar as contingências e perturbações que podem ocorrer no sistema. Essas investigações podem ser realizadas por meio de simulações com o uso do fluxo de potência continuado. Um dos dados que se pode obter do fluxo é a margem de estabilidade de tensão, sendo uma das formas de se verificar se o sistema apresenta estabilidade e segurança. Em situações de instabilidade ou de insegurança, Sistemas Especiais de Proteção com estratégia de Corte de Carga (CC) devem ser implementados para manter a integridade de parte da rede elétrica. Neste trabalho é proposto um novo algoritmo baseado no fluxo de potência continuado que promove o corte de carga para contingências instáveis para um dado carregamento inicial e estáveis que não respeitam os limites da margem de estabilidade de tensão. Com intuito de verificar a eficácia da metodologia foram feitas simulações para os sistemas modelos de 14 e 57 barras do IEEE. Os resultados obtidos demonstraram que o algoritmo desenvolvido foi capaz de implantar o CC para os casos analisados e a tensão foi estabilizada com segurança.

Keywords: Smart Grids, Artificial Intelligence, Power and Eneergy Systems
Abstract: Este artigo apresenta um método preciso para identificar cargas residenciais de forma não intrusiva. O desempenho é analisado a partir de um conjunto de dados com cinco cargas de execução correspondentes a cinco classes. As Estatísticas de Ordem Superior e a Máquina de Vetor de Suporte (SVM) executam a extração e a classificação de recursos, respectivamente. Os resultados mostraram que os modelos foram concluídos satisfatoriamente, destacando-se o desempenho do método de extração dos parâmetros de trabalho utilizando a Estatística de Ordem Superior. Os resultados obtidos neste estudo foram considerados excelentes, apresentando 100% de acerto para a maioria dos dados analisados. No entanto, observou-se que para a classe C2 - LED TV, o desempenho do classificador apresentou índices um pouco menores, mas ainda assim considerado excelente.

Keywords: Electrical Machines and Drives, Cyber Physical Systems, Digital Twins and Knowledge Systems, Industry Applications
Abstract: Induction electric motors are the main load in the industry globally. Their robustness has allowed their widespread use worldwide given their tolerance to different power supply conditions. Voltage variation is a silent disturbance in the power supply of several electric motors and inevitably produces deviations in the motor response. The impacts of this disturbance can be better appreciated from the magnetic assessment and its validation with experimental tests. Based on the above, this paper presents a magnetic assessment of the voltage variation in electric motors using the multiphysical FEM software Elmer, then the results have been validated from measuring campaigns. The impact evaluation of these disturbances becomes important given the existing regions with different supply voltages according to IEC 60038-2009.

Keywords: Electrical Machines and Drives, Artificial Intelligence, Automation and Process Control
Abstract: Speed measurement in Induction Motors (IM) consists in a major barrier to the implementation of vector control in low-power applications. To overcome this issue, many estimation strategies were developed. In particular, the Extended Kalman Filter (EKF) provides good results with a relatively low computational cost. This work proposes the application of a 6th-order EKF to make rotor speed and rotor flux estimation in an RFOC-MPCC (Rotor Field Oriented Control - Model Predictive Current Controller). The effectiveness of this approach was evaluated in two stages. First, the accuracy of the EKF in estimating the rotor speed and rotor magnetic flux was assessed. Then, the performance of the control system using data from the EKF was analyzed. The results from both analyses suggest that proposed scheme is a promising method for sensorless control of induction motors.

Keywords: Electrical Machines and Drives, Industry Applications
Abstract: Centrifugal pumps are utilized in various applications that require flow control during operation. In pumping systems, flow control can be achieved by controlling the speed of the motor coupled to the centrifugal pump or by throttling a valve. In this paper, a comparison of the dynamic performance of a 12/8 switched reluctance motor for the two methods of flow control is presented. To ensure the performance of the switched reluctance motor in the flow control of the pumping system, the optimization of the turn-on angle and the DC-link voltage was performed. These results were compared and the switched reluctance motor under speed control was more energy efficient than the valve control. At a flow rate of 20.57 m³/h, using the flow control method through valve regulation, the electrical power in the motor was 176.28% of the electrical power from the method using speed control.

Keywords: Electrical Machines and Drives, Electrical Vehicle and Energy Storage, Industry Applications
Abstract: Linear generators are complex machines with multiple structural parameters that significantly impact the electrical values generated during energy conversion. Based on this, this study presents a sensitivity analysis of which parameters have the greatest influence on the flux density, a variable that directly affects other electrical values such as flux, electromotive force, and force. In this paper, a total of ten structural parameters were selected for study with the aim of analyzing and selecting a combination that would allow the design of a new machine with improved electrical values compared to the reference machine. By using FEMM software together with MATLAB, it was possible to vary these ten parameters through lines of code, obtaining multiple models for the electric machine. Changes to the structural parameters were applied and their contribution to the improvement of flux, electromotive force, and force values were observed. It was noted that the adjustments of the four most influential parameters in the improvement of electrical values had a slight effect on the final size of the generator. The air gap was reduced by 30%, the slot opening width was increased by 60%, and the heights of the magnets and translator were increased by 30%, which consequently contributed to a slight increase of approximately 7.51% in the amount of magnetic flux compared to the reference machine, an increase of 21.41% in the electromotive force, and also an increase of approximately 68.22% in the force caused by changes in flux densities in both directions (Br and Bz).

Keywords: Electrical Machines and Drives, Industry Applications, Power Quality
Abstract: Voltage imbalance remains one of the secret killers of electric motors considering its multiplying effect on current magnitudes. Its impacts on conventional induction motors have been widely addressed, however considering the new technologies existing commercially as a result of the search for higher efficiencies, the evaluation in non-ideal conditions also becomes of significant interest to users. This paper presents an evaluation of the voltage unbalance impacts on a line-start permanent magnet motor (LSPMM). For the study, the motor was fed under different voltage unbalance conditions and varying the motor load percentage. The results revealed that for high levels of voltage unbalance in the motor phases, the current unbalance in combination with the permanent magnet's constant magnetic field produces negative single-phase power under low-load conditions. Power quality impacts are also addressed from current oscillography.

Keywords: Electrical Machines and Drives, Industry Applications, Power Quality
Abstract: This paper presents a comparison of the performance of squirrel cage induction motors (SCIMs) and permanent magnet motors (LSPMMs) under typical operating conditions. The effects of voltage unbalance and different rated load conditions are examined to determine the motors' performance and tolerance to disturbances. Results show that SCIMs demonstrate slightly higher tolerances to disturbances than LSPMMs in non-ideal voltage conditions, thus providing insight into their effectiveness in various applications. This comparison is crucial for a better understanding of motor performance and its implications for the design and operation of motors in certain operating conditions and regions.

Keywords: Electrical Machines and Drives, Industry Applications
Abstract: No interior das máquinas elétricas são formados acoplamentos capacitivos entre as superfícies dos condutores ali presentes. Considerando o acionamento via conversor de frequência, as componentes de alta frequência da tensão de modo comum excitam esses acoplamentos, que passam a se apresentar como caminhos de baixa impedância, permitindo o surgimento de correntes de alta frequência no interior do motor. Essas correntes podem causar danos para o funcionamento da máquina, que vão desde interferências eletromagnéticas à inutilização de seus mancais. Dessa maneira, a determinação das capacitâncias parasitas é crucial para a elaboração de medidas preventivas visando à solução de alguns problemas relacionados ao acionamento do motor. Porém, estas capacitâncias podem variar dependendo da posição das espiras no interior das ranhuras, o que torna difícil a obtenção destes parâmetros, sobretudo para motores de menor potência, nos quais as bobinas são formadas por espiras sem um posicionamento fixo no interior da ranhura (random-wound winding). Neste contexto, a partir de um modelo em 2-D que utiliza o método dos elementos finitos, este trabalho realiza o cálculo das principais capacitâncias parasitas que são formadas no interior de um motor de indução trifásico com rotor em gaiola. Apresenta-se uma rotina computacional desenvolvida para a determinação das capacitâncias levando em consideração a aleatoriedade no posicionamento das espiras das bobinas. A partir dos resultados obtidos, é possível identificar faixas de valores mais prováveis para as capacitâncias. Além disso, os resultados encontrados com o modelo proposto são comparados com valores determinados a partir de formulações analíticas encontradas na literatura.

Keywords: Power and Eneergy Systems, Machine Learning, Diagnosis, Prognosis and System Identification
Abstract: The investigation of partial discharges (PDs) in transformers windings provides early identification and localization of failures in these equipment, which can avoid considerable financial losses. For the localization techniques development, however, a fundamental step refers to obtaining signals and features, which can directly influence the performance of the applied localization method. In order to identify a better set of features for the PDs location, statistical and principal component features are investigated in this work, regarding the ability to distinguish different signals classes and the applicability in the most employed transformers windings configurations (stacked disks and concentric layers). Employing the Multi-conductor Transmission Lines modelling, simulated PD signals were obtained, from which features were extracted. The results show, although visually, that the statistical features present a better spatial separation of the current signals of the PDs, being more suitable for the location of such discharges, mainly in disk-type transformer windings.

Keywords: Power and Eneergy Systems, Optimization Heuristics and Methods, Renewable Energy
Abstract: The main tools for the operation and planning of electrical power systems are Power Flow (PF) and Optimal Power Flow (OPF). However, the analysis performed by these tools delivers deterministic values, that disregard uncertainties that occur in electrical systems. Such uncertainties are mainly caused by data error of lines, transformers, demand, and generation of alternative sources forecasting. A strategy to incorporate these uncertainties in the tools is through Interval Math (IM), which allows the inclusion of data intervals instead of single points. This strategy is performed only once, unlike probabilistic methods that run exhaustive simulations. Thus, the objective of this work is to incorporate IM into a Multiperiod Optimal Power Flow (MOPF)

to consider load and generation data uncertainties, after the deterministic optimization. The uncertainty ranges are added via Krawczyk Method, which determines an optimal interval of operation. The MOPF realizes the dispatch of power generation and reserve spinning along 24 hours of a hydro-thermal-wind

system. The proposed method was compared with results obtained from traditional exhaustive simulations carried out with the southern Brazilian system with 33 Buses. The results obtained were close to the values of the random analysis, with uniformity.

Keywords: Power and Eneergy Systems, Renewable Energy, Power Electronics
Abstract: Diante da crescente penetração de Geradores Distribuídos (GDs) no sistema elétrico, diversas documentações internacionais vêm sendo estabelecidas com o intuito de definir requisitos mínimos de suportabilidade de GDs para garantir uma operação confiável do sistema de proteção anti-ilhamento. Neste contexto, este trabalho avalia a influência da janela de medição da taxa de variação de frequência na atuação da função ANSI 81R. As análises foram realizadas a partir de simulações de eventos de ilhamento e faltas em um sistema em média tensão contendo Gerador Síncrono Distribuído (GSD) ou Gerador Baseado em Inversor (GBI). Os estudos possibilitaram constatar uma influência significativa da janela nas atuações para faltas, não sendo tão significativa para a detecção do ilhamento, assim como evidenciar a importância da consideração deste parâmetro na parametrização da função de proteção.

Keywords: Power and Eneergy Systems, Renewable Energy, Self Configuration & Self Diagnosis
Abstract: With the increasing adoption of Industry 4.0 in industrial power systems, the integration of distributed generation (DG) into distribution electrical grids is becoming more relevant. However, the presence of DG can have significant impacts on protection parameters, causing sudden power flow variations and increasing fail occurrences. Therefore, detailed studies are essential to determine the effects of DG connections on the electrical grid. This work presents an analysis of the impacts of photovoltaic DG on phase-timed overcurrent relay parameters and proposes the use of an adaptive protection system to optimize the coordination of protection devices. Additionally, it emphasizes the importance of proper protection of industrial power systems and the role of new technologies in improving system efficiency and safety. The numerical results were satisfactory, demonstrating several practical comparisons.

Keywords: Power and Eneergy Systems, Power Quality, Renewable Energy
Abstract: The electrical system is susceptible to faults whose currents have low magnitude and, consequently, do not sensitize the traditional overcurrent protection relays. Generally, high impedance faults (HIF) are responsible for these events and their occurrence is due to the contact of the energized cable with low conductance surfaces, which does not result in a current increase sufficient to exceed the pick-up value of the protection. conventional. A second factor favorable to these defects are distributed photovoltaic generators (GDFV) with low voltage ride through (LVRT) that can contribute to the fault in a possible short circuit on the same line to which the distributed generator is connected. To overcome such failures, this paper will investigate a differential protection that works with the positive sequence components of the current and with the least squares method (MMQ) to detect and isolate short circuits with low currents. This protection philosophy, as well as the active distribution network (RDA) were implemented and simulated in Matlab/Simulink® software.

Keywords: Power and Eneergy Systems, Renewable Energy, Smart Grids
Abstract: The growth of distribution systems with the inclusion of distributed generators (DG) is a reality that is taking place in several countries. The use of DGs in these systems brings benefits and challenges. One of these challenges is the effect of DGs on overcurrent protection, which makes it important to investigate through computational tools that evidence such effects and serve as a support for evaluating the necessary adaptations by the protection engineer. So, this work is a complementing of an earlier conference paper [1] that contains simulation analyses to evaluate these questions. For this, overcurrent protection devices were incorporated into the original system, in the places where DGs were inserted, for an evaluation of the network's behavior. The impact of these DGs was evaluated in the application of symmetrical three-phase faults in specific locations of the system, for a given DG size. An analysis of current variation, selectivity and protection coordination was performed in numerical and graphical form, in different scenarios. From the results obtained, it was observed that new settings for the protection devices are necessary so that the system remains correctly coordinated and protected with the inclusion of DGs.

Keywords: Power and Eneergy Systems, Artificial Intelligence, Smart Grids
Abstract: Several factors have propitiated the expansion of electrical systems and this expansion causes a greater complexity of operation. The operating complexity of current power systems increasingly requires the development of indices capable of identifying instability mechanisms in the long term. The load margin of power systems is an index to evaluate the distance that the power system is from a stability problem. Usually, this index employs static models that govern the power flow of the system and aims to determine the voltage collapse condition. However, low-frequency oscillation modes present in dynamic models can also cause instability in power systems when they are not well damped. Thus, the load margin traditionally calculated by static models can provide information on the existence of operating points in the system, but these operating points can present oscillation modes with low damping rates that can be observed in the dynamic analysis of systems often subject to contingencies and these modes can affect the stability of operation of power systems. Thus, it is important to inform the operation center of a load margin index value that satisfies the requirements of the existence of equilibrium points in voltage stability studies and well-damped oscillation modes in small-signal rotor angular stability studies. This paper proposes a new method based on Recurrent Neural Networks (RNNs) to evaluate the load margin with evaluation of the requirements of small-signal and voltage stability. The proposed method of this paper was applied on IEEE 39-bus for different scenarios. The obtained solutions demonstrate a good ability to generalize the RNN and classify the load margin ranges.