Teses em Engenharia Elétrica (Doutorado) - PPGEE/ITEC
URI Permanente para esta coleçãohttps://repositorio.ufpa.br/handle/2011/2317
O Doutorado Acadêmico inicio-se em 1998 e pertence ao Programa de Pós-Graduação em Engenharia Elétrica (PPGEE) do Instituto de Tecnologia (ITEC) da Universidade Federal do Pará (UFPA).
Navegar
Submissões Recentes
Item Acesso aberto (Open Access) Intent-based radio resource scheduling in ran slicing scenarios using reinforcement learning(Universidade Federal do Pará, 2024-11-04) NAHUM, Cleverson Veloso; KLAUTAU JÚNIOR, Aldebaro Barreto da Rocha; http://lattes.cnpq.br/1596629769697284Network slicing at the radio access network (RAN) domain, called RAN slicing, requires elasticity, efficient resource sharing, and customization to deal with scarce and limited frequency spectrum resources while fulfilling the slice intents in an intent-based system. In this scenario, radio resource scheduling is an essential function to provide the resource management needed to prevent intent violations, hence providing sufficient radio resources for RAN slices to accomplish their intents. The wide variety of scenarios supported in 5G and beyond 5G (B5G) networks makes the radio resource scheduling (RRS) problem in RAN slicing scenarios a significant challenge. This thesis proposes an intent-based RRS for RAN slicing using reinforcement learning (RL) to fulfill the slice intent. The proposed method aims to prevent intent violations by making the management of resource block groups (RBGs) available between slices and users’ equipment (UEs) using inter-slice and intra-slice schedulers, respectively. This thesis also proposes investigating a slice prioritization structure to ensure the intent of more important slices when the available radio resources are insufficient to guarantee all slice’s intents. This thesis proposal presents results obtained using an intent-based RRS with RL for a fixed number of slices and also for multiple network scenarios, aiming to demonstrate the importance of intentbased RRS design for scenarios with RAN slicing. The proposed method outperformed the baselines in fixed and multiple network scenarios, protecting high-priority slices and minimizing the total number of violations.Item Acesso aberto (Open Access) Metodologia de Avaliação do desempenho energético da integração de carros elétricos à edificações(Universidade Federal do Pará, 2024-09-06) SOUZA, Ana Carolina Dias Barreto de; CARVALHO, Carminda Célia Moura de Moura; http://lattes.cnpq.br/1778332169942633; TOSTES, Maria Emília de Lima; http://lattes.cnpq.br/4197618044519148Energy diagnosis methodologies have been incorporating energy consumption and energy generation systems into their analysis, making it possible to classify energy-self-sufficient buildings as Near Zero Energy Buildings (NZEB) or Positive Energy Buildings (PEB). In electric mobility, the increased use of electric vehicles (EVs) brings challenges and opportunities in electricity consumption, management and efficiency. The impact of this robust and growing load when integrated into new and existing buildings is not yet considered in performance assessments. Consequently, the methodologies for obtaining certifications and labels do not consider the load of this system as an individual end-use. For buildings with energy efficiency (EE) and self-sufficiency labels, introducing EVs can result in the rating being downgraded due to increased energy consumption. Therefore, analyzing the impact of integrating EVs into buildings aims to support the formulation or revision of energy diagnosis methodologies that include EV charging systems integrated into buildings. This thesis evaluates the influence of EV charging in buildings with the NZEB/PEB label from the Brazilian Building Labeling Program (PBE Edifica). Using on-site surveys, computer modelling and thermo energetic analysis with software such as OpenStudio and EnergyPlus, an energy rating was carried out on a building in Belém, State of Pará, Brazil. Subsequently, energy flow simulations using probabilistic models with the Monte Carlo method were run in OpenDSS software to examine the impact of integrating EVs without (scenario 01) and with (scenario 02) the implementation of demand-side management techniques. Analysis using the labelling methodology showed that the building has an EE level C rating and NZEB self-sufficiency. Scenario 01 generated a 69.28% increase in energy consumption, reducing the EE level to D and resulting in the loss of the NZEB class. Scenario 02 increased consumption by 40.50%, a lower percentage than scenario 01 and guaranteed the return of the NZEB class lost in scenario 1, but did not return the EE level to class C. The results highlight the need for immediate and comprehensive energy management strategies. However, these strategies are not sufficient if other consumption restrictions or EE measures are not applied to other systems in the building. To this end, EE measures were proposed and evaluated in the air conditioning and lighting systems. Subsequently, an equation was drawn up to indicate the maximum level of energy X consumption that could be increased without compromising the building's energy performance and NZEB rating. Finally, OpenDSS software was used to simulate the increased availability of EV charging after the retrofit. With the proposed retrofit, the building improved its EE ratings by three levels, and the NZEB rating percentage increased by 33.28%. These measures also increased the EV charging load by 20% while maintaining the maximum EE level and NZEB rating.Item Acesso aberto (Open Access) Projeto de controle robusto de ordem fracionária para sistemas com incerteza paramétrica(Universidade Federal do Pará, 2024-10-21) GOMES, Marcus Ciro Martins; AYRES JÚNIOR, Florindo Antonio de Carvalho; http://lattes.cnpq.br/1919442364965261; COSTA JÚNIOR, Carlos Tavares da; http://lattes.cnpq.br/6328549183075122This research introduces a novel methodology that integrates fractional-order control theory with robust control techniques to address parametric uncertainty, aimed at enhancing the performance of linear time-invariant uncertain systems with integer or fractional orders, referred to as Fractional-Order Robust Control (FORC). Unlike traditional methods, this proposed approach offers a new formulation of inequalitiesbased design, broadening the scope for discovering improved solutions through linear programming optimization. Consequently, fractional-order controllers are tailored to ensure desired transient and steady-state performance in closed-loop systems. In order to facilitate the digital implementation of the designed controller, the impulse response invariant discretization of fractional-order differentiators (IRID-FOD) is used to approximate fractional-order controllers to integer-order transfer functions. Additionally, the Hankel reduction order method is applied, making the controllers suitable for hardware deployment. Experimental tests conducted on a thermal system, along with assessment results based on time-domain responses and robustness analysis supported by performance indices and set value analysis, demonstrate the enhanced and robust performance of the proposed FORC methodology compared to classical robust control under parametric uncertainty.Item Acesso aberto (Open Access) Avaliação probabilística do impacto da recarga rápida de veículos elétricos nos sistemas de distribuição de energia elétrica(Universidade Federal do Pará, 2024-11-13) HERNÁNDEZ GÓMEZ, Oscar Maurício; VIEIRA, João Paulo Abreu; http://lattes.cnpq.br/8188999223769913The mass adoption of electric vehicles (EVs) is transforming the automotive sector, driven by environmental concerns and technological advancements. Governments and companies are investing in the expansion of charging networks, focusing on fast charging to meet the growing demand. Developing a robust infrastructure of charging stations is essential to eliminate “range anxiety” and encourage the transition to EVs. Fast charging is crucial for the success of vehicle electrification. It allows batteries to be charged much more quickly than conventional charging, increasing convenience for users and improving the overall user experience. As more fast-charging stations are installed, consumer confidence in EVs grows, paving the way for a more sustainable future. With a well-distributed fast-charging network, EVs become a practical alternative to fossil fuel-powered vehicles, accelerating the transition to greener mobility. However, fast charging of EVs can cause technical impacts on medium voltage networks. The high current demand can result in voltage drops, especially in areas with weaker distribution infrastructure. Transformers can be overloaded, reducing their lifespan and increasing the risk of failures. Excessive heating of conductors due to high current can also cause losses and damage cables. These challenges highlight the need for proper planning and investments in electrical infrastructure to support the increase in fast charging. A probabilistic analysis of the impact of fast charging on medium voltage networks is crucial. Energy demand varies throughout the year due to seasonal factors, such as the use of air conditioning in summer and heaters in winter. Fast charging adds a considerable load to the network, which can coincide with these demand peaks, exacerbating management challenges. The installation of multiple charging points can cause voltage fluctuations and overloads. Probabilistic analysis helps predict these impacts and develop mitigation strategies by simulating charging scenarios and user behaviors. This allows for more precise infrastructure planning, including network reinforcements and improvements to ensure supply reliability. This thesis proposes a probabilistic methodology to evaluate the impact of fast charging of electric vehicles on medium voltage distribution networks, considering voltage drops, network element loading, and regulator tap changes. Using the Power Factory software by DIgSILENT©, a real feeder in Brazil is simulated, analyzing different case studies. Three fast charging stations (FCS) are connected, each with six charging points of 100 kVA, totaling 600 kVA per EP. The charging profile of the EPs is programmed with stochastic variables. Finally, a Volt/Var control strategy is presented to mitigate the impact on voltage drops and regulator tap changes, allowing reactive power injection without the need for communication between charging points.Item Acesso aberto (Open Access) Identificação de sistemas multiforças a partir de dados de vibração e técnicas de aprendizado de máquinas(Universidade Federal do Pará, 2024-11-07) PINHEIRO, Giovanni de Souza; NUNES, Marcus Vinícius Alves; http://lattes.cnpq.br/9533143193581447The emergence of defects in dynamic components tends to produce changes in the forces generated, which can be detected through alterations in the vibration response spectrum of the equipment. Understanding the forces acting on a structure is extremely important, especially in cases where measurement points are limited or inaccessible, as it allows for assessing, among other things, whether the component's lifespan is compromised by the current condition of the machine. In such cases, an inverse problem needs to be solved. Machine Learning techniques have been standing out as a powerful tool for prediction among the solutions developed for this type of problem, being increasingly applied to engineering problems. Therefore, this work aims to evaluate different machine learning models for the identification of a system, composed of a suspended plate with one or more applied forces, based on measured vibration data. In this regard, a computational model was generated and calibrated using vibration responses measured in the laboratory. A robust database was created using Response Surface Methodology together with the Design of Experiment (DOE) and then used to assess the ability of machine learning models to predict the location, excitation frequency, magnitude, and number of forces acting on the structure. Among the six machine learning models evaluated, k-NN was able to predict with an error of 0.013%, and random forests showed a maximum error of 0.2%. Finally, a database, containing a line of experimental data, was used to evaluate the k-NN and Random Forest models, obtaining a score of 0.96 and 0.93, respectively. The innovation of the study lies in the application of the proposed method for parameter identification in multiforce systems.Item Acesso aberto (Open Access) Orquestração de recursos multicamadas para arquiteturas de próxima geração(Universidade Federal do Pará, 2024-11-29) PAIXÃO, Ermínio Augusto Ramos da; CARDOSO, Diego Lisboa; http://lattes.cnpq.br/0507944343674734Due to the significant increase in data traffic and the large number of devices using Internet Protocol (IP), operators and researchers are seeking solutions to meet the growing demand. One of the most attractive solutions is Heterogeneous Cloud Radio Access Networks (H-CRAN), which has the capability to address current generation problems and bring various improvements, such as centralized processing and greater energy efficiency. The central challenge lies in the complexity of managing and optimizing these resources efficiently, especially in highdemand scenarios with dense device populations. However, the orchestration of resources such as radio, mapping between radio and BaseBand Unit (BBU), and load balancing in the BBU pool remains critically important. This thesis presents a framework aimed at reconfiguring the mobile network in areas affected by the variability of tidal effect traffic, ensuring high availability, energy savings, and improved data processing efficiency. The results obtained were compared with other approaches in the literature and demonstrated that the proposed framework optimizes the resources of the Peak Remote Radio Head (PRRH) and the BBU without compromising the user’s minimum QoS. The findings highlight a reduction of up to 9% in the number of active antennas over a 24-hour period and emphasize that the proposed solution consumes up to 14% less energy than the primary reference in this thesis.Item Acesso aberto (Open Access) Ciência de dados e aprendizado de máquina aplicados ao estudo de variáveis epidemiológica hanseníase na Amazônia(Universidade Federal do Pará, 2024-12-18) FALCÃO, Igor Wenner Silva; CARDOSO, Diego Lisboa; http://lattes.cnpq.br/0507944343674734; SERUFFO, Marcos César da Rocha; http://lattes.cnpq.br/3794198610723464Leprosy is a significant public health problem that largely affects low-income populations. Although the World Health Organization (WHO) establishes guidelines for diagnosis, prevention, and treatment, disease detection faces limitations, often resulting in late or inaccurate diagnoses and leading to serious neurological complications and multidrug-resistant cases. Therefore, early diagnosis is essential to reduce the burden of this disease. Machine learning has been widely used in several areas of science and industry, but especially in health, where it plays an essential role in the analysis and treatment of large volumes of data. In this sense, this thesis investigates the application of a model based on Data Science and Machine Learning to act in the specification of the clinical profile of possible leprosy cases in the Amazon Region and, thus, to be able to act preventively in the early diagnosis and treatment of patients under medical followup. The work takes into account clinical data of patients from a non-public dataset, collected between 2015 and 2020 in the North region of Brazil. Therefore, this thesis proposes a learning model to identify groups clinically affected by the disease using Clustering and Random Forest techniques. In the results obtained, the proposed model demonstrated efficiency in evaluating the probability of individuals being ill, achieving an accuracy of 90.39% in the performance evaluation and identifying a probability of 83.46% of an individual being ill, considering a set of epidemiological and non-generic variables. This approach offers a promising vision for the future of health, allowing the formulation of effective strategies for the early identification of possible cases.Item Acesso aberto (Open Access) Metodologias de controle de tensão com justiça de corte da geração fotovoltaica em redes de distribuição de baixa tensão(Universidade Federal do Pará, 2024-01-31) LOPES, Andrey CostaThe growing concern about climate change and global warming have motivated the current Energy Transition, which concerns the shift from fossil fuels to renewables energy sources (RES) in an effort to reduce CO2 emissions. This energy transition has driven the electrification of the economy, fostering significant growth in RES, particularly in photovoltaic solar energy. In this context, the decentralization of the electric sector has enabled the direct integration of these sources into Low Voltage Distribution Networks (LVDNs). However, the massive integration of Micro Photovoltaic Solar Generation (μPVSG) into these networks has caused reverse power flow, resulting in technical challenges such as overvoltage and thermal overload in their assets. Solutions, such as Volt-Watt Control (VWC) in Photovoltaic Inverters (PVIs), have proven effective in addressing voltagerelated issues. However, this control has led to an unfair distribution of active power among the PVIs during VWC operation, penalizing consumers located further from the distribution transformer. Additionally, stability issues related to the convergence in the dynamics of VWC, due to the slope of the Volt-Watt curve, have been considered in various studies. Therefore, this study presents a new methodology for adjusting Volt-Watt curves, ensuring the stability of VWC and simultaneously ensuring a fair power cut among PVIs. This approach is applied in two voltage control architectures, decentralized and centralized, respectively. In the first methodology, a linearized model of the network is used for Volt-Watt curve adjustment, employing local measurements at the connection points of the respective PVIs. In the second methodology, a voltage sensitivity matrix is used for the linearized model of the network when applying the Volt-Watt curve adjustment, where VWC parameters are coordinated in real-time, assisted by local measurements in the respective PVIs. The studies were conducted on a set of LVDNs and evaluated for effectiveness and fairness of power cuts quantitatively, using the Jain’s Fairness Index (JFI) as a metric. The results confirmed the effectiveness of the proposed control in mitigating voltage problems, acting fairly by equally exporting surplus energy to the grid, while ensuring controller stability. Additionally, penalties arising from the local dependence of PVIs in power cuts were eliminated compared to conventional VWC strategies.Item Acesso aberto (Open Access) Uma metodologia temporal para avaliação do desempenho de códigos concatenados em sistemas OFDM para transmissão de vídeo 4K-UHD(Universidade Federal do Pará, 2024-08-16) COSTA, Thiago de Araújo; CASTRO, Bruno Souza Lyra; http://lattes.cnpq.br/1897829604434609; BARROS, Fabrício José Brito; http://lattes.cnpq.br/9758585938727609The communication channel is a critical part of the process of information degradation. In the 4K ultra-resolution video transmission domain, the communication channel is a crucial part where information degradation occurs, inevitably leading to errors during reception. To enhance the transmission process in terms of fidelity, advanced technologies such as digital video broadcasting terrestrial (DVB-T) and its evolutionary successor, digital video broadcasting terrestrial second generation (DVB-T2), are utilized to mitigate the effects of data transmission errors. In the transition, a notable change is the replacement of the concatenated channel coding pairs. Within this scenario, this research presents an innovative methodology for the temporal analysis of 4K ultraresolution video quality under the influence of additive white Gaussian noise (AWGN) and Rayleigh channels. This analytical endeavor is facilitated through the application of concatenated coding schemes, specifically, the Bose-Chaudhuri-Hocquenghem concatenated low-density parity check (BCH-LDPC) and Reed-Solomon concatenated convolutional (RS-CONV) coders. A more comprehensive understanding of video quality can be attained by considering its temporal variations, a crucial aspect of the ongoing evolution of technological paradigms. In this study, the Structural Similarity Index (SSIM) serves as the main metric for quality assessment during simulations. Furthermore, the simulated Peak Signal-to-Noise Ratio (PSNR) values validate these findings, exhibiting consistent alignment with the SSIM-based evaluations. Additionally, the performance of the BCH-LDPC significantly outperforms that of RSCONV under the 64-QAM modulation scheme, yielding superior video quality levels that approximate or surpass those achieved by RS-CONV under QPSK (Quadrature Phase Shift Keying) modulation, leading to an increase in spectral efficiency. This enhancement is evidenced by SSIM gains exceeding 78% on average. The computation of average gains between distinct technologies in video quality analysis furnishes a robust and comprehensive evaluation framework, empowering stakeholders to make informed decisions within this domain.Item Acesso aberto (Open Access) Investigação experimental de estratégias de controle robusto aplicadas à melhoria de desempenho de um conversor de potência CC/CC do tipo buck com estrutura Single Inductor Multiple Output(Universidade Federal do Pará, 2024-08-05) MONTAÑO SAAVEDRA, Alvaro Christian; MEDEIROS, Renan Landau Paiva de; http://lattes.cnpq.br/8081923559538095; BARRA JUNIOR, Walter; http://lattes.cnpq.br/0492699174212608Recently, DC/DC power converters have gained wide attention, especially in industry,telecommunications, and the control of renewable energy sources. The increase in the use of this technology can be explained by the growing demand for high-quality DC voltage regulation in various applications. Additionally, recent advances in power electronics along with control engineering have accelerated the development of DC/DC power converters. Therefore, they looked to optimize these converters in several ways, such as improving conversion efficiency and reducing their weight and cost. In the proposed work, control strategies for voltage regulation in a single-inductor, dualoutput Buck DC-DC converter system (SIDO) are investigated. Based on a nominal multiple-input, multiple-output plant model and performance requirements, both a Linear Quadratic Regulator (LQR) and a Decoupled PI control strategy are designed to control the power converter system under parametric uncertainties such as variation of the voltage source, variations of constant power loads (CPLs) and variations of load resistances. A prototype of a single inductor dual output DC-DC Buck converter was developed for experimental testing. The results indicate that the proposed LQR strategy approach is reasonable and provides adequate performance improvements in SIDO converter controllers under conditions of varying voltage source and varying load resistances, offering robust performance and system stability; however, more research is needed to address variations in constant power loads and in the design of a PI controller for its application in this kind of system.Item Acesso aberto (Open Access) SmartLVEnergy: um framework para gestão energética inteligente e descentralizada de sistemas legados de baixa tensão(Universidade Federal do Pará, 2024-07-11) FERNANDES, Rubens de Andrade; GOMES, Raimundo Cláudio Souza; http://lattes.cnpq.br/4244097441063312; COSTA JÚNIOR, Carlos Tavares da; http://lattes.cnpq.br/6328549183075122Essential for technological and economic progress, electrical energy requires well-founded solutions and strategies for efficient and sustainable management. Existing consumer units, lacking modern technological resources, need gradual alternatives to optimize energy use, making the most of pre-established resources. In this context, retrofit offers an effective update for these infrastructures. Systematic models and strategies can standardize and ensure the replication of these solutions in different contexts through abstractions known as frameworks. However, there is a lack of frameworks to enable the implementation of systematic retrofit strategies for energy management, especially in the low-voltage energy sector. To fill this gap, this thesis presents the SmartLVEnergy framework, proposed to guide the design of innovative retrofit strategies to modernize legacy low-voltage installations with IoT, AIoT, and distributed computing solutions, optimizing energy management with distributed technological resources and advanced predictive capabilities. The experiments conducted in this thesis are presented in the format of aggregated scientific articles, which contributed to the conception of the SmartLVEnergy framework. As a result, it was possible to implement energy management tools in existing building and industrial scenarios in a systematic manner, based on the premises of the proposed framework. The main focus was the analysis and prediction of the energy demand of the installations and their respective circuits, allowing to anticipate and mitigate demand overrun events of the consumer units, following the guidelines of the Brazilian National Electric Energy Agency. The strategies conceived included the development, use, and integration of sensing, communication, and computing resources, distributed locally, in the cloud, and at the edge, according to the principles of the SmartLVEnergy framework, maximizing the use of existing resources according to the specific needs of each installation. The proposed framework is flexible and allows the integration, expandability, and interoperability of technological solutions across legacy systems, enabling operations according to the peculiarities and resources of each pre-existing context. This versatility confirms the relevance of this work as a robust and sustainable proposal to promote energy efficiency today, especially in legacy low-voltage systems.Item Acesso aberto (Open Access) Inibidor bidirecional de eventos de runaway no comutador de tap de reguladores de tensão em redes de distribuição reconfiguráveis com geração distribuída(Universidade Federal do Pará, 2024-06-25) SOUZA, Vanderson Carvalho de; VIEIRA, João Paulo Abreu; http://lattes.cnpq.br/8188999223769913Climate change has intensified over the years, especially as a result of the global energy model that is predominantly based on the use of fossil fuels. Thus, there is an urgent need to boost a low-carbon economy as a response to the climate crisis. In this context, renewable energy sources emerge as the main alternative to fossil fuels. However, the integration of these sources into distribution networks can cause voltage control problems resulting from bidirectional power flow in such networks. An important voltage control problem is the phenomenon known as tap changer runaway condition in step-voltage regulators (SVRs). Nowadays, the problem is further challenging in reconfigurable distribution networks with renewable energy sources connected to both the source-side and load-side of the SVR. This problem occurs when the SVR control cannot adequately distinguish the origin of the active power flow through the SVR and tries to control the voltage on the side of the network with the highest short circuit capacity (strong side), causing under or overvoltage on the side of the network with the lowest short circuit capacity. short circuit (weak side). Current solutions to mitigate the runaway problem are mainly based on three categories: 1) voltage control support by distributed generation (DG); 2) use of remote measurements/information; and 3) use of local measurements/information. However, considering practical aspects, only solutions in the third category are feasible. Even so, these solutions are restricted to application for inhibiting the runaway condition caused exclusively by reverse power flow. In this Thesis, an algorithm is proposed for robust local bidirectional on-line inhibition of the runaway condition based only on a test tap switching with robustness guarantees and without the need for switching of tap test coordinate in cascaded SVRs. The main contributions of the Thesis are the innovative application of the algorithm in robust local bidirectional on-line inhibition of the runaway condition in the tap switch and the introduction to industrial insights. The accuracy and robustness of the proposed algorithm are verified through time series power flow simulations carried out on two test networks, with noise and gross errors in measurements, using extensive Monte Carlo simulations. The uncoordinated operation of test tap switching in cascaded SVRs is examined through case studies on a long real rural distribution network. Finally, the effect of photovoltaic (PV) source variability on the performance of the proposed algorithm is evaluated. The results obtained confirmed the effectiveness of the proposed algorithm in bidirectional inhibition of the runaway conditionItem Acesso aberto (Open Access) Designing feasible deployment strategies for cell-free massive MIMO networks : assessing cost-effectiveness and reliability(Universidade Federal do Pará, 2024-06-14) FERNANDES, André Lucas Pinho; MONTI, Paolo; http://lattes.cnpq.br/4220330196422554; COSTA, João Crisóstomo Weyl Albuquerque; http://lattes.cnpq.br/9622051867672434Cell-free Massive Multiple-Input Multiple-Output (mMIMO) networks are a promising solution for the Sixth Generation of mobile systems (6G) and beyond. These networks utilize multiple distributed antennas to transmit and receive signals coherently, under an apparently non-cellular communication paradigm that eliminates the traditional concept of cells in mobile networks. This shift poses significant deployment challenges, as conventional tools designed for cellular systems are inadequate for planning and evaluating cell-free mMIMO architectures. In this sense, the literature has been developing models specific to cell-free mMIMO that deal with system coordination, fronthaul signaling, required computational complexities of processing procedures, segmented fronthaul, transitioning from cellular network deployments, and integration to Open Radio Access Network (O-RAN) technologies. These advancements are instrumental in transforming cell-free mMIMO from a theoretical system to a practical application. Despite this, further study is needed to integrate existing models and develop practical evaluation tools to assess the feasibility of cell-free mMIMO and its enablers. This thesis addresses these gaps by proposing new tools to evaluate the feasibility of cell-free mMIMO networks regarding reliability and costs. The first tool focuses on evaluating the reliability of cell-free mMIMO. It is used to improve the understanding of possible failure impacts and to develop effective protection schemes for the fronthaul network of cell-free mMIMO networks. Results for an indoor office implementation with an area of 100 m2 and a Transmission-Reception Point (TRP) spacing of 20 m, demonstrate that cell-free systems with segmented fronthaul, i.e., with serial fronthaul connections between TRPs, require protection strategies. It is shown that interconnecting serial chains and partially duplicating serial chains (40% redundancy) are effective protection schemes. Finally, in the considered indoor scenarios, interconnection appears to be the most feasible alternative when the number of serial chains is higher than three. The second tool assesses the Total Cost of Ownership (TCO) of cell-free mMIMO and its enablers, considering essential aspects, like user demands, fronthaul bandwidth limitations, and hardware processing capacities. The tool is used to evaluate the costs of two functional splits from the literature that are equivalent to distributed and centralized processing architectures for cell-free mMIMO networks. Results for an ultra-dense urban scenario covering an area of 0.25 km2 with up to 800 TRPs, reveal that centralized processing is more feasible for most user demands, hardware configurations of TRP, and cost considerations. Despite this, distributed processing may be more feasible in limited cases of low demand (up to 50 Mbps per user) and under massive cost reductions for expenses related to TRPs deployment.Item Acesso aberto (Open Access) Aplicação de redes neurais artificiais para predição de RSSI e SNR em ambiente de bosque amazônico(Universidade Federal do Pará, 2024-06-11) BARBOSA, Brenda Silvana de Souza; ARAÚJO, Jasmine Priscyla Leite de; http://lattes.cnpq.br/4001747699670004; https://orcid.org/0000-0003-3514-0401; BARROS, Fabrício José Brito; http://lattes.cnpq.br/9758585938727609The presence of green areas in urbanized cities is crucial to reduce the negative impacts of urbanization. However, these areas can influence the signal quality of IoT devices that use wireless communication, such as LoRa technology. Vegetation attenuates electromagnetic waves, interfering with data transmission between IoT devices, resulting in the need for signal propagation modeling that considers the effect of vegetation on its propagation. In this context, this research was conducted at the Federal University of Pará, using measurements in a wooded environment composed of the Pau-Mulato species, typical of the Amazon. Two propagation models based on machine learning, GRNN and MLPNN, were developed to consider the effect of Amazonian trees on propagation, analyzing different factors such as the height of the transmitter relative to the trunk, the beginning of the foliage, and the middle of the tree canopy, as well as the LoRa spreading factor (SF) 12 and the copolarization of the transmitter and receiver antennas. The best models were the machine learning ones, GRNN and MLPNN, which demonstrated greater accuracy, achieving root mean square error (RMSE) values of 3.86 dB and 3.8614 dB, and standard deviation (SD) of 3.8558 dB and 3.8564 dB, respectively. On the other hand, compared to classical models in the literature, the best-performing model was the Floating Intercept (FI) model, with RMSE and SD errors around 7.74 dB and 7.77 dB, respectively, while the FITU-R model had the highest RMSE and SD errors, around 26.40 dB and 9.65 dB, respectively, for all heights and polarizations. Furthermore, the importance of this study lies in its potential to boost wireless communications in wooded environments, as it was observed that even at short distances at heights of 12 m and 18 m, the SNR (Signal-to-Noise Ratio) had lower values due to the influence of the foliage, but it was still possible to send and receive data. Finally, it was shown that vertical polarization achieved the best results for the Amazon forest environment.Item Acesso aberto (Open Access) Scalable AP selection strategies for user-centric cell-free massive MIMO networks(Universidade Federal do Pará, 2024-06-06) FREITAS, Marx Miguel Miranda de; COSTA, Daniel Benevides da; http://lattes.cnpq.br/0644241968356756; https://orcid.org/0000-0002-5439-7475; VALCARENGHI, Luca; COSTA, João Crisóstomo Weyl Albuquerque; http://lattes.cnpq.br/9622051867672434User-centric (UC) cell-free (CF) massive multiple-input multiple-output (MIMO) systems are promising technologies for beyond 5G (B5G) networks. In these systems, the user equipment (UE) is associated with a subset of access points (APs) distributed into the coverage area, leading to improvements in macro-diversity and spectral efficiency (SE) compared to conventional cell-based systems. Despite the benefits, challenges such as scalable AP selection strategies, computational complexity (CC), and inter-central processing unit (CPU) coordination may still exist in these systems. In this regard, this thesis proposes a novel and general AP selection framework that affords scalability for UC systems, enabling more efficient use of the network resources, such as transmission power and reduced processing demands. The solution is based on a matched-decision among the most suitable connections for APs and UEs. Moreover, three strategies to fine-tune the AP clusters of UEs are proposed, aiming to reduce the number of APs connected to each UE without compromising the SE. Simulation results reveal that the matched-decision framework improves up to 163% the SE of the 95% likely UEs compared with baseline schemes. A heuristic approach that reduces the effects of inter-CPU coordination is also proposed. It decreases the number of inter-coordinated UEs (i.e., UEs connected to multiple CPUs) on each CPU to reduce signaling demands on backhaul links. Numerical results indicate that the proposed method mitigates inter-CPU coordination while yielding slight degradation in SE and improving energy efficiency (EE). Finally, this thesis investigates the performance of UC systems with limited processing capacity. Specifically, it is assumed that the CC of performing channel estimation and precoding signals does not increase with the number of APs. Thus, the UE can only be associated with a finite number of APs. Furthermore, a method is proposed for adjusting the AP clusters according to the network implementation, i.e., centralized or distributed. The results show that UC systems can keep the SE under minor degradation even if the CC up to 96%. Besides, the proposed method for adjusting the AP cluster leads to further reductions in CC.Item Acesso aberto (Open Access) Um método baseado em cruzamentos por zero para localização de faltas de alta impedância em redes aéreas de distribuição(Universidade Federal do Pará, 2024-06-05) HUAQUISACA PAYE, Juan Carlos; CARDOSO JUNIOR, Ghendy; http://lattes.cnpq.br/6284386218725402; https://orcid.org/0000-0002-1423-6968; VIEIRA, João Paulo Abreu; http://lattes.cnpq.br/8188999223769913The location of High-Impedance Faults (HIFs) is an increasingly relevant reliability issue in the electric power distribution industry. The development of practical and accurate single-terminal fault locating methods is vital for reducing the time and cost of restoring long-duration interruptions. However, the need to estimate both the parameters of the fault model and the fault current signal can compromise the accuracy and practicality of existing HIF location methods. This is due to the larger number of parameters that need to be estimated when a HIF model is included in the formulation, as well as the assumption that load currents at the network bars are constant during a pre- and post-HIF interval. In other words, the use of the fault model and waveform implies that the location method depends on the random characteristics and magnitudes of the fault current, which are determined by environmental, technical conditions, and the type of surface where the HIF occurs, including even the way the contact between the surface and the conductor occurs. This thesis proposes a fault-model-free iterative method based on zero-crossings of signals to locate HIFs in overhead distribution networks. Two insights on voltage signal relationships are provided to eliminate the need to estimate fault model parameters and the fault current signal in the HIF location process. The first insight is based on zero-crossings of the calculated voltage drop per unit length signal to estimate two parameters of the voltage signal at the fault point. The other insight is based on the zero-crossing of the voltage signal at the fault point, where the two parameters were previously estimated, to calculate the fault distance from the k-th node. Simulation results on a modified IEEE 34-node test feeder validate the high accuracy and robustness of the proposed method, considering the effect of various factors on the estimation of the HIF distance. Additionally, the convergence performance of the proposed method is evaluatedItem Acesso aberto (Open Access) Detecção de danos em superfícies geotécnicas com redes neurais convolucionais de baixa complexidade(Universidade Federal do Pará, 2024-05-29) ARAÚJO, Thabatta Moreira Alves de; FRANCÊS, Carlos Renato Lisboa; ttp://lattes.cnpq.br/7458287841862567Most natural disasters result from geodynamic events, such as landslides and collapse of geotechnical structures. These failures are catastrophic that directly impact the environment and cause financial and human losses. Visual inspection is the main method for detecting surface flaws in geotechnical structures. However, visits to the site can be risky due to the possibility of soil’s instability. Furthermore, the terrain design, hostile environment and remote installation conditions make access to these structures impractical. When a quick and safe assessment is necessary, computer vision analysis becomes a potential alternative. However, studies on computer vision techniques still need to be explored in this field due to the particularities of geotechnical engineering, such as limited, redundant and scarce public data sets. In this context, this thesis presents a redes neurais convolucionais, do inglês Convolutional Neural Network (CNN) approach for identifying defects on the surface of geotechnical structures to reduce dependence on human-led on-site inspections. To this end, images of surface failure indicators were collected on slopes on the banks of a Brazilian highway, with the help of UAVs and mobile devices. Next, low-complexity CNN architectures were explored to build a binary classifier capable of detecting flaws apparent to the naked human eye in images. The architecture composed of three convolutional layers, each with 32 filters, followed by two fully connected layers, each composed of 128 neurons and output with one neuron, showed an accuracy of 94.26%. The performance evaluation of the model with the test set obtained AUC metrics of 0.99, confusion matrix, and a AUPRC curve that indicates robust performance of the classifier in detecting damage, while maintaining a low computational complexity, making it suitable for applications field practices. The contributions of the thesis include the provision of an image database, the obtaining of a classification model suitable for scarce data and limited computational resources, and the exploration of strategies for remote inspection and detection of signs of failure in geotechnical structures.Item Acesso aberto (Open Access) Previsão de séries temporais no sistema elétrico brasileiro utilizando preditores baseados em aprendizado de máquina: uma análise empírica(Universidade Federal do Pará, 2024-04-05) CONTE, Thiago Nicolau Magalhães de Souza; OLIVEIRA, Roberto Célio Limão de; http://lattes.cnpq.br/4497607460894318; https://orcid.org/0000-0002-6640-3182The overview of electric energy in Brazil is influenced by a variety of complex factors and nonlinear relationships, making forecasting challenging. With the increasing demand for energy and growing environmental concerns, it is crucial to seek solutions based on clean and renewable energy practices, aiming to make the energy market more sustainable. These practices aim to reduce waste and optimize the efficiency of processes involved in the operation of electricity distribution and generation technologies. A promising approach to enable sustainable energy is the application of forecasting techniques for various variables in the energy market. This thesis proposes an empirical analysis of the use of regressors to make predictions in the databases of the Price of Settlement Differences (PLD) in the Brazilian market and wind speed in wind turbines in Northeast Brazil, through principal component analysis. We aim to provide significant information about machine learning techniques that can be employed as effective tools for time series prediction in the electric sector. The results obtained may encourage the implementation of these techniques to extract knowledge about the behavior of the Brazilian energy system. This is particularly relevant, given that energy prices often exhibit seasonality, high volatility, and peaks, and wind power generation is widely influenced by weather conditions. To model the prediction of these two time series, we use the database on the Price of Settlement Differences (PLD), focusing especially on the average energy price of the Brazilian National System. The most relevant variables are related to hydrological conditions, electrical load, and fuel prices for thermal units. For collecting variables related to wind energy, two distinct locations in the Northeast region of Brazil were considered: Macau and Petrolina. For the prediction study, we use a Multilayer Perceptron Neural Network (MLP), a Long Short Term Memory (LSTM), Auto-Regressive Integrated Moving Average (ARIMA), and Support Vector Machine (SVM) to determine baseline results in prediction. To enhance the results of these regressors, we employ two different prediction approaches. One approach involves combining deep artificial neural network techniques based on the Canonical Genetic Algorithm (AG) meta-heuristic to adjust the hyperparameters of MLP and LSTM regressors. The second strategy focuses on machine committees, which include MLP, decision tree, linear regression, and SVM in one committee, and MLP, LSTM, SVM, and ARIMA in another. These approaches consider two types of voting, voting average (VO) and voting weighted average (VOWA), to assess the impact on the performance of the machine committee.Item Acesso aberto (Open Access) Classification and characterization methods of non-tchnical losses on smart grid scenarios(Universidade Federal do Pará, 2024-03-28) BASTOS, Lucas de Lima; ROSÁRIO, Denis Lima do; http://lattes.cnpq.br/8273198217435163; https://orcid.org/0000-0003-1119-2450; CERQUEIRA, Eduardo Coelho; ttp://lattes.cnpq.br/1028151705135221Nowadays, grid resilience as a feature has become non-negotiable, significantly when power interruptions can impact the economy and society. Smart Grids (SGs) widespread popularity enables an immense amount of fine-grained e lectricity consumption data to be collected. However, risks can still exist in the Smart Grid (SG), since SG systems exchange valuable data, the distribution system loses substantial electrical energy. We divide this loss into two categories: technical and non-technical loss. A substantial amount of electrical energy is lost throughout the distribution system, and these losses are divided into two types: technical and non-technical. Non-technical losses (NTL) are any electrical energy consumed that is not invoiced. They may occur due to illegal connections, fraudulent activities, issues with energy meters such as delay in the installation or reading errors, contaminated, defective, or non-adapted measuring equipment, very low valid consumption estimates, faulty connections, and disregarded customers. Non-technical losses are the primary cause of revenue loss in the SG. Annually, electrical utilities incur billions in losses due to non-technical reasons. This thesis presents two detection methods of NTL: classification a nd c haracterization. We c reate a n ensemble predictor-based time series classifier t o c lassify N TL d etection. T his p redictor u ses the user’s energy consumption as a data input for classification, f rom s plitting t he d ata to executing the classifier. A lso, i t a ssumes t he t emporal a spects o f e nergy consumption data during the pre-processing, training, testing, and validation stages. The classification method has the advantage of classifying heterogeneous features in data. The characterization method proposes a study based on Information Theory Quantifiers (ITQ) to mitigate this challenge. First, we use a sliding window to convert the user’s energy consumption time series into a Bandt-Pompe (BP) probability distribution function. Then, we extract the used ITQ. Finally, we apply each metric to the Probability Density Function (PDF) and map the layers to characterize their behavior. The characterization method is advantageous to be used when we have big data. Overall, our best results have been recorded in the fraud detection-based time series classifiers (TSC) model, improving the empirical performance metrics by 10% or more over the other developed models. Our results show that users with normal and abnormal energy consumption can be distinguished using only Information Theory Quantifiers by considering the range of values for each metric.Item Acesso aberto (Open Access) Development of machine learning-based frameworks to predict permeability of peptides through cell membrane and blood-brain barrier(Universidade Federal do Pará, 2024-03-27) OLIVEIRA, Ewerton Cristhian Lima de; LIMA, Anderson Henrique Lima e; http://lattes.cnpq.br/2589872959709848; https://orcid.org/0000-0002-8451-9912; SALES JUNIOR, Claudomiro de Souza de; http://lattes.cnpq.br/4742268936279649Peptides comprise a versatile class of biomolecules with diverse physicochemical and structural properties, in addition to numerous pharmacological and biotechnological applications. Some groups of peptides can cross biological membranes, such as the cell membrane and the human blood-brain barrier. Researchers have explored this property over the years as an alternative to developing more powerful drugs, given that some peptides can also be drug carriers. Although some machine learning-based tools have been developed to predict cell-penetrating peptides (CPPs) and blood-brain barrier penetrating peptides (B3PPs), some points have not yet been explored within this theme. These points encompass the use of dimensionality reduction (DR) techniques in the preprocessing stage, molecular descriptors related to drug bioavailability, and data structures that encode peptides with chemical modifications. Therefore, the primary purpose of this thesis is to develop and test two frameworks based on DR, the first one to predict CPPs and the second to predict B3PPs, also evaluating the molecular descriptors and data structure of interest. The results of this thesis show that for the prediction of penetration in the cell membrane, the proposed framework reached 92% accuracy in the best performance in an independent test, outperforming other tools created for the same purpose, besides evidencing the contribution between the junction of molecular descriptors based on amino acid sequence and those related to bioavailability and cited in Lipinski’s rule of five. Furthermore, the prediction of B3PPs by the proposed framework reveals that the best model using structural, electric, and bioavailability-associated molecular descriptors achieved average accuracy values exceeding 93% in the 10-fold cross-validation and between 75% and 90% accuracy in the independent test for all simulations, outperforming other machine learning (ML) tools developed to predict B3PPs. These results show that the proposed frameworks can be used as an additional tool in predicting the penetration of peptides in these two biomembranes and are available as free-touse web servers.