O crescente interesse por sistemas de transmissão do tipo VSC-based HVDC (VSCHVDC) como alternativa aos atuais tem levado à necessidade de estudos sobre o principal desafio associado à sua implantação, que é o desenvolvimento de um esquema de proteção que seja rápido e que identifique a zona de proteção na qual uma eventual falta tenha ocorrido, respeitando-se assim o princípio de seletividade. Para estudo de um sistema de proteção que atenda a estes requisitos, foi utilizada neste trabalho uma rede com quatro terminais e transmissão High-Voltage Direct Current (HVDC), implementada com conversores modulares multiníveis (Modular Multilevel Converter (MMC)) e simulada em situações de contingência. Grandezas medidas nos elos de Corrente Contínua (CC) ligados a um dos terminais foram utilizadas para identificar a ocorrência de uma contingência na rede e então para decidir se trata-se de uma falta interna à zona de proteção considerada na análise. Para classificar as faltas em interna ou externa, observaram-se os espectros dos sinais de tensão CC em um terminal, calculados sobre uma janela do sinal através de Fast Fourier Transform (FFT), para várias situações de falta. Destes espectros, notou-se que uma característica varia de acordo com a distância da falta até o terminal, com o fato de a falta ser interna ou externa e com a resistência de falta. Através desta característica, denominada aqui \"frequência de centro de massa\", foi possível distinguir as faltas internas com resistências de falta baixas do restante dos eventos testados. Considerando-se apenas este parâmetro, não foi possível classificar corretamente faltas internas com resistência de falta muito alta. / The growing interest in VSC-based HVDC (VSC-HVDC) transmission systems as an alternative to the current ones has required studies about the the main challenge associated with its application, which is the development of a protection scheme that is able to both fast and efficiently determine the protection zone where a fault may have happened so as to achieve selectivity. In order to study a protection system that complies with those requirements, a High-Voltage Direct Current (HVDC) transmission system with four terminals implemented with Modular Multilevel Converter (MMC) converters was simulated under contingency condition. Electrical quantities measured in one of the terminal\'s Direct Current (DC) links were used to determine the occurrence of a fault event in the system, as well as to decide whether that event belonged to the protection zone considered in the analysis. Spectra obtained via Fast Fourier Transform (FFT) over a window of the DC voltage measured in one of the terminals were observed in order to discriminate between internal and external fault events. From these observations, it was noticed that a certain characteristic, named here \"center of mass frequency\", varies according to the distance between the fault location and the terminal, fault resistance, as well as whether that the fault is either internal or external to the analysed protection zone. By using this characteristic, it was possible to tell internal faults with low resistance fault from other events. Considering this parameter only, it was not possible to correctly classify internal faults with high fault resistance.

O crescente interesse por sistemas de transmissão do tipo VSC-based HVDC (VSCHVDC) como alternativa aos atuais tem levado à necessidade de estudos sobre o principal desafio associado à sua implantação, que é o desenvolvimento de um esquema de proteção que seja rápido e que identifique a zona de proteção na qual uma eventual falta tenha ocorrido, respeitando-se assim o princípio de seletividade. Para estudo de um sistema de proteção que atenda a estes requisitos, foi utilizada neste trabalho uma rede com quatro terminais e transmissão High-Voltage Direct Current (HVDC), implementada com conversores modulares multiníveis (Modular Multilevel Converter (MMC)) e simulada em situações de contingência. Grandezas medidas nos elos de Corrente Contínua (CC) ligados a um dos terminais foram utilizadas para identificar a ocorrência de uma contingência na rede e então para decidir se trata-se de uma falta interna à zona de proteção considerada na análise. Para classificar as faltas em interna ou externa, observaram-se os espectros dos sinais de tensão CC em um terminal, calculados sobre uma janela do sinal através de Fast Fourier Transform (FFT), para várias situações de falta. Destes espectros, notou-se que uma característica varia de acordo com a distância da falta até o terminal, com o fato de a falta ser interna ou externa e com a resistência de falta. Através desta característica, denominada aqui \"frequência de centro de massa\", foi possível distinguir as faltas internas com resistências de falta baixas do restante dos eventos testados. Considerando-se apenas este parâmetro, não foi possível classificar corretamente faltas internas com resistência de falta muito alta. / The growing interest in VSC-based HVDC (VSC-HVDC) transmission systems as an alternative to the current ones has required studies about the the main challenge associated with its application, which is the development of a protection scheme that is able to both fast and efficiently determine the protection zone where a fault may have happened so as to achieve selectivity. In order to study a protection system that complies with those requirements, a High-Voltage Direct Current (HVDC) transmission system with four terminals implemented with Modular Multilevel Converter (MMC) converters was simulated under contingency condition. Electrical quantities measured in one of the terminal\'s Direct Current (DC) links were used to determine the occurrence of a fault event in the system, as well as to decide whether that event belonged to the protection zone considered in the analysis. Spectra obtained via Fast Fourier Transform (FFT) over a window of the DC voltage measured in one of the terminals were observed in order to discriminate between internal and external fault events. From these observations, it was noticed that a certain characteristic, named here \"center of mass frequency\", varies according to the distance between the fault location and the terminal, fault resistance, as well as whether that the fault is either internal or external to the analysed protection zone. By using this characteristic, it was possible to tell internal faults with low resistance fault from other events. Considering this parameter only, it was not possible to correctly classify internal faults with high fault resistance.

<p>  </p> <p>This dissertation is composed of three main articles, which are detailed below.</p> <p>First article: The proliferation of prosumers generates the opportunity to have a more decentralized and open energy market. Given this opportunity, the Peer-to-Peer (P2P) trading energy paradigm appears, where consumers and prosumers can exchange energy without an intermediary. Because P2P energy trading plays a fundamental role in the proliferation of renewable energies and the system flexibility for a low-carbon energy transition, this article provides a review of the P2P energy trading that is necessary to understand the current approaches, challenges, and future research that should be conducted in this area. As a result, areas for consideration were identified and grouped into the following six topics: (1) trading platform, (2) blockchain, (3) game theory, (4) simulation, (5) optimization, and (6) algorithms. The study identified several challenges that may give way to future research, such as integrating generation, transmission, and distribution into studies, large-scale studies, and modeling consumer and prosumer complex behavior. Given that P2P energy trading is a relatively new topic, there is still much work to be done to implement the real-world model successfully.</p> <p>Second article: Along with the <a href="https://www.sciencedirect.com/topics/engineering/exponential-growth" target="_blank">exponential growth</a> of distributed energy sources in the last decade, net-metering programs have expanded to encourage investment in renewable energy. However, several countries and some states in the United States are ending these programs. Therefore, it is needed to explore alternatives to net-metering programs to continue encouraging the adoption of renewable energies. In this paper, Peer-to-Peer (P2P) models are studied as viable options to net-metering. In particular, the evaluation and comparison of the net-metering model to two P2P models is proposed. The first P2P model used the power grid for electricity exchange, and the other used electric vehicles. Simulations of a 50 household <a href="https://www.sciencedirect.com/topics/engineering/micro-grids" target="_blank">microgrid</a> with access to electric vehicles and <a href="https://www.sciencedirect.com/topics/engineering/photovoltaics" target="_blank">photovoltaic</a> generation were run to achieve this objective. Technical and economic indexes were established to measure the performance of the three models. The results indicated that the P2P model using the grid shows the best performance, followed by the P2P model using electric vehicles. </p> <p>Third article: Solar generation has increased rapidly in recent years worldwide, and it is projected to continue to grow exponentially. A problem exists in that the increase in solar energy generation will increase the probability of grid disturbances. This article focused on analyzing the grid disturbances caused by the massive integration to the transmission line utility-scale solar loaded to the balancing authority high voltage transmission lines in four regions of the United States electrical system: (1) California, (2) Southwest, (3) New England, and (4) New York. Statistical analysis of equality of means was carried out to detect changes in the energy balance and peak power. Results show that when comparing the difference between hourly net generation and demand, energy imbalance occurs in the regions with the highest solar generation: California and Southwest. No significant difference was found in any of the four regions in relation to the energy peaks. The results imply that regions with greater utility level solar energy adoption must conduct greater energy exchanges with other regions to reduce potential disturbances to the grid. It is essential to bear in mind that as the installed solar generation capacity increases, the potential energy imbalances created in the grid increase. </p>

Orientador: Maria Cristina Dias Tavares / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-25T04:28:30Z (GMT). No. of bitstreams: 1 ZevallosAlcahuaman_MiltonElvis_D.pdf: 2479285 bytes, checksum: 4c1c5f42e8da1a073caee7cb88308479 (MD5) Previous issue date: 2013 / Resumo: Uma solução bastante utilizada para reduzir a corrente de arco secundário nas linhas aéreas de transmissão em extra-alta-tensão após a ocorrência de faltas monofásicas não permanentes é a inserção de um reator de neutro no arranjo dos reatores dos bancos de compensação reativa em derivação da linha. Muitas vezes estes reatores de neutro não são otimizados, de modo que reatores com valores típicos são especificados nos projetos de linhas de transmissão aéreas, baseados simplesmente no seu nível de tensão, mas sem considerar as especificidades das linhas, como as suas características físicas ou o seu comprimento. A especificação inadequada do reator de neutro pode resultar em correntes de arco secundário elevadas, o que, em última análise, pode reduzir a taxa de sucesso da manobra de religamento monopolar. Além disto, o custo do sistema de compensação pode ser maior do que o necessário. A tese apresenta uma extensa análise de sensibilidade dos principais parâmetros que influenciam no dimensionamento de um tronco de transmissão longo, especificamente do nível de compensação mais adequado e do valor ótimo do reator de neutro. Destaca-se na pesquisa o desenvolvimento de uma representação dos componentes de um sistema de transmissão através de quadripolos trifásicos com o intuito de obter a resposta sustentada durante o desequilíbrio. Através desta representação foi possível identificar a relação entre o reator de neutro ótimo e os parâmetros transversais da linha de transmissão, assim como a influência direta entre o nível de compensação do sistema com o valor do reator de neutro otimizado. Adicionalmente foram identificados comprimentos limites para linhas não compensadas de diferentes níveis de tensão de modo a ainda apresentar alta probabilidade de sucesso para a manobra monopolar. Simulações de faltas monofásicas foram realizadas utilizando dados provenientes de linhas típicas do sistema elétrico brasileiro / Abstract: A widely used solution to reduce secondary arc current in the overhead transmission lines in EHV after single-phase non permanents faults is the insertion of a neutral reactor in the arrangement forming part of the shunt reactive compensation of the line. Frequently in Brazilian electrical system, neutral reactors are not optimized, so that reactors with typical values are specified in the project of overhead transmission lines based on their voltage level, but without considering the specifics line parameters, as their physical characteristics or their length. The specification of unsuitable neutral reactor can result in high secondary arc currents, which eventually diminishes the success rate of single-phase automatic reclosure (SPAR) maneuver. Moreover, the cost of the compensation system may be larger than necessary. This paper presents an extensive sensitivity analysis of the main parameters that influence the project of a long transmission trunk, specifically the most appropriate level of compensation and the optimal value of the neutral reactor. It is noteworthy the development of a transmission system components representation using three-phase two-ports networks in order to obtain sustained response during unbalance. Through this analysis it was possible to identify the relationship between the shunt reactor and the optimized neutral parameter and the influence of the compensation level with the neutral reactor optimized value. Additionally the maximum lengths of non-compensated lines that would have high probability of success of SPAR maneuver were identified. Simulations of single-phase faults were performed using data from typical lines of the Brazilian electric system / Doutorado / Energia Eletrica / Doutor em Engenharia Elétrica

Orientadores: José Pissolato Filho, Pierre Kaufmann / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-23T01:41:59Z (GMT). No. of bitstreams: 1 Duro_MagdaAparecidaSalgueiro_D.pdf: 3861523 bytes, checksum: f5896319d2a3653a7903f977210f3ac1 (MD5) Previous issue date: 2013 / Resumo: A atual dependência da sociedade aos diversos sistemas tecnológicos em funcionamento na Terra traz uma constante preocupação pela sua vulnerabilidade a fatores menos conhecidos. As possíveis falhas no fornecimento de energia elétrica podem ocasionar transtornos de grande impacto às comunidades podendo acarretar perdas financeiras expressivas. Diversos fatores podem ocasionar falhas nas linhas de transmissão, entre eles, os fatores climáticos. Nesse contexto, são pouco conhecidas as influências qualitativas de distintos fatores, destacando-se os efeitos do denominado clima espacial. Há indicações que o clima espacial influencia a alta atmosfera terrestre, com repercussão no clima bem como em sistemas de engenharia em operação na Terra. As redes de transmissão de alta tensão representam um grande circuito elétrico pouco acima do solo sujeito a uma série de sobrecargas temporárias de vários tipos, algumas das quais podem levar a falhas. Algumas destas falhas podem estar relacionadas ao ambiente geofísico. Neste trabalho foi analisada uma base de dados sem precedentes de falhas em linhas de transmissão por um longo período (nove anos) numa grande malha localizada no Estado de São Paulo (sudeste do Brasil). O período estudado (1998-2006) compreende uma significativa fração do ciclo de atividade solar 23. A concessionária responsável (ISA.CTEEP) pela operação das linhas estudadas classifica as falhas em noventa e cinco tipos distintos de causas possíveis, sendo que a maior parte está relacionada às descargas atmosféricas. Este estudo está relacionado aos desligamentos devido às descargas atmosféricas, em duas redes de alta voltagem da companhia, de 138 kV e de 440 kV. No período estudado, estes desligamentos correspondem a 1.957 (42,80%) num total de 4.572 para a linha de 138 kV e de 170 (22,28%) sobre 763 para a linha de 440 kV. Equivale a menos de um a cada dez mil do número total de descargas atmosféricas ocorridas na mesma área, o que demonstra alta resiliência das redes de potência em relação aos raios. Durante a época das chuvas, há uma maior concentração de desligamentos. Para todo o período estudado houve redução de 67% e 77% no número de desligamentos devido às descargas atmosféricas, para as linhas de 138 kV e 440 kV, respectivamente, havendo uma boa correspondência com a redução do número de manchas solares. Nenhuma correlação foi encontrada em relação às atividades geomagnéticas caracterizadas pelo índice planetário (Kp) e pelas grandes tempestades (Dst) tanto em longos quanto em curtos prazos. Uma explicação sugerida associa a diminuição da atividade solar ao aumento da condutividade na atmosfera causada pelo maior fluxo de raios cósmicos. Consequentemente poderá ocorrer uma redução do limiar de voltagem necessário para produzir descargas atmosféricas para provocar desligamentos nas redes de alta voltagem. O circuito elétrico global descrito pelo acoplamento ionosfera-terra (a eletrosfera) exerce um papel importante para explicar a redução dos desligamentos. Com o aumento da condutividade atmosférica as descargas são menos potentes, resultando em menor número de desligamentos com o decréscimo do ciclo solar / Abstract: The current society dependence on the operating technological systems on the Earth brings a permanent concern for their vulnerability to not well known factors. Possible failures in electricity supply can cause inconveniences with large impact to the communities, which may cause significant financial losses. Different factors can cause failures in the transmission networks, including, climate factors. In this context, the influences of various qualitative factors, especially the effects of space weather are not well known. There are indications that space weather affects the upper atmosphere, with repercussions on the climate as well as in engineering systems in operation on Earth. High-voltage transmission networks represent large electrical circuits just above the ground which are subjected to a number of transient overcharges of various kinds, some of which may lead to failures. Some failures might be related to anomalies of the geophysical environment. In the present study we analyze a database consisting in a one unprecedented long series of transmission grid failures (nine years) on high-voltage networks located in São Paulo state (southeastern Brazil). The studied period (1998-2006) includes an important fraction of the solar activity cycle 23. The company responsible by the power grid operator (ISA.CTEEP) classifies causes in ninety-five distinct failure classes to explain the transmission grid shut downs. Most of the failures were attributed to atmospheric discharges. We have studied the failures attributed to atmospheric discharge, in the two power grids of the company, 138 kV and 440 kV. The failures attributed to atmospheric discharge correspond to 1.957 (42.80%) for a total 4.572 at 138 kV and to 170 (22.8%) out of 763 at 440 kV. They correspond to less than one ten thousandth of the actual number of atmospheric discharges recorded in the same area, demonstrating the grid's high resilience to breakdowns due to lightning. A clear concentration of failures in the region's thunderstorm season has been found. A significant 67% and 77% reduction in the number of failure rates has been found for the 138 and 440 kV grids, respectively, for the period studied, in good correspondence with the decay in the sunspot numbers. No obvious correlation was found between power failures and geomagnetic activity, represented by (Kp) the planetary index or major geomagnetic storms (Dst) in the period, either on short or on long time scales. One suggested explanation associates the decrease in solar activity to the increased conductivity in the atmosphere caused by increase cosmic ray flux. Consequently there may be a reduction in the threshold voltage necessary to produce discharge atmospheric to cause failures in high-voltage grids. The global electric circuit described by the ionosphere-ground coupling (the electrosphere), plays an important role in explaining the reduction the failures. With increase in conductivity atmospheric discharges are less potent, resulting in fewer failures with the decreasing solar cycle / Doutorado / Energia Eletrica / Doutor em Engenharia Elétrica

Orientadores: Alberto Luiz Francato, Carlos Alberto Mariotoni / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Civil, Arquitetura e Urbanismo / Made available in DSpace on 2018-08-18T13:32:01Z (GMT). No. of bitstreams: 1 Barbosa_CarlosDavidFranco_D.pdf: 7182306 bytes, checksum: 6a8fd84aba169e2fceffb93e27a4c5b5 (MD5) Previous issue date: 2011 / Resumo: Esta tese tem como objetivo estudar métodos de planejamento da expansão da transmissão de energia elétrica em mercados de eletricidade re-estruturados, tanto em um ambiente tradicional quanto com a adoção de inovações tecnológicas. No trabalho são estudados procedimentos de campo para as linhas de transmissão, considerando economia e confiabilidade, principalmente em emendas a compressão para cabos ACSR (Aluminum Core Steel Reinforced). O trabalho descreve os resultados de pesquisas efetuadas com fabricantes e concessionárias usuárias das emendas a compressão e ainda sugere alguns procedimentos e rotinas de inspeções para melhorar a confiabilidade do sistema. Os resultados dos testes com as emendas, realizados em laboratório, serviram para dar suporte para a analise de falhas, subsidiando a avaliação das causas de falhas mais frequentes no ambiente competitivo. Por fim foram apresentadas tecnologias emergentes com possibilidade de redução da taxa de falhas nos sistemas de transmissão de energia elétrica bem como as recomendações de boas praticas para o setor / Abstract: This thesis has as objective to study electricity transmission expansion planning In re-structured markets, both traditional as well as with technological innovations. In the work are studied transmission lines field procedures considering economy and reliability, mainly ACSR (Aluminum Core Steel Reinforced) compression splices. The work describes compression splices field surveys with suppliers and users concessionaries and also suggest some procedures and inspections routines to improve system reliability. The results of laboratory splice tests helped the failure analysis and frequently failure causes on competitive environment. Finally it is presented emerging technologies with failure reduction possibilities for the electricity transmission systems and good practices recommendations for sector / Doutorado / Recursos Hidricos, Energeticos e Ambientais / Doutor em Engenharia Civil

<p dir="ltr">The horizon for inclusion of data-driven algorithms in cyber-physical systems is rapidly expanding due to evermore availability of high-performance computing tools and the inception of novel mathematical models in the fields of deep learning and reinforcement learning. In this regard, energy systems are a suitable candidate for data-driven algorithms utilization due to rapid expansion of smart measuring tools and infrastructure. Accordingly, I decided to explore the capabilities of deep reinforcement learning in control, security, and restoration of smart energy systems to tackle well-known problems such as ensuring stability, adversarial attack avoidance, and the black start restoration. To achieve this goal, I employed various reinforcement learning techniques in different capacities to develop transfer learning modules based on a rule-based approach for online control of the power system, utilized reinforcement learning for procedural noise generation in adversarial attacks against contingency detection in a power system and exploited multiple reinforcement learning algorithms to fully restore an energy system in an optimal manner. Per the results of these endeavors, I managed to develop a rule-based transfer learning logic to control the power system under various disturbance types and intensities. Furthermore, I developed an optimal adversarial attack module using a reinforcement-learning-based procedural noise generation to avoid detection by conventional deep-learning-based detection. Finally for the system restoration, the proposed intelligent restoration module managed to provide sustainable results for the black start restoration in energy system.</p>

<p dir="ltr">The safe and reliable delivery of electricity is critical for the functioning of our modern society. However, high-impact, low-probability (HILP) catastrophic events (such as extreme weather caused by climate change, or cyber-physical attacks) pose an ever-growing threat to the power grid. At the same time, modern advancements in computational capabilities, communication infrastructure, and measurement technologies provide opportunities for new operation and control strategies that enhance the resilience of electric power systems to such HILP events. In this work, optimization-based operation and control approaches are proposed to improve resilience in two power systems applications. First, a real-time linearized-trajectory model-predictive controller (LTMPC) is developed for ensuring voltage, frequency, and transient (rotor angle) stability in systems engineered to operate as microgrids. Such microgrids are capable of seamlessly transitioning from grid-connected operation to an islanded mode and thus, enhance system resilience. The proposed LTMPC enables rapid deployment of such systems by reducing engineering costs and development time while maintaining stable operation. On the other hand, some power systems, such as distribution feeders, are not designed to operate as standalone microgrids. For these cases, a method is proposed for forming ad-hoc microgrids from intact sections of the damaged feeder in the aftermath of a HILP event. A feeder operating center-on-a-laptop (FOCAL) is introduced that coordinates the control of possibly hundreds of inverter-interfaced distributed energy resources (e.g., rooftop solar, battery storage) to improve system resilience. Theoretical analysis as well as numerical case studies and simulations of the proposed strategies are presented for both applications.</p>

<p dir="ltr"><b><i>Background</i></b><b>:</b> The increase in EV deployment is presenting numerous energy challenges to the utility distribution infrastructure. The energy demands created by EV charging sessions and the growing call to develop a network of DCFC charging facilities increases operational risk to the utilities in the ability to provide safe and reliable electricity to all customers.</p><p dir="ltr"><b><i>Purpose:</i></b> The purpose of this study is to identify the extent of impact to the utility distribution grid from an increasing EV (electric vehicle) adoption.</p><p dir="ltr"><b><i>Setting</i></b><b>: </b>In total, there were 3,020 rows of distribution circuit feeder data collected from the PG&E DIDF and National Grid NY System Reporting Tool between 2022 – 2023. Additionally, 48 documents, engineering reports, rate filings, articles, research studies, and utility whitepapers were examined.</p><p dir="ltr"><b><i>Research Design:</i></b> Impact analysis using a mixed methodology.</p><p dir="ltr"><b><i>Data Collection and Analysis:</i></b> A single research question was used to formulate an impact analysis to the utility distribution infrastructure under a mixed methodology. A quantitative analysis to determine circuit burden based on historical feeder capacity data and conduct hypothetical impact testing based on a set of ten variables. A qualitative analysis was administered to support these results and further design recommendations for the utility system under a logic model.</p><p dir="ltr"><b><i>Findings:</i></b> The PG&E and Utility National Grid EV and Circuit Impact Analysis demonstrated high susceptibility to overburden under a moderate number of level 2 EV chargers and significantly more when the loading impact was the result of DCFC facilities. The additional exploratory research yielded a consistent theme of mitigation strategies applicable to all electric utilities.</p><p><br></p><p dir="ltr"><b><i>Conclusions</i></b><i>:</i> Portions of the electric distribution infrastructure, operated by hundreds of utilities across the United States must be analyzed, upgraded, and adequately managed under systematic programs which promote facility upgrades, energy management, technology integration, such as AMI. Further, the execution of regulatory strategies for smart policy development and investment into hosting capacity tools are critical to reducing EV impact to the utility.</p><p dir="ltr"><b><i>Keywords</i></b><i>: </i>EV, electric utility, EV grid impacts, EV grid analysis, EV managed charging, EV AMI infrastructure.</p>

<p dir="ltr">In recent years, a surge in data-driven computation has significantly impacted security analysis in cyber-physical systems (CPSs), especially in decentralized environments. This transformation can be attributed to the remarkable computational power offered by high-performance computers (HPCs), coupled with advancements in distributed computing techniques and sophisticated learning algorithms like deep learning and reinforcement learning. Within this context, wireless communication systems and decentralized computing systems emerge as highly suitable environments for leveraging data-driven computation in security analysis. Our research endeavors have focused on exploring the vast potential of various deep learning algorithms within the CPS domains. We have not only delved into the intricacies of existing algorithms but also designed novel approaches tailored to the specific requirements of CPSs. A pivotal aspect of our work was the development of a comprehensive decentralized computing platform prototype, which served as the foundation for simulating complex networking scenarios typical of CPS environments. Within this framework, we harnessed deep learning techniques such as restricted Boltzmann machine (RBM) and deep convolutional neural network (DCNN) to address critical security concerns such as the detection of Quality of Service (QoS) degradation and Denial of Service (DoS) attacks in smart grids. Our experimental results showcased the superior performance of deep learning-based approaches compared to traditional pattern-based methods. Additionally, we devised a decentralized computing system that encompassed a novel decentralized learning algorithm, blockchain-based learning automation, distributed storage for data and models, and cryptography mechanisms to bolster the security and privacy of both data and models. Notably, our prototype demonstrated excellent efficacy, achieving a fine balance between model inference performance and confidentiality. Furthermore, we delved into the integration of domain knowledge from CPSs into our deep learning models. This integration shed light on the vulnerability of these models to dedicated adversarial attacks. Through these multifaceted endeavors, we aim to fortify the security posture of CPSs while unlocking the full potential of data-driven computation in safeguarding critical infrastructures.</p>