Este trabalho aborda a temática de segurança de barragens de gravidade e apresenta uma metodologia para análise por instrumentação colocada no maciço destas estruturas. Como estudo de caso, utilizou-se a barragem da Usina Hidrelétrica de Dona Francisca, construída no final do ano de 2000, no Rio Jacuí, na região central do Estado do Rio Grande do Sul, Brasil. Sua estrutura é de concreto compactado com rolo e possui cerca de 63 metros de altura máxima e 660 metros de comprimento. As condições geológico-geotécnicas da fundação da UHE Dona Francisca remetem às rochas da Formação Caturrita, constituídas principalmente por arenitos e níveis de siltito e argilito intercalados. Devido à preocupação com o comportamento deste tipo de fundação em relação à estabilidade das estruturas, a UHE Dona Francisca foi contemplada com um amplo plano de instrumentação, composto principalmente por medidores de vazão, piezômetros, pêndulos diretos, extensômetros e medidores triortogonais de junta. Esta dissertação apresenta e analisa os dados destes dispositivos. Os objetivos principais foram avaliar a evolução dos registros ao longo do tempo e frente às variações de carregamento e condições ambientais, além de comparar as leituras com as previsões de projeto e com os dados de instrumentação de outras barragens. Observou-se que a maioria da instrumentação é sensível à variação do nível do reservatório e, portanto, o mesmo foi tomado como referencial nas apresentações das leituras dos instrumentos. De um modo geral, a barragem esteve estável durante o período auscultado de 17 anos. As poropressões foram controladas satisfatoriamente pelo sistema de drenagem. Na maioria das vezes, os deslocamentos, as deformações na fundação e vazões foram superiores aos registros de outras barragens de maior porte. Entretanto, essas grandezas tenderam a reduzir ou se estabilizar ao longo dos anos. Atualmente, alguns instrumentos exigem atenção visto que suas leituras não estão estáveis ou não correspondem às previsões de comportamento estabelecidas durante a fase de projeto. A metodologia para análise adotada neste trabalho mostrou-se eficaz e permitiu a definição das tendências atuais de leituras de cada instrumento. / This work deals with the safety of gravity dams and presents a methodology for analysis by instrumentation inserted in the massif of these structures. The Dona Francisca Hydroelectric dam was used as a case study. This dam was built in the end of 2000, on the Jacuí River, in the central region of the Rio Grande do Sul State, Southern Brazil. Its structure is made of rolled concrete and has approximately 63 meters of maximum height and 660 meters in length. The geological and geotechnical conditions of the Dona Francisca HPP foundation refer to the Caturrita Formation rocks, made up mainly of sandstones and intercalated levels of siltstone and claystone. Due to the concern about the behavior of this type of foundation in relation to the stability of the structures, the Dona Francisca HPP was contemplated with a broad instrumentation plan, composed mainly of flow meters, piezometers, direct pendulums, extensometers and triorthogonal meters. This dissertation presents and analyzes the data of these devices. The goals were to evaluate the evolution of these registers over time and in relation to loading variations and environmental conditions, moreover, compare the readings with the project forecasts and the instrumentation data from other dams. It was observed that most of the instrumentation is sensitive to variation of the reservoir level and, therefore, it was taken as reference in the presentations of the instruments readings. In general, the dam of the Dona Francisca HPP was stable during this auscultated period of 17 years. The pore water pressures were successfully controlled by the drainage system. Most of the time, the displacements, the deformations in the foundation and flows were higher to the records of other larger dams. However, these magnitudes have tended to decrease or stabilize over the years. Nowadays, some auscultation devices require attention because their readings are not stable or do not match the behavior predictions established during the design stage. The analysis methodology adopted in this study proved to be effective and allowed the definition of each instrument reading trends.

Hydroelectric power is a commonly used alternative source of energy in developing countries. In this regard, concrete gravity dams are the most preferred dam type especially with the developments in the engineering industry. Roller compacted concrete became more popular in dam construction due to its advantages of speed and economy. Several methods are used for the design of concrete gravity dams by analyzing the dam response under static and dynamic loads. This study provides three dimensional linear dynamic analysis of roller compacted concrete gravity dam with a complete dam-foundation-water interaction by using EACD-3D-08 program. Foundation flexibility was included with damping and mass using boundary elements. Three dimensional solid elements were used for the idealization of the dam and water with using finite element methods. Compressibility of water with reservoir absorption was studied. In the light of USACE, performance criteria of linear analyses were assessed. Parametric study was conducted to determine the most influential parameters on the dam response. The importance and necessity of three dimensional analyses were investigated by comparing with linear two dimensional analyses. Linear analyses were then compared with three dimensional nonlinear analyses. In conclusion, the realistic dam seismic response can only be obtained by using three dimensional linear analyses with full interaction of dam-foundation-water.

Internationellt sett är merparten av de dammbrott som ägt rum kopplat till grundläggningen. Det är oftast kopplat till att det finns osäkerhet i de rådande grundförhållandena. Som extra säkerhet, installerades bergförankringar som en åtgärd att behandla osäkerheterna i de rådande grundförhållandena. I Sverige är merparten av kraftverksdammarna uppförda mellan 1940-och1960 talet. Det var inte ovanligt att cementingjutna slakarmerade kamstål sattes under dammarna som en extra säkerhet. När dessa bergbultar installerades var kunskapen om dess verkningssätt begränsat, då teorier om bergbultens verkningssätt började utvecklas under 70- och 80-talet. Vid dimensionering av bergförankring till dammbyggnader i nutid tillämpas RIDAS, Kraftföretagens riktlinjer för dammsäkerhet, som är upprättade av de Svenska dammägarna. I RIDAS framgår att när det endast föreligger en risk för glidning får rostfri slak armering användas som bergförstärkning och dimensioneras enligt BBK, Boverkets handbok för betongkonstruktioner, kap 3.11 "kraftöverföring genom fog". Examensarbetets inledande del är en litteraturstudie som är inriktad på den forskning som har bedrivits rörande bergbultens principiella beteende när den går till brott. Vidare i litteraturstudien har en beräkningsteori som utvecklats av Holmberg (1991) vilken beskriver en bergbults bidrag till bärförmågan för en bergspricka studerats i detalj. Teorin har även jämförts mot experimentellt utförda skjuvförsök av bergbultar. Med hjälp av den teori som presenterades i litteraturstudien genomfördes en stabilitetsanalys av en betongmonolit i utskovsdammen på en typisk Svensk vattenkraftstation. Målet var att undersöka möjligheten för bergbultarna, som är installerade i berggrunden att samverka, med ett antaget horisontell sprickplan under dammbyggnaden. Det bidrag till bärförmågan som bergbultarna och bergsprickan utgör till totalstabiliteten jämförs med de krav som ställs i RIDAS. I slutet av arbetet förs en diskussion om de presenterade teoriernas förmåga att beskriva bultens bärförmåga under olika förhållanden samt hur en bergförankring bör utformas för att säkerställa dragbrott i bulten. Efter det ges förslag på forskning rörande bergbultning som kan utföras, för att bättre förstå bergbultens principiella beteende. / Internationally, the majority of dam failures that have occurred are related to the foundation. It is usually linked to the large uncertainty in the actual propertys of the foundation. A common measure was to install rock bolts as an extra precaution in the subgrade to handle the uncertainties of the subgrade. In Sweden, most of the hydroelectric dams were constructed between 1940 and 1960. It was not uncommon that cemented reinforced rock bolts were installed in the subsoil as an extra precaution. When these rock bolts were installed, understanding of their stress handling was limited, considering the fact that the theory of the rock bolts behavior was first studied between 1970 and 1980. The first part of this thesis is a thorough review of the literature which focuses on the research that has been conducted on the fundamental behavior of the rock bolt when it fails. In this study, a computational theory describing a rock bolts load contribution to a rock joint is presented, verified against experimental shear tests performed on rock bolts. In the design of rock anchoring to the construction of reservoirs in the present application RIDAS, Power Companies' guidelines for dam safety, the guidelines is formed by the Swedish dam owners. Accordingly to RIDAS where there is only a risk for a slide in a rock joint stainless untensioned rock bolts are allowed to be used fore support and dimension according to BBK, Boverkets handbok för betongkonstruktioner, kap 3.11 "kraftöverföring genom fog". Using the theory presented in the literature review, a stability analysis of a monolith on Långbjörn PowerStation was carried out. The goal was to explore the possibility of rock bolts installed in the subgrade to interoperate with a rock joints under the dam building. The load contribution of rock bolts and the rock joints has to overall stability is compared with the requirements of the Swedish power company’s guidelines for dam safety, RIDAS. At the end of the thesis, the presented theories’ applicability under different conditions and how a rock anchor should be designed to be as efficient as possible is discussed. In conclusion, suggestions for research that can be performed on bolting, to gain better understanding of the fundamental behavior of the rock bolt are presented.

A subpressão consiste em um dos maiores problemas no que diz respeito à segurança de obras hidráulicas. A execução de vedações com cortinas de injeção de cimento, associadas à drenagem é de suma importância para coibir as ações da água percolada através de suas fundações. Torna-se, portanto, indispensável conhecer elementos para a verificação do bom funcionamento destes sistemas. Grande parte dos critérios de subpressão utilizados hoje no dimensionamento de barragens é baseada em métodos determinísticos e empíricos, com origem em experiência acumulada sob certas condições específicas. Quando tais critérios são utilizados para lugares específicos, com feições geológicas peculiares, podem não refletir os dados utilizados anteriormente. Tanto projetos sem segurança como antieconômicos podem resultar desta prática. Neste trabalho, procura-se avaliar a eficácia dos sistemas de vedação e drenagem de maneira probabilística, utilizando métodos estatísticos e dados obtidos durante a execução de obras já prontas, de modo a considerar a heterogeneidade dos maciços de fundação. A primeira parte da pesquisa consistiu, portanto, na compilação e ajuste de curvas de distribuição estatística aos valores de absorções de água e calda de cimento das barragens da Usina Hidrelétrica de Salto Caxias e da derivação do Rio Jordão. As duas localizam-se no Estado do Paraná e são de propriedade da COPEL - Companhia Paranaense de Energia. Ambas são fundadas em rochas basálticas, mas possuem diferentes características de permeabilidade. A eficácia de uma cortina de injeção pode ser dividida em duas: na dos trabalhos de injeção, com o controle dos valores de calda de cimento absorvida, e na da cortina em si, através da verificação da diminuição das subpressões e das vazões de percolação. Assim, através da análise da tendência das curvas de distribuição encontradas, foi possível comprovar o bom resultado dos trabalhos de injeção realizados nas duas obras. Na segunda parte da pesquisa, é proposta uma metodologia de análise probabilística da eficácia das cortinas de vedação e de drenagem. Esta metodologia baseia-se nos princípios do método de Monte Carlo, gerando-se cenários aleatórios de acordo com as curvas de distribuição estatísticas obtidas da primeira parte do trabalho. Através de simulações numéricas com elementos finitos, com valores de permeabilidades das fundações obedecendo àquelas distribuições estatísticas, foi possível obter uma distribuição estatística das subpressões. Com isso, pode-se proceder a análises de confiabilidade, e probabilidades de falha podem ser obtidas utilizando leituras de instrumentação ou mesmo critérios de subpressão. / One of the biggest problems for hydraulic works safety is the uplift pressure. The execution of grouting, associated with drainage, is very important to control seepage through concrete dam foundations. Therefore, appropriate means to check the efficiency of those systems are essential for dam safety evaluation. Most of the uplift pressure criteria used nowadays for dam design is based on deterministic and empirical methods, originated from accumulated experience under certain specific conditions. When such criteria are used for a specific site, peculiar geological features may not be reflected in the data basis used as previous experience. Either unsafe or uneconomic designs may result from such practice. In this work, grouting and drainage systems effectiveness are evaluated in a probabilistic way, using statistical methods and data obtained during grouting execution, which provide good assessment for foundation properties variability. The first part of the research consisted, therefore, in the compilation and statistical distribution curves adjustment of data from Salto Caxias and Jordão river Hydroelectric Power Plants water pressure tests and grouting takes. The two dams are located in the state of Paraná, Brazil, and are owned by COPEL - Companhia Paranaense de Energia. They are founded on basaltic rocks, but have different permeability characteristics. Grout curtain effectiveness can be divided in two: in the grouting operation, with the treatment of cement grout take data, and in the curtain, which is measured by uplift and seepage reductions. Thus, by analyzing statistical distribution trends, it was possible to verify treatments effectiveness in the two dams. In the second part of the research, a probabilistic analysis methodology of drainage and grouting curtains effectiveness is proposed. This methodology is based on Monte Carlo method principles, randomly generating scenarios according to statistical distributions obtained in the first part of the work. By means of numerical simulations with finite elements, using foundation permeability coefficients following those statistical distributions, it was also possible to obtain a statistical distribution of uplift forces. Reliability analyses can be performed, and failure probabilities can be obtained using instrumentation readings, or even any uplift pressure criterion.

A subpressão consiste em um dos maiores problemas no que diz respeito à segurança de obras hidráulicas. A execução de vedações com cortinas de injeção de cimento, associadas à drenagem é de suma importância para coibir as ações da água percolada através de suas fundações. Torna-se, portanto, indispensável conhecer elementos para a verificação do bom funcionamento destes sistemas. Grande parte dos critérios de subpressão utilizados hoje no dimensionamento de barragens é baseada em métodos determinísticos e empíricos, com origem em experiência acumulada sob certas condições específicas. Quando tais critérios são utilizados para lugares específicos, com feições geológicas peculiares, podem não refletir os dados utilizados anteriormente. Tanto projetos sem segurança como antieconômicos podem resultar desta prática. Neste trabalho, procura-se avaliar a eficácia dos sistemas de vedação e drenagem de maneira probabilística, utilizando métodos estatísticos e dados obtidos durante a execução de obras já prontas, de modo a considerar a heterogeneidade dos maciços de fundação. A primeira parte da pesquisa consistiu, portanto, na compilação e ajuste de curvas de distribuição estatística aos valores de absorções de água e calda de cimento das barragens da Usina Hidrelétrica de Salto Caxias e da derivação do Rio Jordão. As duas localizam-se no Estado do Paraná e são de propriedade da COPEL - Companhia Paranaense de Energia. Ambas são fundadas em rochas basálticas, mas possuem diferentes características de permeabilidade. A eficácia de uma cortina de injeção pode ser dividida em duas: na dos trabalhos de injeção, com o controle dos valores de calda de cimento absorvida, e na da cortina em si, através da verificação da diminuição das subpressões e das vazões de percolação. Assim, através da análise da tendência das curvas de distribuição encontradas, foi possível comprovar o bom resultado dos trabalhos de injeção realizados nas duas obras. Na segunda parte da pesquisa, é proposta uma metodologia de análise probabilística da eficácia das cortinas de vedação e de drenagem. Esta metodologia baseia-se nos princípios do método de Monte Carlo, gerando-se cenários aleatórios de acordo com as curvas de distribuição estatísticas obtidas da primeira parte do trabalho. Através de simulações numéricas com elementos finitos, com valores de permeabilidades das fundações obedecendo àquelas distribuições estatísticas, foi possível obter uma distribuição estatística das subpressões. Com isso, pode-se proceder a análises de confiabilidade, e probabilidades de falha podem ser obtidas utilizando leituras de instrumentação ou mesmo critérios de subpressão. / One of the biggest problems for hydraulic works safety is the uplift pressure. The execution of grouting, associated with drainage, is very important to control seepage through concrete dam foundations. Therefore, appropriate means to check the efficiency of those systems are essential for dam safety evaluation. Most of the uplift pressure criteria used nowadays for dam design is based on deterministic and empirical methods, originated from accumulated experience under certain specific conditions. When such criteria are used for a specific site, peculiar geological features may not be reflected in the data basis used as previous experience. Either unsafe or uneconomic designs may result from such practice. In this work, grouting and drainage systems effectiveness are evaluated in a probabilistic way, using statistical methods and data obtained during grouting execution, which provide good assessment for foundation properties variability. The first part of the research consisted, therefore, in the compilation and statistical distribution curves adjustment of data from Salto Caxias and Jordão river Hydroelectric Power Plants water pressure tests and grouting takes. The two dams are located in the state of Paraná, Brazil, and are owned by COPEL - Companhia Paranaense de Energia. They are founded on basaltic rocks, but have different permeability characteristics. Grout curtain effectiveness can be divided in two: in the grouting operation, with the treatment of cement grout take data, and in the curtain, which is measured by uplift and seepage reductions. Thus, by analyzing statistical distribution trends, it was possible to verify treatments effectiveness in the two dams. In the second part of the research, a probabilistic analysis methodology of drainage and grouting curtains effectiveness is proposed. This methodology is based on Monte Carlo method principles, randomly generating scenarios according to statistical distributions obtained in the first part of the work. By means of numerical simulations with finite elements, using foundation permeability coefficients following those statistical distributions, it was also possible to obtain a statistical distribution of uplift forces. Reliability analyses can be performed, and failure probabilities can be obtained using instrumentation readings, or even any uplift pressure criterion.

Water structure Ivanské jezero was built on the Javorník Creek back between 1905-1907. The main purpose of the dam was accumulation of high volumes of water and protection of the territory downstream. After the adjustment for higher flows on the river Kněžna in Rychnov nad Kněžnou, which was completed in 1937, the water structure lost the protective function. On the basis of an assessment made by Povodí Labe in 1993 - due to the low volume of the reservoir there is no flood wave transformation and already while flow rates greater than Q50 pass through dam, it results in overflowing the dam. To this date water structure Ivanské jezero is in very poor condition due to degradation of the binding material in the dam. The goal of the thesis is to propose a procedure for removal of the existing structure and the subsequent construction of the design and construction of a new – gravity, concrete dam. This work will contain technological removal process of current dam and possible usage options for the dam material. In the next step a new structure will be designed in the place of existing dam. After new design is completed there will be hydraulic assessment of the functional objects, hydrology and stability on the dam.

Concrete has for decades been one of the main materials used to construct important structuressuch as dams and bridges. However, after years of service, the concrete structuresstart deteriorating and signs of damages start showing on the structure. The inspectionof such structures is compulsory to assess their state and plan repairing operations ifnecessary. The main inspection method has for long been field inspection, however, thismethod presents several problems, namely, the complexity to access some parts of thestructures and the subjectivity of the decisions. These difficulties make the operationtime-consuming and prone to error, thus a need for a new methodology that would benumerical and more automated.One of the damages that affect the carrying capacity the most in concrete structuresare surface-cracks. In this work, the focus is brought on this type of damage and a solutionfor the inspection methodology is presented and applied to a case study.The methodology first consists in using the photogrammetry technology, which allows theproduction of a 3D point cloud model of the structure by taking multiple pictures of it usingUAVs to facilitate access to its complex parts. From this point cloud, surface-openingcracks could be visualized and an orthoimage featuring the damages can be produced.Then, an image-based crack detection program will be used on the orthoimage to extractthe crack. From this, a program is developed to find the coordinates of the crack in thestructure and to process the model of the crack. Then a program is made to importthe crack model to a finite element software, and from there the extended finite elementmethod (XFEM) combined with the concrete damaged plasticity (CDP) method will beused to assess the carrying capacity of the structure and to study the evolution of itscrack pattern.The methodology was tested in a case study. In this case study, a steel-reinforced concretebeam was damaged and all of the methodology steps were applied. An adaptationto some steps was needed due to challenges raised by the test set-up.A three-point bending experience was carried out on the beam while recording the forceand displacement data and the crack pattern evolution. Numerically, this experience wasreproduced and relevant results were extracted and compared to the experimental ones.Differences in the carrying capacity were observed and multiple numerical analyses werecarried out to test the taken assumptions and detect the source of error. On the otherhand, for the crack pattern, satisfying results were achieved. Moreover, the degree of thedetailing in the crack model and its effect on the results is discussed.Globally, it can be concluded that the methodology can effectively in a numericalsemi-automated way capture the surface cracks in concrete structures and import theirmodel to a finite element software to apply analyses to assess the capacity of the damagedstructure and study the evolution of the cracks.This methodology could be further developed in the future by including more technologiessuch as lasergrammetry to make it go from a 2D surface-based damage analysis to a 3Danalysis. Moreover, criteria specialized for inspection and repairing purposes could becreated and implemented in the methodology.

It is of great importance to ensure the structural safety of dams during earthquakes since a failure may cause catastrophic consequences. Conventional computation of the structural response of dams is based on a simplified approach where the foundation is considered as massless. However, recent developments have produced several new analysis methods that consider the foundation mass, modelled with absorbing boundaries and free-field forces. These newer methods are intended to simulate the seismic structural response more accurately, optimize the design and minimise future unnecessary reparations. The aim of the thesis was to investigate the influence of foundation modelling in seismic time history analyses. This was done by comparing the established massless foundation approach to two approaches with foundation mass and free-field forces included; the analytical approach presented by Song et al. (2018) and the direct FE approach by Løkke (2018). Both the efficiency of the seismic wave propagation simulation and the structural response of the dam were of interest, and points on the dam and foundation were studied to accurately compare these modelling approaches. The time history analyses showed that the massless approach corresponded perfectly with the ideal theoretical velocity at the foundation surface when studying only the foundation block, as expected. The analytical and direct FE however, differed slightly from the theoretical value but still gave an accurate representation. Both methods using free-field forces obtained equivalent and realistic structural responses when studying the dam-reservoir-foundation model. The massless method however,strongly overestimated the dam response and was therefore found to not capture the actual behavior of the dam accurately, despite modifications such as increased material damping in the concrete. Additionally, another aim was to analyse the influence of modelling in 2D versus 3D for determining the dynamic characteristics of the dam such as natural frequencies and eigenmodes of the dam. These frequency analyses were made using models with and without foundation mass considered and was compared to experimental data.The massless 3D model was found to be the most effective modelling approach for deriving the dynamic characteristics of the dam since the use of a 3D model was necessary in order to study the behaviour of the whole dam and post-processing was simpler when using the massless model. / Det är nödvändigt att säkerställa dammars säkerhet mot jordbävningar i design-processen eftersom ett dammbrott kan få katastrofala konsekvenser. Traditionellt används förenklade beräkningar där dammens strukturella respons beräknas med en berggrund där bergets massa är försummad. Den senaste tiden har flera nya analysmetoder tagits fram, som tar hänsyn till bergets massa och är modellerade med absorberande randvillkor och free-field forces. De nyare metoderna förväntas modellera de seismiska krafterna mer exakt för att optimera designen och minimera onödiga reparationer. Syftet med projektet var att undersöka inverkan från olika metoders sätt att beakta berggrunden vid seismiska analyser. Det utfördes genom att jämföra den etablerade masslösa metoden med två metoder som beaktar bergmassan och free-fieldforces; den analytiska metoden av Song et al. (2018) och Direct FE-metoden av Løkke (2018). Både effektiviteten i den seismiska vågutbredningssimuleringen och dammens strukturella respons var av intresse. Modelleringsmetoderna jämfördes genom att studera punkter på både dammen och berget. När enbart berggrunden studerades med den masslösa metoden så erhölls, som förväntat, god överenstämmelse med den ideala teoretiska hastigheten på bergsytan. De analytiska och Direct FE metoderna skiljde sig marginellt från det teoretiska värdet men gav fortfarande en korrekt hastighet på bergsytan. Vid analys av modeller med dam och reservoar inkluderade, gav metoderna som använde free-field forces ekvivalenta och realistiska strukturella responser. Den masslösa metoden däremot, överskattade kraftigt dammens respons och ansågs därför inte modelleradet verkliga beteendet hos dammen på ett korrekt sätt, trots modifieringar med ökad materialdämpning i betongen. Ett annat syfte var att analysera påverkan av modellering i 2D kontra 3D för att bestämma dammens dynamiska egenskaper, som egenfrekvenser och egenmoder. Dessa frekvensanalyser gjordes med hjälp av modeller som både beaktade och försummade bergets massa, och jämfördes med experimentella data. Den masslösa 3D-modellen visade sig vara den mest effektiva modelleringsmetoden för att erhållade dynamiska egenskaperna hos dammen. Det eftersom en 3D-modell var nödvändig för att studera hela dammens beteende och hantering av utdata var förenklad vid användning av den masslösa modellen.