Evaluating the safety of unreinforced concrete structures, such as concrete dams, requires an accurate prediction of cracking. Developing a suitable constitutive material model and a reliable computational procedure for analysing cracking processes in concrete has been a challenging and demanding task. Although many analytical methods based on fracture mechanics have been proposed for concrete dams in the last few decades, they have not yet become part of standard design procedures. Few of the current research findings are being implemented by practising engineers when evaluating dam safety. This research is focused on the development of a suitable crack modelling and analysis method for the prediction and study of fracturing in concrete gravity dams, and consequently, for the evaluation of dam safety against cracking. The research aims to contribute to the continuing research efforts into mastering the mechanics of cracking in concrete dams. An analytical method for the purpose of establishing a crack constitutive model and implementing the model for the fracture analysis of concrete structures, in particular massive concrete gravity dams under static loading conditions, has been developed, verified and applied in the safety evaluation of a concrete gravity dam. The constitutive material model is based on non-linear fracture mechanics and assumes a bilinear softening response. The crack model has various improved features: (1) an enhanced mode I bilinear strain-softening approach has been put forward; (2) a new formula for bilinear softening parameters has been developed and their relation with linear softening has been outlined; (3) the influence of bilinear softening parameters on the cracking response has been studied; and (4) an enhanced modification to the shear retention factor which depends on the crack normal strain is included. The material model has been incorporated into a finite element analysis using a smeared crack approach. A sub-program was specially coded for this research. The validity of the proposed cracking model and the computational procedure developed for the purpose of analyzing the tensile fracture behaviour of concrete structures has been confirmed by verification on various concrete structures, including beams, a dam model and actual gravity dams. The crack modelling technique developed was successfully used in evaluating the safety of an existing concrete gravity dam in South Africa and adequately predicted the cracking response of the dam structure under static loadings. The main conclusions drawn are as follows: <ul><li>Both mode I and mode II fracture have been modelled successfully.</li> <li>The proposed bilinear softening model remains relatively simple to implement but significantly improves on predicting the softening response of “small-scale” concrete structures.</li> <li>Both plane stress and plane strain crack analyses have been considered and can be confidently adopted in two-dimensional applications.</li> <li>The proposed method is mesh objective.</li> <li>The crack modelling method developed can correctly predict the crack propagation trajectory and the structural behaviour with regard to fracturing in concrete structures.</li> <li>If not considering shear stress concentration near the tip of a crack, constitutive crack analysis normally indicates a higher safety factor and a higher Imminent Failure Flood (IFF) than the classical methods in the analysis of concrete gravity dams for safety evaluation.</li></ul> / Thesis (PhD(Civil Engineering))--University of Pretoria, 2007. / Civil Engineering / PhD / unrestricted

I Sverige finns det idag ett stort antal dammar och nära hälften av Sveriges elproduktion kommer ifrån vattenkraft. Höga säkerhetskrav ställs på dammarna då konsekvenserna av ett dammbrott kan orsaka stor ekonomisk skada både i form av minskad produktion och som uppbyggnadskostnader. Dammbrott kan även utgöra risk för skador på människor och omkringliggande miljö samt byggnader. Det finns flera dimensioneringsmetoder för att uppfylla gällande säkerhetskrav på konstruktioner. De tillämpningsvägledningar som används vid stabilitetsanalys av betongdammar baseras på gamla deterministiska metoder med säkerhetsfaktorer. Det arbetas med att ta fram nya tillämpningsvägledningar baserat på sannolikhetsbaserade metoder. De sannolikhetsbaserade metoderna tar hänsyn till osäkerheterna i enskilda variabler, vilket förväntas ge effektivare konstruktioner. Stabilitetsanalys med sannolikhetsbaserade metoder i varje enskilt fall är förmodligen det bästa ur säkerhetssynpunkt, men det kan vara väldigt tidskrävande. Ett sett att underlätta stabilitetsberäkningarna, i vanligt förekommande konstruktioner, är användandet av partialkoefficienter. Partialkoefficienter är en semi-probabilistisk metod som kan kalibreras med hjälp av tillförlitlighetsanalysen och appliceras på systemets olika variabler och parametrar. På så vis beaktas osäkerheten i enskilda variabler och parametrar bättre än i deterministiska metoder med säkerhetsfaktorer. Målet med det här examensarbetet var att undersöka om stabilitetsanalys av betongdammar med partialkoefficienter kan vara ett alternativ till de metoder som redan finns. Rapporten kan delas in i tre delar. Den första delen av rapporten beskriver teorin bakom kalibrering av partialkoefficienter med hjälp av tillförlitlighetsteori, FORM. I den andra delen beskrivs metoder och tillämpningsvägledning för stabilitetsanalys av betongdammar med probabilistiska metoder. I den tredje delen används kunskaperna från de två första delarna för att beräkna tillförlitligheten på 15 utav Sveriges dammar. Beräkningarna används sedan för att kalibrera partialkoefficienter. En del kraftiga avgränsningar har gjorts i arbetet, bland annat har beräkningarna utförts för ett statiskt lastfall och bara fokuserat på stabilitetsvillkoret för glidning. Resultatet visar att det inte är rekommenderat att använda sig av partialkoefficienter vid stabilitetsanalys, åtminstone inte med den information och kunskap som idag finns tillgänglig. Osäkerheterna kring vissa modeller och variabler behöver minskas för att partialkoefficienter skall vara ett alternativ. Teorin kring partialkoefficienter kräver även en viss likhet i konstruktioner och konstruktionselement. De dammar som användes i denna studie kan ha varit för olika för att erhålla tillfredsställande resultat med avseende på likartade partialkoefficienter. / In Sweden today there are a large number of dams and nearly half of Sweden's electricity is produced from hydropower. The safety requirements on the dams are high due to the consequences that a dam failure can cause. There are several design methods to achieve the expected safety requirements. The design guidelines used in the stability analysis of concrete dams in Sweden is based on the deterministic methods with safety factors. However, a new proposition for design guidelines based on probabilistic methods is being developed. Compared to the deterministic approach the probabilistic method takes into account the uncertainties in individual variables, which are expected to provide more efficient structures. One problem with stability analysis using probabilistic methods is that it can be very time consuming. Another method which combines the simplicity of the deterministic approach with the effectiveness of the probabilistic method is the use of partial factors. The use of partial factors is a semi-probabilistic method that can be calibrated from the reliability analysis in the probabilistic method and applied to the individual variables and parameters in the system similar to the safety factor. The aim of this thesis was to investigate if stability analysis of concrete dams with partial factors can be an alternative to the methods already available. The report can be divided into three parts. The first part of the report describes the theory behind the calibration process of partial factors using the first order reliability method, FORM. The second part describes existing Swedish methods and application guidelines for stability analysis of concrete dams. The knowledge from the first two parts is then used in order to calculate the reliability of 15 of Sweden's dams. It should be observed that the work contains some limitations, for instance only the sliding stability is studied using one static load case. The results from the calculation of this load case are then used to calibrate the partial factors. The result shows that it is not recommended to use the partial factors for stability analysis of concrete dams, at least not with the information and knowledge available today in the field. The uncertainties surrounding certain models and variables need to be reduced in order for partial factors to be an option. The theory behind partial factors requires a certain degree of uniformness when it comes to the structure or elements used. The difference among the dams analyzed in this study may have impacted the results negatively.

Dams are designed and assessed based on traditional factor of safety methodology. Several drawbacks of this approach exist; for example varying failure probability for structures where the factor of safety is the same. This traditional factor of safety methodology imposes conservative assumptions in terms of both design and analysis. A probability-based analysis has been suggested to account for the omission of uncertainties and provide a less conservative analysis (Westberg &amp; Johansson, 2014). Through the stability analyses of three existing dam structures, a minimum level of reliability or maximum failure probability may be calculated with the ultimate goal of defining a target safety index (β-target) for buttress and gravity dams. These analyses shall in turn contribute to the formulation of a probability-based guideline for the design and assessment of Swedish concrete dams. This probability-based guideline shall be known as the ‘Probabilistic Model Code for Concrete Dams.’ The calculations carried out in this study adhere to the methodologies and specifications set out in the preliminary draft of the Probabilistic Model Code for Concrete Dams. These methodologies encompass analyses within two dominating failure modes for concrete dams; sliding stability and overturning stability. Various load combinations have been modelled for each dam structure to account for the probabilistic failure of each dam under commonly occurring circumstances. A parametric study has been carried out in order to provide insight into the contribution that existing rock bolts provide to the stability of each dam. Furthermore, a study has been carried out into the existence of a persistent rock joint or failure plane in the rock foundation and the effects its presence would have on the sliding stability of a dam. Finally a discussion had been carried out in order to provide suggestions into the formulation of a target safety index through the data envisaged by our analyses for the design and assessment of Swedish concrete dams.

La modélisation des problèmes d'interactions sol-structure et fluide-structure couvre plusieurs domaines de recherche très actifs qui traitent une multitude d'aspects tels que la géométrie non bornée du sol et dans certains cas du fluide stocké, les effets dissipatifs visqueux et radiatifs, l'application du chargement sismique, le choix des variables de base, les propriétés algébriques des systèmes d'équations résultant du couplage,... etc. Dans le présent travail, différents modèles numériques de couplage sol-structure et fluide-structure ont été examinés. Les limites de troncature géométrique du sol et du fluide on été traitées avec des éléments infinis dont les performances ont été comparées à celles des conditions de radiations. Le problème de vibrations libres couplées des systèmes fluide-structure a été résolu en introduisant de nouvelles techniques de symétrisation efficaces. De plus, une nouvelle formulation symétrique en éléments de frontière a été proposée. Cette formulation permet de produire une matrice symétrique définie positive et aboutit ainsi à un système algébrique similaire à celui qui découle de la discrétisation en éléments finis. La matrice bâtie dite "raideur équivalente" peut facilement être assemblée ou couplée avec les matrices de la formulation en éléments finis. Toutes les applications qui ont servi soit à comparer des modèles soit à valider les programmes développés, ont été effectué es dans le cas des barrages poids en béton. Ce cas constitue un problème de couplage fluide-sol-structure typique / Modeling fluid-structure and soil-structure interaction problems covers several research fields dealing with multiple aspects such as : unbounded geometry of soil media and in some cases of retained fluid, viscous and radiation dissipative effects, application of seismic loading, choice of the basic variables, algebraic properties of the resulting coupled system, ... etc. In this work, different numerical models of soil-structure and fluid-structure coupling have been studied. The truncation boundaries of the soil and of the fluid domains have been considered by using infinite elements whose performances were compared to those of radiation conditions. The problem of the free coupled vibrations of the fluid-structure systems has been solved by introducing efficient symmetrization techniques. In addition, a new symmetric boundary element formulation is proposed. It allows to produce a positive definite and symmetric matrix and therefore to conduct to an algebraic system similar to the one obtained from finite element discretization. The produced matrix called "equivalent stiffness matrix" can easily be assembled or coupled to finite element matrices. All of the applications which have been done in order to compare models or to validate the developed programs were built in the case of concrete gravity dams, which constitute typical coupled fluid- soil-structure problems

The most common failure of concrete gravity dams is sliding along the foundation surface. This thesis studies progressive failure of the Longtan dam on the upper Hongshuie river in china.Two methods are used in this thesis; the Safety Reserve Factor (SRF) method and the Overload method. The SRF-method is used as a tool to study sliding failure along the foundation weak layer of the Longtan dam. Strength reduction coefficients decrease the cohesion and friction angle values for the weak layer of the foundation. Simulations with reduced shear strength parameter values gives information about the development of the plastic zone. The ultimate bearing resistance and the failure path along the foundation are obtained. The safety reserve coefficient is established through the strength reduction coefficients, when the plastic zone of the foundation is totally coalescent. To analyse the development of the plastic zone along the dam foundation with the strength reserve method, the commercial finite element software MSC.Marc is used. The results of the Safety Reserve Factor method (SRF) show that the failure of the dam is highly related to the strength of the interface between the dam and rock foundation. The strength reserve factor is determined to 2.4. The Overload method gives a visual deformation shape of the dam structure and pressure load at the moment of failure. / Dammar har används i mer än 5000 år (Yang et al. 1999) och är fortfarande en viktig källa för energiutvinning. Det största antalet dammar finns i Kina och man tror att det finns över 80,000 dammar i landet (Shapiro 2001). En ökning av dammars kapacitet och antal sker, vilket resulterat i ett behov av bättre sätt för att utvärdera säkerhetsparametrar som betongkvalité, styvhet och homogenitet av bergmassa. Simuleringar av dammkonstruktioner kan ge värdefull information om dessa parametrar och kan därigenom bidra till att förbättra en damms stabilitet och sänka konstruktionskostnader. I det här examensarbetet utvärderas och modelleras Longtan dammen i övre Hongshui floden i Kina. Den vanligaste orsaken till haveri av betongdammar är glidning mellan betonglager och bergmassan. I den här rapporten simuleras ett svagare lager mellan betong och bergmassa och utvecklingen av kontaktbrottsvägen visualiseras och utvärderas med progressiva haveri metoden. Det svagare lagrets hållfasthetskoefficienter; friktionsvinkel och kohesion, reduceras för att analysera utbredningen av den plastiska zonen. Resultaten visar att en överskridning av draghållfastheten för betong börjar vid dammens häl och att kompressionsbristningsgränsen överskrids vid dammens tå när hållfasthetskoefficienterna reduceras. Säkerhetsreservskoefficienten för Longtan dammen erhålls då gränsytan mellan betong och berg är helt plastisk. För att analysera utbredningen av den plastiska zonen längs med dammfundamentet med progressiva haveri metoden används den kommersiella finita element mjukvaran MSC.Marc. Resultaten från säkerhetskoefficientfaktormetoden visar att ett haveri av gravitations dammar är i hög grad relaterad till hållfastheten mellan betongen och bergets gränsyta. Säkerhetskoefficientfaktorn bestäms till 2.4. För att bestämma Longtan dammens maximala vattenbelastningskapacitet används Överbelastningsmetoden. Det maximala vattentryck som Longtan dammen klarar av utan att haverera simuleras med hjälp av en vätska vars densitet ökas mellan modellkörningar. Dammens förskjutning når till sist en punkt där den ökar kraftigt och dammens mutationsdeformationstillstånd har uppnåtts. En visuell deformationsbeskrivning av dammen ges genom modellkörningar och överbelastningsfaktorn bestäms.