Publications

The Baixo Sabor hydroelectric power plant, located in the northeast of Portugal, includes a reservoir with a storage capacity of 630 million m3 of water created by a concrete double-curvature arch dam, 100 m high, which was equipped with a 20-channel vibration based structural health monitoring system.This paper, apart from describing the dam, presents the main characteristics of the monitoring system and the results obtained during the first 6 months of operation and the first filling of the reservoir. These results include in particular the characterization of the accelerations amplitude and the evolution of the dam modal parameters (natural frequencies, modal damping ratios and mode shapes). It is particular interesting to observe the influence of the water level on the modal parameters and the development of statistical models to mitigate this effect.

Earthquakes cause severe damage to masonry structures due to inertial forces acting in the normaldirection to the plane of the walls. The out-of-plane behavior of masonry walls is complex anddepends on several parameters, such as material and geometric properties of walls, connectionsbetween structural elements, the characteristics of the input motions, among others. Differentanalytical methods and advanced numerical modeling are usually used for evaluating the out-ofplanebehavior of masonry structures. Furthermore, different types of structural analysis can beadopted for this complex behavior, such as limit analysis, pushover, or nonlinear dynamic analysis.Aiming to evaluate the capabilities of different approaches to similar problems, blind predictionswere made using different approaches. For this purpose, two idealized structures were tested on ashaking table and several experts on masonry structures were invited to present blind predictions on theresponse of the structures, aiming at evaluating the available tools for the out-of-plane assessment ofmasonry structures. This article presents the results of the blind test predictions and the comparisonwith the experimental results, namely in terms of formed collapsed mechanisms and control outputs(PGA or maximum displacements), taking into account the selected tools to perform the analysis.

The analysis of the shaking table test of a 3-wall stone masonry structure performed with a discreteelement model is presented. The numerical model, created with the code 3DEC, employed a rigidblock representation and a Mohr-Coulomb joint model. Joint stiffness calibration to match theexperimental natural frequencies is discussed, as well as the boundary conditions to simulate theshake table. Comparisons are made with themeasured displacements at key locations, and themodesof deformation and fracture of the walls. The DEM model was able to reproduce important features ofthe shaking table tests. The experimental deformation and near collapse patterns were clearlyidentifiable in the numerical simulations, which produced displacements within the observed ordersof magnitude, for the various levels of excitation.

This paper presents a back analysis procedure for identification of the elastic parameters of transversely isotropic rock cores, containing an overcoring triaxial strain probe, from the strains measured during a biaxial test. A three-dimensional finite element model was developed to simulate the biaxial test on the overcored rock specimen and to compute the strains at the location of the strain gauges. Different optimisation algorithms were tested and the most suitable one was selected. The back analysis procedure was tested for identification of the five elastic parameters and the two orientation angles that characterise a transversely isotropic rock core. In spite that, with the developed methodology, convergence was reached and all those parameters could be identified, sensitivity analyses demonstrated that the results obtained were not stable and therefore they were not reliable. By introducing constrains based on common practice and previous experience, a stable and robust methodology was achieved: the three elastic parameters, E1, E2 and

The 1:10 scale model of the 15th century Mustafa Pasha Mosque in Skopje, that underwent a comprehensiveshake table program, is modeled by the discrete elements approach. A rigid block model withnonlinear behavior concentrated at the joints was developed and calibrated by comparison with theobserved response. Time domain analyses of the discrete model were performed under the various levelsof dynamic excitation used in the shake table test. Under the lower levels of input, the time and frequencydomain characteristics of the shake table experimental response were fairly well simulated by thenumerical model. This model also predicted well the zones and the level of damages. For the higher inputlevels, the comparison was less satisfactory. Overall, the discrete element approach showed the capabilityto handle the dynamic nonlinear modeling of relatively complex masonry structures.

This work investigates tensile crack propagation in concrete gravity dams by using some recently developed numerical techniques (crack-path field and strain injection techniques). The work carefully addresses aspects related to mesh independence, robustness and computational cost, which are the main issues in fracture modeling. The novel technique consists of a procedure to insert, in the selected domain areas, specific strain fields for enhancing the performance of the underlying finite elements in modeling fracture. Representative numerical simulations of concrete dams show the accuracy and robustness of the methodology.

Block models have been shown to provide a realistic representation of the behavior of many types of masonry structures under static and dynamic loads. When the strength of the units is such that movements along the joints govern the behavior, it is acceptable to make the simplifying assumption that blocks act as rigid bodies. This assumption is particularly useful when dealing with seismic problems, for which the computational times for time domain analysis may be substantial. In this paper, the application of discrete element models for dynamic analysis of masonry structures is addressed. The emphasis is on the seismic behavior of block stone masonry, but the treatment is general to cover other types of masonry. First, the assumptions involved in the choice of a block representation are discussed, stressing in particular the case of rigid block models. Numerical issues are examined, including contact models, calculation of natural frequencies, time stepping algorithms, damping and boundary conditions. A review is presented of modeling examples published in the literature for various types of masonry structures. The choice of numerical representation and its main features are discussed for each case.

Much research in recent years has focused on the seismic analysis of concrete and earthfill dams,and few works have addressed the case of masonry dams. The structural behavior of masonrydams is controlled essentially by its discontinuous nature, which may induce significant nonlinearresponse during an intense earthquake. In this article, a numerical tool based on the Discrete ElementMethod is presented, aimed at the static, dynamic, and hydromechanical analysis of masonry gravitydams. The use of discontinuous models is mandatory for the study of failure mechanisms involvingthe masonry discontinuities, the dam-rock interface or the rock mass joints. The Discrete ElementMethod is able to assemble continuous and discontinuous meshes simultaneously in the same model,providing a versatile tool to consider various assumptions and levels of analysis, ranging from simplifiedto detailed structural representations. A comprehensive study of the seismic behavior of LagoaComprida Dam, located in Portugal, is presented. Both continuous and discontinuous models weredeveloped to assess the main failure mechanisms, including overstress, partial and global sliding,and overturning.

The analysis of the shaking table test of a 3-wall stone masonry structure performed with a discreteelement model is presented. The numerical model, created with the code 3DEC, employed a rigidblock representation and a Mohr-Coulomb joint model. Joint stiffness calibration to match theexperimental natural frequencies is discussed, as well as the boundary conditions to simulate theshake table. Comparisons are made with themeasured displacements at key locations, and themodesof deformation and fracture of the walls. The DEM model was able to reproduce important features ofthe shaking table tests. The experimental deformation and near collapse patterns were clearlyidentifiable in the numerical simulations, which produced displacements within the observed ordersof magnitude, for the various levels of excitation.

A computationally affordable modeling of dynamic fracture phenomena is performed in this study by using strain injectiontechniques and Finite Elements with Embedded strong discontinuities (E-FEM). In the present research, classical strain localizationand strong discontinuity approaches are considered by injecting discontinuous strain and displacement modes in the finite elementformulation without an increase of the total number of degrees of freedom. Following the Continuum Strong DiscontinuityApproach (CSDA), stress

Neste trabalho apresentam-se os primeiros resultados obtidos com um sistema de observação do comportamento dinâmico instalado na barragem do Cabril, em que se mostra a evolução no tempo das primeiras quatro frequências naturais da obra durante um período de observação de cerca de um ano. Descreve-se um conjunto de rotinas implementadas para a identificação automática das frequências naturais da estrutura; a filosofia de gestão dos dados que são adquiridos em contínuo; o controlo remoto do sistema e a transmissão de dados via internet, entre o centro de controlo do sistema instalado em obra e o centro de análise situado no Laboratório Nacional de Engenharia Civil (LNEC). Este sistema, baseado na medição de vibrações em contínuo, permite a caracterização do comportamento dinâmico desta obra ao longo do tempo, tendo em conta os diversos tipos de excitação, tais como o vento, a vibração provocada pelo funcionamento dos grupos de produção de energia e dos órgãos de descarga, micro tremores e sismos de média e elevada intensidade.

Rock blasting in urban settings is always extremely demanding in what regards the control of ground vibrations and other environmental impacts. Extensive underwater blasting was performed in the Leixões Harbour, an important harbour surrounded by a dense urban mesh. Macro-seismometers were installed in the perimeter and recorded more than one thousand blasts. After completion of the works classical curve fitting statistical analysis of peak particle velocity (PPV) as a function of charge and distance to the blast was performed. PPV spatial distributions were determined for the duration of the construction and maps of ground peak vibration velocity drawn. The results showed that the accuracy of the attenuation law is within the usual range for this kind of phenomenon. However, factors like terrain morphology or geology variations cannot easily be included in a regression law but can be treated in a simple numerical model. A dynamic finite-difference model was developed and the influence of these features was studied thus increasing the understanding of the problem and providing valuable insights into some aspects of the usually large dispersion in the PPV.

The presented work explores a combination of strain localization methods and strong discontinuity ap-proaches to remove the flaws (stress-locking and mesh bias dependence) of the classical strain localization methods, and reduce the sophistication of the strong-discontinuity based methods. The concept of strong discontinuity is initially substituted by that of weak discontinuity capturing the displacement jump through its distribution (smearing) in a finite length: the oriented with of a finite element through which the fracture passes in a rather diffuse manner. In other words, the classical strain localization concept is brought to the approach, no strong discontinuity enrichment is made and standard constitutive models are used in a classi-cal strain localization setting. The standard mesh refinement recovers the strong discontinuity concept by taking the element width to zero. However, in order to remove the spurious mesh orientation dependence, in a second stage specific localization modes are injected, via mixed finite elements formulations, to the path of elements that are going to capture the cracks. These modes, as well as their injection time, are character-ized from the information provided by the SDA. Then, the obtained approach enjoys the benefits of the strong discontinuities one, at a complexity similar to the classical, and simpler, localization methods.

A 3D spherical rigid particle discrete element formulation based on the Discrete Element Method, DEM, which has been used to model concrete, is described. The model has been shown to give a good agreement in terms of peak strength, fracture localization and crack patterns for various load conditions. An explicit formulation for the reinforcement model which allows the model to be applied to reinforced concrete is presented. The reinforced rigid particle model is validated against known shear transfer tests due to dowel action. It is shown that the model is able to predict the influence of the dowel stiffness and the localized fracture in the vicinity of the initial crack.

The paper presents the 2 main methods for the determination of the in-situ state of stress in rock masses, developed by the LNEC, and refers to a practical case in which they were used, at an underground powerhouse in northern Portugal.

Abstract. The application of discrete element models based on rigid block formulations to theanalysis of masonry walls under horizontal out-of-plane loading is discussed. The problemsraised by the representation of an irregular fabric by a simplified block pattern are addressed.A test problem provides a comparison of various regular and random block patterns, showingtheir influence on the failure loads. An example of application of a rigid block model to a wallcapacity problem is presented.

This paper presents a study, which describes the influence of the dynamic behaviour of the intake tower on the experimental results obtained from operational modal analysis of Cabril dam. A 3D finite element model, for the intake tower of Cabril dam, was developed in order to study the dynamic behaviour of this neighbour structure of the dam. The main dynamic parameters computed and evaluated experimentally for this structure are summarized and discussed, to understand how they can interfere in the modal identification of Cabril dam.

The paper aims at presenting some achievements of joint statistics. The different statistical distributions used for the description of the main parameters of the joint sets (attitude, intensity, area, and aperture) are introduced. For the attitude, a comparison of the existing distribution functions is made. For the intensity, the formulas for its determination are given, the applicable statistical models are presented, and the Poisson model for the occurrence of the discontinuities in rocks is discussed in depth. The statistical models for the distribution of the volume of the blocks of a rock mass are introduced. For the area, again, the formulas for its determination are given, and the applicable statistical models presented. Finally, the distribution function for the aperture is referred to, as well as a note on the roughness and waviness.

A libertação do estado de tensão inicial nos maciços rochosos é a principal acção a considerar em escavações subterrâneas. Como o estado de tensão é muito influenciado pela topografia e por condições geológicas particulares de cada local, a sua determinação a partir de ensaios de campo necessita de uma abordagem particular que, tendo em conta a natureza pontual dos ensaios e a sua variabilidade, permita estimar o estado de tensão global no maciço. Apresentam-se os ensaios realizados para o projecto de uma central subterrânea, com destaque para a modelação numérica que permitiu integrar os resultados dos ensaios e determinar o estado de tensão global mais provável. Os modelos usados permitiram considerar no estado de tensão inicial o efeito erosivo do rio ao escavar o vale e da abertura das galerias onde os ensaios foram executados.

This paper deals with issues related to the evaluation of design parameters from some types of in situ and laboratory tests performed, during the geotechnical investigations, to characterize the most relevant geotechnical properties of the rock mass and also the actions. Some considerations regarding safety evaluation and limit state design are presented. Several types of Rock Mechanics tests are introduced in some detail, and a few application examples covering different Rock Engineering projects are presented to show how test results can be analysed to determine their values to be used in the design. The following examples are shown: triaxial tests to determine the strength of the intact rock, a joint network study for a dam foundation, laboratory shear tests to assess the shear strength of rock joints, flat jack and overcoring tests to estimate the in situ state of stress for an underground cavern, and borehole dilatometer and large flat jack tests to evaluate the deformability of a concrete dam foundation.

Masonry is a combination of units such as stones, bricks or blocks usually laid in a cementitious or limemortar. It is probably the oldest material used in construction and has proven to be both simple to buildand durable. Over the years, existing masonry constructions have inevitably suffered damage with time.Earthquakes, soil settlements, material degradation and lack of maintenance are the main reasons forthat. Careful and periodic assessment of such structures is necessary in order to evaluate their structuralcapacity and safety levels. However, performing the structural analysis of masonry construction is notan easy task. A review is presented of the main models based on the discrete element method and the available related numerical techniques that have been proposed for the analysis of masonry. The essential assumptions adopted by these models and numerical implementation issues are discussed. Differences between available models are illustrated by applications to various masonry problems including static and dynamic analysis of masonry arch bridges, walls, vaults, domes and ancient colonnades.This book is composed of 17 chapters authored/co-authored by 25 outstanding researchers from 11countries (Canada, France, Greece, Hungary, Iran,Italy, Mexico, Norway, Portugal, UK, USA), whichwere reviewed by 60 referees.

N/A

This chapter consists of two parts. First a general overview is given: the exact definition of the Discrete Element Method is provided, and the main types of existing models are introduced; then the theoretical fundaments of the most important approaches available today for masonry analysis are given. The second part of the chapter focuses on 3DEC, the most widespread commercial DEM code for masonry analysis. This second part begins with a theoretical overview. Application issues for masonry structures are then considered (block material behaviour, contact representation, structural elements like cables or bars etc.). Among the practical examples arches, barrel vaults, walls, coloumns and complete structures (like houses) are presented, including the issues of quasi-static versus dynamic analysis and of block fracture as well. The readers can find helpful advice on how to simulate their own problems most effectively.

N/A

Discrete element models have become a major tool for masonry analysis, allowing a suitable representation of its discontinuous nature and marked nonlinear behaviour. The chapter reviews the fundamental assumptions of the main DEM formulations available. It focuses more specifically on the widely used UDEC and 3DEC codes, explaining their relation with the other approaches, and the most relevant issues arising in their application to masonry. Applications to the main types of structures are reviewed, considering both static and dynamic analyses. Recent research providing comparisons of the numerical models with experimental and field data is particularly covered. The discussion of the different examples is aimed at clarifying the key capabilities of DEM and demonstrating the most effective ways of using it in a variety of masonry analysis situations.

This chapter is focused on the Portuguese experience on the developmentand exploration of systems for continuously monitoring dam vibrations, usingaccelerometers. The pioneer system for seismic and structural health monitoring(SSHM) installed in Cabril dam (the highest Portuguese arch dam: 132 m high) isdescribed in detail. The design of this system was the result of a long-term LNECresearch program, still ongoing. These monitoring systems should include softwaredeveloped to automatically perform the analysis of collected data, including theautomatic comparison with numerical results from 3DFE models. In view of thegood results obtained with the system in operation in Cabril dam since 2008, similarsystems have been installed in other large dams in Portugal, particularly in recentlybuilt dams. Finally, Baixo Sabor dam is presented as an example of a new Portuguesedam with a complete SSHM system, in operation since 2015. Themain experimentalresults obtained for both dams are shown, namely the evolution of natural frequenciesover time, mode shapes and the measured seismic response to earthquake events.

Para verificação do cumprimento das exigências de segurança das barragens,tanto relativamente à funcionalidade (cenários correntes), como em relação àsegurança ao colapso (cenários de rotura), é essencial utilizar modelos de simulação oude interpretação (no caso de obras existentes) do seu comportamento. Estes modelossão idealizações (simplificações) da realidade, tanto no que diz respeito às acções(modelos das acções), como no que respeita à componente estrutural (modelos estruturais).A solução destes modelos, através de adequados métodos de análise, quantificaparâmetros da resposta das estruturas que permitem avaliar a sua segurança, utilizandocritérios de segurança previamente definidos.A validação destes modelos pode ser feita, no caso de cenários correntes,através da sua comparação com os resultados da observação do comportamento debarragens. No caso de cenários de rotura, não existe muita experiência de comparaçãodestes resultados com situações reais: é pequeno o número de acidentes ocorrido e,mesmo nestes, o volume de informação é reduzido. Desta forma, o recurso a métodosexperimentais, para o estudo de cenários de rotura, tem uma importância acrescida.Assim o principal objectivo deste trabalho é contribuir para melhorar a capacidade deavaliação da segurança de barragens de betão envolvendo cenários de rotura pela fundação.Neste trabalho desenvolveram-se dois ensaios experimentais em modelosfísicos: um primeiro envolvendo uma barragem abóbada, para o estudo de cenários derotura da fundação para acções estáticas; e um segundo de uma barragem gravidadeensaiado em mesa sísmica, com vista a estudar a rotura pela fundação para acçõesdinâmicas.A previsão e interpretação dos ensaios foram efectuadas com modelosmatemáticos adequados, que permitiram a sua validação e posterior utilização no estudode outros cenários.

A brief description of the phenomena and problems associated with underground structures in hydraulic projects is presented, namely in what concerns the powerhouses, surge chambers and pressure tunnels and shafts. An analysis of the calculation methodologies followed in the structural design of this works is presented, as well as of the numerical models usually adopted, namely finite and boundary element models. Finally, some applications of computational mechanics are presented. First, a parameter study for underground powerhouses, with special emphasis in the definition of shapes, existence of several caverns,and in the tridimensional equilibria is described. Afterwards, examples of application to some hydraulic projects are illustrated: a large underground powerhouse in Mozambique, the Alto Lindoso hydroelectric power scheme in construction in the north of Portugal, and the Castelo do Bode tunnel, in operation, integrated in the water supply system to Lisbon.

O presente trabalho aborda a origem do estado de tensão interna dos maciços rochosos e apresenta os factores que o influenciam. Apresentam-se as técnicas e a aparelhagem correntemente utiiizada na determinação do estado de tensão, bem como a interpretação dos ensaios mais comuns, dando-se maior relevo aos ensaios com macacos planos pequenos (S F J) e com o extensómetro tridimensional (S T T),efectuados correntemente pelo LNEC.

Descreve-se sucintamente uma técnica para amostragem linear de descontinuidades em faces expostas de maciços rochosos e referem-se métodos para tratamento dos dados de campo com vista à definição defamílias de descontinuidades e caracterização de uma forma estatística da sua orientação, espaçamento e extensão. Apresentam-se os resultados obtidos para o caso de uma galeria subterrânea no aproveitamentohidroeléctrico do Alto Lindoso. Para as famílias de descontinuidade identificadas apresenta-se um estudo dos blocos tetraédricos que se poderão formar nas diferentes faces de rocha e analisam-se as suas condições de estabilidade.

N/A

Apresentam-se, neste relatório, modelos numéricos baseados no método dos elementos da fronteira, desenvolvidos no LNEC para análise de estruturas subterrâneas. Tecem-se algumas considerações sobre o método dos elementos de fronteira, descrevem-se os modelos desenvolvidos para equilíbrios planos e tridimensionais e apresentam-se alguns exemplos de aplicação.

N/A

N/A

N/A

This document presents a back analysis procedure for identification of the elastic parameters of transversely isotropic rock cores, containing an overcoring triaxial strain probe, from the strains measured during a biaxial test. A three-dimensional finite element model was developed to simulate the biaxial test on the overcored rock specimen and to compute the strains at the location of the strain gauges. Different optimisation algorithms were tested and the most suitable one was selected. The back analysis procedure was tested for identification of the five elastic parameters and the two orientation angles that characterise a transversely isotropic rock core. In spite that, with the developed methodology, convergence was reached and all those parameters could be identified, sensitivity analyses demonstrated that the results obtained were not stable and therefore they were not reliable. By introducing constrains based on common practice and previous experience, a stable and robust methodology was achieved: the three elastic parameters, E1, E2 and ?2, are reliably identified using the value of G2 calculated with Saint-Venant’s expression and a fixed value of ?1, while the orientation parameters are obtained from observation of overcored rock. Analysis of the results shows that application of this methodology represents an enormous step forward when compared with the traditional use of isotropy. Besides, the methodology is general and can also be used with other types of overcoring equipment. The five elastic parameters and the two orientation angles obtained can then be used, together with the overcoring strains, to compute the complete in situ state of stress.

This document presents the work developed by the PhD student Margarida Espada in the scope of the course unit Risk Analysis in Geotechnical Works of the Doctoral Program in Civil Engineering of the Faculty of Engineering of the University of Porto (FEUP). This work presents a reliability analysis, using the response surface method (RSM), for the case study of the left bank excavation slopes of the Baihetan arch dam foundation, which is one of the case studies of Margarida Espada's PhD thesis. The aim was to approximate an implicit limit state function by computing safety factors, using a discrete element model of the Baihetan left bank developed in 3DEC software, where the shear strength parameters of two important rock discontinuities were considered as random variables. The probability of failure was then obtained through an iterative process, using approximation methods. This work presents the 3DEC model developed to analyse the stability conditions in the Baihetan left bank and the results of the reliability analysis.

O problema da solução das equações de Navier Stokes é abordado neste trabalhopropondo-se uma técnica de representação das incógnitas deslocamentos e pressõesvariáveis no espaço e no tempo em série de Delta de Dirac e suas derivadas no tempocujos coeficientes se determinam com base na solução de sistemas de equaçõesdiferenciais definidos no espaço (problema com três variáveis espaciais independentesque pode ser estudado pelo método de Monge).

Neste relatório apresentam-se as equações básicas da elastodinâmica linear, as soluçõesfundamentais e o teorema da representação no domínio de tempo e de frequência.Para movimentos finitos de casca formula-se o teorema dos trabalhos virtuais e a suaimplementação pelo método dos elementos finitos.

Neste relatório apresentam-se os principais aspectos da modelação sísmica de barragensabóbada. Assim propõe-se: modelos numéricos de geração de sismos por uma falha; aforma como o sismo actua na barragem; modelos para estudar a influência da albufeirana vibração do conjunto; métodos para solução das equações do movimento; e novasformulações de elementos finitos para o problema estrutural.

In pressure tunnels and shafts, in addition to the mechanical effect that excavation has on the rock mass and the disturbance that it introduces in the water flow, also the contained water can interact from the mechanical and hydraulic points of view with the rock mass. Seepage in fractured rock masses is dominated by the flow that takes place through the joints, and this is highly d(;pendent on their properties, namely aperture and roughness. Joint aperture, in tum, i::; very sensitive to stress changes. Special attention must therefore be paid to the joints in hydromechanical analyses. In this report are presented some studies that were carried out for the hydromechanical characterization of joints of the Alto Lindoso rock mass and numerical simulations of the hydromechanical behaviour of two selected sections of the high pressure circuit.