Submitted by Pedro Lemos Camarero null (camarero@folha.com.br) on 2016-05-23T22:57:01Z No. of bitstreams: 1 Análise de integridade Física de Barragens de Terra a partir da integração do método geofísico da eletrorresistividade com ensaios geotécnicos.pdf: 5497568 bytes, checksum: 87ea6f438628a89617555e0a6fb716dc (MD5) / Rejected by Ana Paula Grisoto (grisotoana@reitoria.unesp.br), reason: Solicitamos que realize uma nova submissão seguindo a orientação abaixo: O arquivo submetido não contém a Folha de aprovação. A versão submetida por você é considerada a versão final da dissertação/tese, portanto não poderá ocorrer qualquer alteração em seu conteúdo após a aprovação. Corrija esta informação e realize uma nova submissão contendo o arquivo correto. Agradecemos a compreensão. on 2016-05-25T19:40:42Z (GMT) / Submitted by Pedro Lemos Camarero null (camarero@folha.com.br) on 2016-05-26T17:51:45Z No. of bitstreams: 1 Análise de integridade Física de Barragens de Terra a partir da integração do método geofísico da eletrorresistividade com ensaios geotécnicos.pdf: 5603284 bytes, checksum: 57cc0506d6e9e96bea9a3ed5dc59a2e7 (MD5) / Approved for entry into archive by Ana Paula Grisoto (grisotoana@reitoria.unesp.br) on 2016-05-30T17:17:09Z (GMT) No. of bitstreams: 1 camarero_pl_me_rcla.pdf: 5603284 bytes, checksum: 57cc0506d6e9e96bea9a3ed5dc59a2e7 (MD5) / Made available in DSpace on 2016-05-30T17:17:09Z (GMT). No. of bitstreams: 1 camarero_pl_me_rcla.pdf: 5603284 bytes, checksum: 57cc0506d6e9e96bea9a3ed5dc59a2e7 (MD5) Previous issue date: 2016-04-15 / Dams are structures that dam rivers and streams for a variety of purposes. These structures, often need to be sturdy to withstand the force of the impoundment and the high values of accumulated water load. The constant maintenance of these structures is essential, since a possible accident can lead to damage of catastrophic proportions. This study presents an inexpensive alternative, simple and quick application for investigation of seepage of water in Earth dams, built with distinct loan material. The research methods used were: geotechnical tests as granulometric analysis, determination of the physical indexes of soil, permeability test with permeameter of Guelph and the geophysical method of electric resistivity, from electrical resistivity tomography. At each dam, were acquired three geophysical lines parallel to the longitudinal axis of the dam. The spacing between electrodes was 2m and the array used in the study was Wenner. The results are presented from geophysical images with 2D and 3D electrical resistivity values measured and modeled, where it was possible to identify areas of low relative values of electrical resistivity, interpreted as saturated areas and likely infiltration of the body of the dam. The quantitative data of the Geotechnical testing contribute to greater understanding of the internal water flow. / Barragens são estruturas que represam rios e córregos para diversas finalidades. Essas estruturas, geralmente precisam ser robustas para suportar a força do represamento e os altos valores de carga de água acumulados. A constante manutenção dessas estruturas é imprescindível, uma vez que um acidente eventual pode gerar danos de proporções catastróficas. Esse estudo apresenta uma alternativa barata, de simples e rápida aplicação para investigação de zonas de percolação de água em barragens de terra, construídas com material de empréstimo distintos. Os métodos de investigação utilizados foram: ensaios geotécnicos como análise granulométrica, determinação dos índices físicos do solo, ensaio de permeabilidade com o permeâmetro de Guelph e, o método geofísico da eletrorresistivade, a partir da técnica do imageamento elétrico. Em cada barragem, foram adquiridas três linhas geofísicas paralelas ao eixo longitudinal da barragem. O espaçamento entre eletrodos foi de 2m e o arranjo utilizado no estudo foi o Wenner. Os resultados são apresentados a partir de imagens geofísicas 2D e 3D com valores de resistividade elétrica medida e modelada, onde foi possível identificar zonas de baixos valores relativos de resistividade elétrica, interpretadas como zonas saturadas e de provável infiltração no corpo da barragem. Os dados quantitativos dos ensaios geotécnicos contribuem para maior entendimento do fluxo hídrico interno.

Embankment dams encounter several problems in terms of dam safety. One of those problems is called internal erosion. This phenomenon is induced by the movement of fine particles within the dam due to seepage forces. If the dam is not able to self-heal, the eroded zones will increase which will eventually cause the dam to fail. Thus standards have been created by Svensk Energi and summarized in the Swedish dam safety guideline RIDAS 2012. These standards are used as a basic in the risk analysis of existing dams and provide guidelines for proper design of future dams.A dam in Sweden has presented recurring incidents related to internal erosion within the core. The impact of this internal erosion is analysed in this thesis with the use of Finite Element Method/Analysis (FEM/A). FEA models simulate the in situ stresses in the dam and calculate the strength. It also enables the analysis of changing hydraulic conductivity and its effect on the overall effective strength due to changing pore pressure and seepage forces. The analysis using numerical methods was performed in the program PLAXIS2D and SEEP/W while limit equilibrium analysis was done in SLOPE/W.The calculation in PLAXIS2D was performed by using the Mohr-Coulomb constitutive model. The in situ stresses are initially calculated using gravity loading since this is the preferred method on an uneven terrain instead of a K0-calculation. Then, through a set of phases in the program, zones where erosion is assumed to have occurred are changed. These zones have a higher permeability and will thus affect the pore pressures in the dam following Darcy’s law with permeability through a set medium.The increased permeability is set to follow an increased void ratio due to loss of fine material in the core. How this increase of void ratio affects the permeability is investigated through using Ren et al’s (2016) proposed equation for calculating permeability with a set void ratio. Their equation, apart from the usually used Kozeny-Carman equation, considers both effective and ineffective void ratio where the ineffective void ratios refers to the volume of pores that is immobile when flow is considered.The increased flow in the eroded zones of the core did not seem to impact the strength of the dam in much regard. The phreatic surface and thus the pore pressure did not change enough to influence the overall effective strength of the dam. It raises the question if the stability of an earth-rock fill dam will be affected due to increased pore pressure at all due to its draining properties and if it would rather fail due to increased seepage forces.

The Lossen dam is an embankment dam in the Swedish river Ljusnan. The dam is founded on thick layers of stratified glaciofluvial sediments and till. Ever since construction, there have been problems with high pore pressures, large seepage flows and springs downstream of the right part of the dam. After the first filling of the reservoir, a large drainage trench was constructed downstream of the dam to lower pore pressures. Sinkholes and settlements downstream of the dam have occurred repeatedly over the lifetime of the dam, particularly in the area surrounding the large drainage trench. This study aims to investigate the causes of the sinkholes and assess the risks of internal erosion in the foundation and in the soil downstream of the dam. A model of the groundwater flow has been created in Visual MODFLOW. Calculations for assessing the soils susceptibility for contact erosion and suffusion have been performed, using soil gradation curves (both from the 1960s and from new samples). The results from the erosion calculations show that the probability of contact erosion being the sole cause of the internal erosion is minor. However, a majority of the soil samples tested are internally unstable (susceptible for suffusion). The possibility of backward erosion piping can not be dismissed but would need further investigation in order to be properly assessed. Due to the very heterogeneous structure of the soil in the area downstream of the dam, it is hard to predict the extent of the eroding soil layers and, thus, the future development of the erosion. A conclusion is that the soil downstream of the dam is subject to internal erosion, which is also the cause of the observed sinkholes and settlements. The internal erosion is probably not a threat to dam safety at the moment, but can possibly evolve as a problem in the future. Therefore close monitoring of the changes in seepage patterns and sediment transport is recommended. Possible solutions to stop the erosion include extending the drainage system with more pressure relief wells, and placing a filter blanket in the seepage exit area. / Lossendammen är en jorddamm i älven Ljusnan.  Dammen är byggd ovanpå tjocka lager av morän och isälvssediment. Sedan första dämningen 1962 har problem med höga portryck, stora läckageflöden och källsprång nedströms dammen uppkommit. Efter första dämningen grävdes ett stort dränagedike (kallat Slits 19) ut nedströms om dammen, för att sänka portrycken i området. Sjunkhål och sättningar har uppkommit återkommande under dammens livstid, speciellt i området kring slitsen. Målet med denna studie är att undersöka orsaken till de uppkomna sjunkhålen och sättningarna, samt att bedöma risken för fortsatt intern erosion i dammens undergrund och i jorden nedströms dammen. En modell över grundvattenflödet i området har skapats i Visual MODFLOW. Jordens fallenhet för kontakterosion och suffusion har beräknats med hjälp av kornfördelningskurvor från jordprover (majoriteten av vilka är från 60-talet, samt några nya). Resultaten från beräkningarna visar att risken för att kontakterosion initieras är liten. Däremot visar en majoritet av de testade kornfördelningskurvorna att jorden är internt instabil (har fallenhet för suffusion). Sannolikheten för att bakåtskridande erosion initieras kan inte uteslutas baserat på det underlag som förelegat, men en pålitlig bedömning skulle kräva ytterligare undersökningar. Eftersom jorden under och nedströms dammen är väldigt heterogen, och de olika lagrens storlek och utbredning är okänd, är det svårt att förutspå hur erosionen kommer att utvecklas i framtiden. Slutsatsen är att jorden nedströms dammen utsätts för fortlöpande inre erosion. Inre erosion är också orsaken till de uppkomna sjunkhålen och sättningarna. I nuläget är erosionen ingen risk för dammsäkerheten, men kan komma att bli det i framtiden. Noggrann övervakning av eventuella förändringar i läckagemönstren och sedimenttransport är viktig. Bland möjliga lösningar för att avstanna erosionen finns utökning av dränagesystemet med nya filterbrunnar, samt att lägga ett filtrerande skyddslager över området där källsprången sker.

There are still uncertainties in the safety of existing embankment dams. For instance, the majority of embankment dams in Sweden were built between 1950s and 1970s, designed and constructed to standards that might be unacceptable nowadays. Particularly, Vattenfall’s records stated that 40% owned embankments dams developed sinkholes (Nilsson, 1999). Moreover, internal erosion and its failure mechanisms of initiation and development are still not fully understood (Bowles et al., 2013). Also, internal threats are difficult to detect and interpret even using new instrumentation techniques. The aim of this Master Thesis is to identify failure mechanisms of embankment dams prior to breaching and hence, verify the reliability of a risk analysis after the breaching of the dam. The methodology consisted of monitoring an embankment dam prone to fail by internal erosion mechanisms. Finally the results were modelled using FEM to identify the risk of internal erosion prior to breaching.

Seepage-related erosion is one of the predominant mechanisms responsible for incidents and failures of dams and levees. Current geotechnical engineering practice consists of comparing expected exit gradients with the critical gradient of the soil at the seepage exit point. The critical gradient is generally considered as the ratio of soil buoyant unit weight and the unit weight of water, suggesting that the critical gradient only depends on the void ratio and specific gravity of the solids. However, in the field and in research, it has been observed that piping can initiate at average gradients much lower than unity due to concentrations in flow and non-vertical exit faces. Therefore, there is a need for deeper understanding of the granular scale mechanisms of the piping erosion process. This thesis presents the results of a laboratory study to assess the effects that soil properties and exit face configurations have on the potential for initiation of piping and the piping mechanisms. By using a laboratory device designed and constructed specifically for this study, the critical gradients needed to initiate piping in a variety of sandy soils were measured to assess the effects that parameters such as gradation, grain size, and grain shape have on the critical gradients. The tests are also used to observe the grain scale mechanisms of piping erosion initiation. The ultimate goal of the study is to develop an empirical, but mechanism-based, grain-scale model that can take into account the effects of converging flows, non-horizontal exit faces, and soil properties while assessing the potential for piping erosion to occur.

L'érosion de contact est un type d'érosion interne qui se développe à l'interface entre deux couches de matériaux de granulométries différentes. Les particules d'un sol fin (sable, limon, argile…) sont détachées par l'écoulement et entraînées à travers les pores du sol grossier au contact (gravier…). Bien que l'on suspecte sa présence dans de nombreux ouvrages, ce processus d'érosion a été peu étudié jusqu'à présent. Aussi ce travail de thèse s'est attaché à mieux le comprendre pour parvenir à le modéliser. A l'échelle du pore, l'écoulement à l'interface entre deux milieux poreux a été caractérisé, grâce à un dispositif expérimental développé au Cemagref d'Aix-en-Provence qui combine fluorescence induite par laser, méthode PIV et milieu iso-indice. L'importance de la variabilité des sollicitations hydrauliques a ainsi été soulignée. A l'échelle de l'échantillon, des essais sur sols réels et sols reconstitués ont été menés au LTHE, afin d'identifier les phénomènes en jeu. Ont également été obtenues sur ce dispositif des mesures du taux d'érosion en fonction de l'intensité de l'écoulement, pour différents types de sols fins et de sols grossiers. Une modélisation stochastique de ces essais a ensuite été proposée. Enfin, des essais à grande échelle ont été conduits au laboratoire de la Compagnie Nationale du Rhône pour étudier l'éventuelle influence d'effets d'échelle ainsi que les conséquences de cette érosion de contact sur le comportement global et l'intégrité d'un ouvrage. Il a ainsi été mis en évidence la possibilité qu'une érosion de conduit soit initiée par érosion de contact. Cette thèse a été réalisée dans le cadre du projet ERINOH (ERosion INterne dans les Ouvrages Hydrauliques), en convention CIFRE avec le Centre d'Ingénierie Hydraulique d'EDF. / Contact erosion is a type of internal erosion which occurs at the interface between two soil layers of different particle sizes. Particles from the fine soil (sand, silt, clay…) are pulled and driven by the flow through the pores of the surrounding coarse soil (gravel ...). Although its presence is suspected in many earthfill embankments, this process has been little studied so far. The aim of this thesis was to better understand and, possibly, to model contact erosion. At pore scale, the flow at the interface between two porous media was characterized thanks to an experimental device, developed in Cemagref, combining Planar Laser Induced Fluorescence, PIV method and Refractive Index Matching. The spatial variability of flow shear stress has been emphasized. At sample scale, several tests on real and reconstituted soils were carried out in LTHE, in order to identify the phenomena involved in contact erosion. Erosion rates as a function of flow magnitude have also been measured for different types of fine and coarse soils. Then, a stochastic model was proposed to account for these experiments. Finally, large scale tests were conducted in the laboratory of Compagnie National du Rhône, to study the influence of scale effects as well as the consequences of contact erosion on global behavior and integrity of the embankment structure. The possibility of piping initiation by contact erosion was underlined. This thesis is part of the project ERINOH project and was funded by EDF-CIH.

L’érosion interne est la cause principale de ruptures des ouvrages hydrauliques en terre, tels que les digues et les barrages. Les conséquences de telles ruptures sont conséquentes et couteuses. L’objectif de cette thèse est de mieux comprendre une catégorie particulière de processus d’érosion, la suffusion, en en caractérisant l’initiation et l’évolution. Cette recherche, à dominante expérimentale, s'appuie sur une modélisation physique unidimensionnelle. Les essais sont réalisés dans une conduite en plexiglas de 18 cm de diamètre intérieur, équipée de capteurs de pression, d’un débitmètre et d’un turbidimètre. Un dispositif de collecte des particules érodées est installé à l’aval de la conduite pour quantifier les particules érodées à des intervalles de temps choisis. Dans cette étude, on a réalisé une série d’essais sur des matériaux pulvérulents. Dans ce type de matériaux, l’érosion se manifeste par suffusion. Elle correspond à l’arrachement et au transport des particules fines à travers l’espace poral des particules grossières. Dans un premier temps, un protocole de reconstitution de matériaux pulvérulents est élaboré. Le principe consiste à réaliser des mélanges humides avec une teneur en eau relative à la masse des particules fines. Pour surmonter les problèmes rencontrés dans la réalisation des essais, la conduite est disposée verticalement, et les matériaux testés sont reconstitués à partir de mélanges de particules sableuses. Une étude paramétrique est menée sur l’influence de certains paramètres, tels que la granulométrie, la pression d’écoulement, la nature et la teneur en particules argileuses ainsi que la densité de mise en place. Les résultats obtenus montrent que l’érodabilité des matériaux dépend de la forme des courbes granulométrique. En effet, l’érosion est plus accentuée pour les matériaux contenant moins de particules fines. Elle est également plus forte quand la courbe granulométrique est discontinue. L’ajout de particules argileuses augmente la résistance à la suffusion des matériaux testés. Trois catégories de matériaux argileux ou fins sont testées : une illite (argile verte de Velay), une kaolinite (Speswhite), et du sable broyé (C10). Les résultats montrent que les matériaux contenant de l’illite sont plus résistants à la suffusion que les matériaux contenant de la kaolinite, tandis que les matériaux contenant du sable broyé sont les moins résistants. Cette résistance à la suffusion augmente avec la teneur en particules argileuses. L’érosion des matériaux dépend aussi de la densité initiale des matériaux testés. En effet, pour la même pression d’entrée, la quantité des particules érodées diminue en augmentant la densité de mise en place / Internal erosion is the main cause of failure of hydraulic structures such dykes and dams. The consequences of such failures are substantial and costly. The objective of this thesis is to better understand one of the erosion phenomena, the suffusion, and to characterize the initiation and the evolution of this phenomenon. A physical modeling approach is used for this study. Tests are carried out in a Plexiglas pipe of 18 cm inner diameter, equipped with pressure sensors, flowmeter and turbidimeter. A device for collecting the eroded particles, at selected time intervals, is installed in the downstream part of the device. In this study we performed a series of tests on coarse cohesionless soil. In this type of material, erosion is manifested by suffusion. It corresponds to the detachment and transport of fine particles through the pore space of the coarse particles. First, a cohesionless soil reconstitution protocol is developed. The idea is to make wet mixtures with water content depending on the fine particle content. To overcome some problems encountered in carrying out the tests, the device is arranged vertically, and the materials tested are reconstituted from mixtures of sand particles. A parametric study on the influence of several parameters such as particle size distribution, flow pressure, nature and content of clay particles and initial density are conducted. The results show that the erodibility of the material depends on the shape of the particle size distribution. Indeed erosion is more pronounced for materials containing less fine particles. This erodability is more pronounced when the curve is gap-graded. Erodibility of the tested soils increases with the hydraulic load. The addition of clay particles increases the resistance to suffusion of the soils. Three categories of clayed or fine soils are tested: illite (Argile Verte de Velay), kaolinite (Speswhite), and crushed sand (C10). The results show that materials containing illite are more resistant to suffusion than materials containing kaolinite, whereas materials containing crushed sand are less resistant. This resistance to suffusion increases with clay particle content. The erosion of materials also depends on the initial density of the soil tested. Indeed, for the same applied pressure, the quantity of eroded particles decreases with increasing the initial density

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第20684号 / 工博第4381号 / 新制||工||1681(附属図書館) / 京都大学大学院工学研究科都市社会工学専攻 / (主査)教授 澤田 純男, 教授 清野 純史, 准教授 後藤 浩之 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

La plupart des digues sont constituées de matériaux granulaires compactés. Elles sont ainsi perméables et constamment soumises à des écoulements d’eau dans leur volume. Dans certaines conditions, ces écoulements peuvent altérer leur microstructure par érosion interne et générer des instabilités mécaniques responsables de ruptures inopinées lors de crues. Cette thèse s’intéresse à l’analyse multi-échelle des instabilités mécaniques dans les matériaux granulaires soumis à l'érosion interne. Dans ce travail, le comportement mécanique de ces matériaux est simulé en 3D à l’échelle de volumes élémentaires représentatifs, et ce, pour différents états de contraintes et gradients hydrauliques. Grâce à l’utilisation du critère du travail du second ordre et d’outils micromécaniques, leur stabilité est analysée avant et après l’application d’un écoulement interne. Il est établi que l’origine micro-inertielle des instabilités observées provient du déconfinement et de la flexion des chaînes de force ainsi que des déformations plastiques importantes résultant de leur effondrement. Par leur capacité à enrayer rapidement le développement de telles déformations plastiques, il est montré que les particules libres contribuent à assurer la stabilité mécanique des matériaux granulaires. Ce résultat est fondamental pour analyser les conséquences de l’érosion interne en termes de stabilité mécanique car les particules libres sont facilement transportables sous l’action d’un écoulement interne. Selon si elles sont colmatées ou érodées, un écoulement interne aura un effet stabilisateur ou déstabilisateur vis-à-vis du comportement mécanique des matériaux granulaires soumis à l’érosion interne / Dikes are most of the time built of compacted granular materials that are permeable and continuously subjected to internal fluid flows. In some cases, microstructure modifications resulting from internal erosion generate mechanical instability that will lead to unexpected failures in case of serious flooding. This thesis focuses on multi-scale analysis of mechanical instability in granular materials subjected to internal erosion. In this work, the mechanical behavior of such materials is simulated in three dimensions at the scale of representative elementary volumes subjected to different stress states and hydraulic gradients. Thanks to the use of the second order work criterion and micromechanical tools, the mechanical stability of these materials is tested before and after internal erosion. It is established that the micro-inertial origin of the observed instabilities is linked to force chain deconfinement and bending as well as to the development of large plastic strains resulting from force chain collapse. By preventing the development of such plastic strains, it is shown that rattlers contribute to ensure the mechanical stability of granular materials. This key finding is of a particular significance in relation with internal erosion as rattlers can be easily transported under the action of an internal fluid flow. Depending on whether they get clogged or eroded, an internal fluid flow has thus either a stabilizing or a destabilizing effect on the mechanical behavior of granular materials subjected to internal erosion

L'érosion interne est le résultat de l'arrachement des particules d'un sol sous l'action d'un écoulement hydraulique. Ce phénomène est à l'origine des ruptures hydrauliques des barrages et des digues. Sur le plan mondial, 46% des désordres observés, sur les ouvrages en terre, ont pour origine l'érosion interne. En France, 70 cas critiques ont déjà été détectés. Malheureusement, à cause du phénomène d'érosion lui-même, la prévision du risque reste délicate et les autorités ont du mal à mettre en place des plans d'urgence. Compte tenu du risque potentiel, la compréhension des phénomènes à l'origine de l'érosion interne apparaît comme un enjeu scientifique majeur. L'objectif du travail de thèse est de mettre au point une méthode de diagnostic de l'érosion interne dans les digues et autres ouvrages en terre. Nous cherchons à établir un protocole qui nous guidera sur les essais à faire et qui nous aidera à interpréter les résultats. Cette méthode permettra d'expertiser les risques d'érosion interne. (i) Dans un premier temps, les principaux résultats des reconnaissances géomécaniques permettent d'identifier des couches géotechniques de sol de caractéristiques homogènes dans lesquelles sont réalisés les essais pressiométriques. L'analyse cluster des résultats des essais pressiométriques permet de déterminer le type, la nature de sol, ainsi que les caractéristiques géotechniques moyennes. (ii) Les résultats obtenus servent ensuite à déterminer si le sol est instable et quel est son seuil de sensibilité à l'aide du logiciel expert. Ce logiciel permet de calculer les critères d'érosion interne et de qualifier la sensibilité du sol à l'érosion en déterminant son seuil érosif. (iii) Un nouveau dispositif expérimental appelé l'Essai d'Erosion Transverse (Cross Erosion Test - CET) permet de caractériser expérimentalement le sol par rapport au risque d'érosion interne. Cet essai consiste à injecter de l'eau dans un forage, puis à pomper cette eau dans un forage parallèle au premier. L'eau extraite est chargée en particules produites par l'érosion interne du sol. L'érosion est estimée en se basant sur la mesure de l'évolution de la masse des particules recueillies au cours du temps. Les résultats montrent que cette expérience permet de caractériser l'érosion interne dans un sol spécifique. L'avantage de cette technique c'est qu'elle est transposable in situ et qu'elle peut être utilisée pour prédire le risque de suffusion dans les barrages et les digues. (iv) La validation de l'expérience est faite par un modèle numérique 3D réalisé avec le logiciel Comsol Multiphysics 4.3b. Ce modèle permet de montrer que sous les conditions de perméabilité observées, le gradient hydraulique se concentre autour des points d'injection et de pompage. Une modélisation de base en 2D est développée. Cette première approche permet de décrire le phénomène d'érosion et le transport des particules fines dans le milieu poreux du sol. / Internal erosion is the displacement of the fine particles of a soil under the action of an internal flow. This mechanism could be the origin of the damage on embankments and earth dams. In the word, 46% of failure observed on earthwork, are caused by internal erosion. In France, 70 critical cases have been detected. Unfortunately, due to internal erosion, the prediction of risk remains difficult. The understanding of this phenomenon of internal erosion appears as a major scientific challenge. The aim of this thesis is to develop a diagnostic method of internal erosion in dams and other earthworks. We seek to establish a protocol that will guide us to the test and help us to interpret the results. This method allows concluding about the risk of internal erosion appearance. (i) Initially, the main results of geotechnical survey allow identifying the layers of homogeneous characteristics of soil which are measured by the pressuremeter test. The results of cluster analysis of pressuremeter tests allow determining the localisation, the nature and the mean geotechnical characteristics of the soil. (ii) The results are used to determine the threshold of sensibility to internal erosion with the help of expert software. This software allows to calculate the criteria of internal erosion and to describe the sensitivity of the soil to internal erosion. (iii) A new experimental device is developed in our laboratory. This experiment called “Cross Erosion Test" (CET) allows determining the experimental resistance of the soil to the risk of internal erosion. The test consists of the injection, in a first drilling, of clear water and of the recovery, in another drilling, of water charged with eroded particles. For different initial state of the soils, it is possible to measure and to characterize the internal erosion by visualisation of the water flow and the measurement of the weight of the extracted eroded particles. The results show that this experience allows characterizing the internal erosion in a specific soil. The advantage of this technique is that it can be used in situ to predict the risk of suffusion in dams and dikes. (iv) A validation of experiences, with a 3D finite element method is carried out with the help of the Comsol Multiphysics 4.3b software. This model shows that under experiment hydraulic conditions, the hydraulic gradient is concentrated around the injection and the pumping. A finite element 2D model is developed to simulate erosion process. This approach describes the phenomenon of internal erosion and transport of fines particles in the porous medium of soils tested.