Dans le présent travail, nous étudions l'impact sur un obstacle rigide d'un écoulement transitoire à surface libre de fluide viscoplastique. Cette étude est conduite numériquement à l'aide de la méthode SPH (Smoothed Particle Hydrodynamics), en y intégrant le modèle rhéologique de Herschel-Bulkley. Le code employé est adapté à nos besoins et validé sur des cas test classiques. Les caractéristiques locales de l'écoulement à proximité de l'obstacle sont analysées et deux régimes d'impact sont mis en évidence en fonction de la pente d'écoulement. L'étude des pressions exercées sur l'obstacle, conduite spatialement et temporellement en fonction de ces régimes d'impact, nous permet de mettre en évidence les rôles respectifs des composantes gravitationnelle et cinétique de la pression. Nos résultats sont comparés systématiquement à des résultats expérimentaux issus de travaux précédents et sont cohérents avec ces derniers. Une étude comparative de nos écoulements de fluide viscoplastique avec des écoulements de matériau granulaires de propriétés similaires nous conduit à mettre en évidence des caractéristiques communes entre les deux matériaux. / In this work, we study the impact of a transient free-surface flow of viscoplastic fluid on a rigid obstacle. This study is conducted numerically using the SPH (Smoothed Particle Hydrodynamics) method, and the Herschel-Bulkley rheological model. The SPH code is adapted to our needs and validated on classic benchmarks. The local characteristics of the flow near the obstacle are analysed and two impact regimes are highlighted depending on the slope angle. By studying of the pressure exerted on the obstacle, both spatially and temporally, with regards to these impact regimes, we evidence the respective roles of the gravitational and kinetic components of the pressure. Our results are systematically compared with experimental data from a previous work and are shown to be consistent. A comparative study conducted on both our viscoplastic flows and flows of granular material of similar properties highlights common characteristics of the two materials.

Dans le présent travail, nous étudions l'impact sur un obstacle rigide d'un écoulement transitoire à surface libre de fluide viscoplastique. Cette étude est conduite numériquement à l'aide de la méthode SPH (Smoothed Particle Hydrodynamics), en y intégrant le modèle rhéologique de Herschel-Bulkley. Le code employé est adapté à nos besoins et validé sur des cas test classiques. Les caractéristiques locales de l'écoulement à proximité de l'obstacle sont analysées et deux régimes d'impact sont mis en évidence en fonction de la pente d'écoulement. L'étude des pressions exercées sur l'obstacle, conduite spatialement et temporellement en fonction de ces régimes d'impact, nous permet de mettre en évidence les rôles respectifs des composantes gravitationnelle et cinétique de la pression. Nos résultats sont comparés systématiquement à des résultats expérimentaux issus de travaux précédents et sont cohérents avec ces derniers. Une étude comparative de nos écoulements de fluide viscoplastique avec des écoulements de matériau granulaires de propriétés similaires nous conduit à mettre en évidence des caractéristiques communes entre les deux matériaux. / In this work, we study the impact of a transient free-surface flow of viscoplastic fluid on a rigid obstacle. This study is conducted numerically using the SPH (Smoothed Particle Hydrodynamics) method, and the Herschel-Bulkley rheological model. The SPH code is adapted to our needs and validated on classic benchmarks. The local characteristics of the flow near the obstacle are analysed and two impact regimes are highlighted depending on the slope angle. By studying of the pressure exerted on the obstacle, both spatially and temporally, with regards to these impact regimes, we evidence the respective roles of the gravitational and kinetic components of the pressure. Our results are systematically compared with experimental data from a previous work and are shown to be consistent. A comparative study conducted on both our viscoplastic flows and flows of granular material of similar properties highlights common characteristics of the two materials.

Este trabajo fue realizado en la quebrada Huaycoloro el cual forma parte de la cuenca que lleva el mismo nombre ubicada en el distrito de Lurigancho-Chosica en la provincia de Lima, donde se busca explicar los criterios de modelación de flujo de detritos aplicando dos modelos matemáticos de análisis para este tipo de flujos. El primero de libre acceso KANAKO-2D y el segundo de acceso de paga FLO-2D, ambos modelos requieren datos de entrada en común como la topografía digital del terreno, valores de rugosidad del canal, hidrogramas para diferentes periodos de retorno, propiedades reológicas del fluido a dos fases (agua-sedimento) entre otros. Para esta investigación se realizó un Modelo de Elevación Digital (DEM por sus siglas en inglés) con resolución espacial de 5mx5m, donde al hidrograma liquido se le añade el porcentaje de concentración de sedimentos para la representación del flujo de escombros en los modelos correspondientes, los parámetros reológicos (viscosidad, porcentaje de concentración, esfuerzo de cedencia) son tomados de manera referencial comparando resultados de estudios hechos en la quebrada con muestras validadas recopiladas de la literatura. Para finalmente con el procesamiento de estos datos en ambos programas, obtener resultados como profundidad de flujo, velocidad, concentración de sedimentos, fuerza de impacto, etc. De los cuales, este artículo se centró en la comparación de la profundidad del flujo de escombros de los respectivos modelamientos con un evento real, donde para el modelo Kanako-2D, Flo-2D y la medida en campo se presenta una profundidad promedio de 2.20, 2.50 y 2.60 metros respectivamente, por lo que se concluye que el programa Flo-2D es más preciso en su modelamiento para esta situación en especial. / This work was carried out in the Huaycoloro stream which is part of the basin that bears the same name located in the district of Lurigancho-Chosica in the province of Lima, where it seeks to explain the criteria of debris flow modeling applying two mathematical models analysis for this type of flows. The first one with free access KANAKO-2D and the second one with paid access FLO-2D, both models require common input data such as digital terrain topography, channel roughness values, hydrographs for different return periods, rheological properties of the two-phase fluid (water-sediment) among others. For this research, a Digital Elevation Model (DEM) was carried out with spatial resolution of 5mx5m, where the percentage of sediment concentration is added to the liquid hydrograph to represent the debris flow in the corresponding models, the Rheological parameters (viscosity, concentration percentage, yield stress) are taken as a reference, comparing results of studies carried out in the stream with validated samples collected from the literature. Finally, with the processing of these data in both programs, obtain results such as flow depth, velocity, sediment concentration, impact force, etc. Of which, this article focused on the comparison of the depth of the debris flow of the respective modeling with a real event, where for the Kanako-2D, Flo-2D and the field measurement an average depth of 2.20 is presented , 2.50 and 2.60 meters respectively, so it is concluded that the Flo-2D program is more accurate in its modeling for this particular situation. / Trabajo de investigación

La presente tesis aplica un análisis de tendencias asociados a la variación de la energía de impacto y localización que tienen las barreras flexibles ante eventos de flujos de detritos. Esto para establecer y considerar la variabilidad anteriormente mencionada como un parámetro de diseño para barreras flexibles y, por consiguiente, desarrollar una propuesta de ingeniería eficaz que permita maximizar el comportamiento de las barreras flexibles y sea aplicable a diferentes proyectos que consideren como propuesta estructuras de contención dinámica para frenar el impacto de los flujos de detritos. Todo esto tomando en cuenta los parámetros geotécnicos e hidro morfológicos de la zona, con principal importancia en la pendiente y la precipitación como desencadenante de eventos de flujos de detritos. El estudio se desarrolló en la quebrada Primero de Mayo ubicado en el distrito de Lurigancho, departamento de Lima, Perú. Tomando en cuenta que es una quebrada sin presencia de estructuras de contención previas y con un crecimiento poblacional no controlado que aumenta el riesgo de pérdidas económicas y sociales en la zona. Las simulaciones del comportamiento de las barreras flexibles ante diferentes situaciones de energía y localización se desarrollaron en el software Rocfall para establecer el número de barreras necesario para contener el flujo de detritos de la zona y plantear cuatro diferentes iteraciones para el estudio. / The proposal of a probabilistic factor is presented to determine the percentage of failure that can occur in shallow foundations, in our case footings. In this article we will present a comparison between the safety factor (F.S), defined by the standard E050 Soils and Foundations, and the probability of failure. This method will be executed through the simulation of input data, such as specific weight, cohesion, angle of friction and load stress. For this simulation process, the Monte Carlo method will be applied. Once the soil resistance values ​​have been obtained, considering its normal distribution and the stress, the probabilities of failure will begin to be analyzed using the strength and stress density curves. Finally, we will analyze the relationship that will occur between the F.S and the probability of failure to observe the behavior when increasing the parameters for resistance. / Trabajo de investigación

El presente trabajo consiste en el modelamiento hidrológico e hidráulico de la quebrada Huaycoloro ubicado en el distrito de Lurigancho – Chosica con fin de proponer soluciones de mitigación estructural y no estructural que integren los modelos. Comenzando por un análisis de eventos de flujo de escombros en el Perú con marco teórico que aborda conceptos básicos relacionados al estudio de hidrología e hidráulica. Para el modelo hidrológico, se realizó la caracterización de la cuenca de la quebrada obteniendo sus parámetros principales, así también se realizó el análisis de precipitaciones y así realizar un modelamiento hidrológico mediante HEC-HMS en colaboración con ArcGIS y determinar caudales máximos con periodo de retorno de 100, 200 y 500 años. El modelo hidráulico fue trabajado con el software FLO-2D PRO y el software libre japonés KANAKO -2D, cuyos parámetros de entrada son hidrogramas calculados por el modelo hidrológico, la topografía digital del terreno, los parámetros físicos y reológicos del flujo, entre otros. Para luego obtener velocidades y profundidades máximas para cada uno de los periodos de retorno mencionados anteriormente cuyos resultados serán comparados entre sí en determinados puntos de control y así precisar cuál de ambos softwares es el mejor, para posteriormente modelar los escenarios incluidas las estructuras y poder determinar su óptimo pre-dimensionamieto y analizar su nivel de mitigación frente al impacto de los flujos de escombros. Finalmente, se realizó la comparación de las modelaciones con estructuras y se propuso de manera conceptual y básica medidas no estructurales, realizando su respectivo análisis, viabilidad y eficiencia para mitigar este tipo de eventos. / The present work consists of the hydrological and hydraulic modeling of the Huaycoloro stream located in the district of Lurigancho - Chosica in order to propose structural and non-structural mitigation solutions that integrate the models. Starting with an analysis of debris flow events in Peru with a theoretical framework that addresses basic concepts related to the study of hydrology and hydraulics. For the hydrological model, the characterization of the basin of the creek was carried out, obtaining its main parameters, as well as the analysis of precipitation and thus perform a hydrological modeling using HEC-HMS in collaboration with ArcGIS and determine maximum flows with return period of 100, 200 and 500 years. The hydraulic model was worked with the FLO-2D PRO software and the Japanese free software KANAKO -2D, whose input parameters are hydrographs calculated by the hydrological model, the digital topography of the terrain, the physical and rheological parameters of the flow, among others. To then obtain maximum velocities and depths for each of the aforementioned return periods, the results of which will be compared with each other at certain control points and thus determine which of the two softwares is the best to subsequently model the scenarios including the structures and be able to determine their optimum pre-dimensioning and analyze their level of mitigation against the impact of debris flows. Finally, a comparison of the modeling with structures was carried out and non-structural measures were proposed in a conceptual and basic manner, performing their respective analysis, feasibility and efficiency to mitigate this type of events. / Tesis

Для решения проблемы низкой точности, слабой адаптивности и плохой интерпретируемости существующих моделей прогнозирования опасности схода грязевых потоков предлагается новый метод прогнозирования. В качестве примера рассматриваются 159 точек бедствий в бассейне реки Нуцзян в Китае. Выбраны 15 факторов влияния, и с использованием метода комбинированного взвешивания тремя сторонами проводится оценка опасности точек риска схода грязевых потоков. Затем для прогнозирования опасности схода грязевых потоков используется модель CNN-BiGRU-Attention. Для оптимизации гиперпараметров применяется улучшенный алгоритм KOA (IKOA). В конечном итоге для повышения интерпретируемости результатов прогнозирования модели введена рамка SHAP. Результаты показывают, что по сравнению с 13 текущими наиболее часто используемыми моделями прогнозирования, модель IKOA-CNN-BiGRU-Attention демонстрирует наилучшие результаты прогнозирования. / To address the issues of low accuracy, poor adaptability, and weak interpretability in existing models for predicting debris flow hazards, a new prediction method is proposed. Using 159 disaster points in the Nujiang River Basin in China as a case study, 15 influencing factors are selected, and a tripartite combined weighting method is used to evaluate the risk levels of debris flow points. Subsequently, the CNN-BiGRU-Attention model is used to predict the hazard of debris flows. The improved KOA algorithm (IKOA) is employed for hyperparameter optimization. Finally, the SHAP framework is introduced to enhance the interpretability of the model's prediction results. The results show that compared to the 13 currently commonly used prediction models, the IKOA-CNN-BiGRU-Attention model exhibits the best predictive performance.

In environmental impact assessment (EIA), accident risk should be considered as one impact. However, few studies have compared assessed risk levels with documented accidents. There is thus a need for assessing accuracy of risk assessment methodologies. The aim of this study was to develop and examine a risk assessment guideline based on environmental assessment. The guideline examination was carried out along the railway and the road between Kiruna and Narvik, Northern Scandinavia. The examined hazards were avalanches, debris flows, rock falls, slush flows, torrents, moose and reindeer. Documented accidents were quantified into an accident level. For each hazard type, six sections/sites were assessed. The propensities of these hazards were analysed through environmental assessments, especially by utilising geomorphology. Then, the accident densities along the sections/sites were compared with assessed accident phase indices describing risk and hazard in both spatial and temporal dimensions. The average uncertainty factors for both the risk and the hazard indices were about 5. This implies a fairly good precision. Risk and hazard indices were also rated and compared to the accident density rates. Thereafter, railway and road information was split up as different proposal alternatives in an EIA context. The total risk for each proposal was compared with the total accident density. The risk assessment guideline proved to be cost-efficient, rapid and relatively reliable in comparing accident risk.

In environmental impact assessment (EIA), accident risk should be considered as one impact. However, few studies have compared assessed risk levels with documented accidents. There is thus a need for assessing accuracy of risk assessment methodologies. The aim of this study was to develop and examine a risk assessment guideline based on environmental assessment. The guideline examination was carried out along the railway and the road between Kiruna and Narvik, Northern Scandinavia. The examined hazards were avalanches, debris flows, rock falls, slush flows, torrents, moose and reindeer. Documented accidents were quantified into an accident level. For each hazard type, six sections/sites were assessed. The propensities of these hazards were analysed through environmental assessments, especially by utilising geomorphology. Then, the accident densities along the sections/sites were compared with assessed accident phase indices describing risk and hazard in both spatial and temporal dimensions. The average uncertainty factors for both the risk and the hazard indices were about 5. This implies a fairly good precision. Risk and hazard indices were also rated and compared to the accident density rates. Thereafter, railway and road information was split up as different proposal alternatives in an EIA context. The total risk for each proposal was compared with the total accident density. The risk assessment guideline proved to be cost-efficient, rapid and relatively reliable in comparing accident risk.

L’objectif de cette thèse est de développer un modèle numérique 3D afin d’étudier le phénomène de laves torrentielles ; écoulements visqueux fortement chargés en matière solide, surgissant en montagne lors d’orages violents. Aujourd’hui, la prévision des zones de vulnérabilité s’appuie sur des outils de calcul 0D, 1D ou 2D. Or ces outils ne peuvent représenter pleinement le comportement à surface libre des écoulements du fait de nombreuses approximations et hypothèses. Ainsi cette thèse met en œuvre un code numérique 3D pour étudier ces écoulements. Ce travail aboutit à la construction d’un modèle 3D à partir de données réelles de terrain. Plusieurs scénarios ont été étudiés et comparés à des résultats issus d’un modèle 2D. Les résultats mettent en évidence les apports non négligeables de la modélisation 3D : zones d’étalement et de dépôt, phénomènes d’encombrement, modélisation fine des écoulements dans les zones chenalisées. Préalablement, le modèle 3D a été validé en comparant les résultats numériques à des données expérimentales issues de la littérature, pour des typologies d’écoulement représentatives de celles observées sur des sites grandeur nature. / The objective of this thesis is to develop a 3D numerical model to assess debris flow. These viscous flows, heavily loaded with solid matter, form when heavy rain occurs in mountains. Today, forecasts of potentially impacted areas are based on 0D, 1D and 2D numerical tools. However, these tools cannot fully represent the free surface behaviour of debris flows due to the approximations and assumptions on which they are based. Thus, this work utilises a 3D numerical code to study this phenomenon. A specific model is built with real field data. Several flow scenarios are studied and compared with a 2D numerical model. The results highlight the significant benefits of a 3D approach by providing information on the fine representation of flow dynamics over the catchment area. The model also predicts the impact of debris flow (overflowing on a road bridge) and the zones of deposition and spreading. It highlights possible congestion phenomena and reproduces flows in the channels by fully accounting for parietal friction, capabilities not provided by 2D models. Prior to this application, the 3D model was evaluated with five sets of experimental data to validate its ability to represent viscoplastic flows. Different types of flows are studied and are representative of those observed on real sites when debris flow occur.

Debris flows, which occur in mountain settings worldwide, have been particularly damaging in the glaciated basins flanking the stratovolcanoes in the Cascade Range of the northwestern United States. This thesis contains two manuscripts that respectively investigate the (1) initiation processes of debris flows in these glaciated catchments, and (2) debris flow occurrence and its effect on valley bottoms over the last thousand years. In a 2006 storm, seven debris flows initiated from proglacial gullies of separate basins on the flanks of Mount Rainier. Gully heads at glacier termini and distributed collapse of gully walls imply that clear water was transformed to debris flow through progressive addition of sediment along gully lengths. In the first study, we analyze gully changes, reconstruct runoff conditions, and assess spatial distributions of debris flows to infer the processes and conditions necessary for debris flow initiation in glaciated catchments. Gully measurements suggest that sediment bulking requires steep gradients, abundant unstable material, and sufficient gully length. Reconstruction of runoff generated during the storm suggests that glaciers are important for generating the runoff necessary for debris flow initiation, particularly because infiltration capacities on glacial till covered surfaces well exceed measured rainfall rates. Runoff generation from glaciers and abundant loose debris at their termini explain why all debris flows in the storm initiated from proglacial areas. Proglacial areas that produced debris flows have steeper drainage networks with significantly higher elevations and lower drainage areas, suggesting that debris flows are associated with high elevation glaciers with relatively steep proglacial areas. This correlation reflects positive slope-elevation trends for the Mount Rainier volcano. An indirect effect of glacier change is thus the change in the distribution of ice-free slopes, which influence a basin’s debris flow potential. These findings have implications for projections of debris flow activity in basins experiencing glacier change. The second study uses a variety of dating techniques to reconstruct a chronology of debris flows in the Kautz Creek valley on the southwest flank of Mount Rainier (Washington). Dendrochronologic dating of growth disturbances combined with lichenometric techniques constrained five debris flow ages from 1712 to 1915 AD. We also estimated ages of three debris flows ranging in age from ca. 970 to 1661. Run-out distances served as a proxy for debris flow magnitude, and indicate that at least 11, 2, and 1 debris flow(s) have traveled at least 1, 3, and 5 km from the valley head, respectively since ca. 1650. Valley form reflects the frequency-magnitude relationship indicated by the chronology. In the upper, relatively steep valley, discrete debris flow snouts and secondary channels are abundant, suggesting a process of debris flow conveyance, channel plugging, and channel avulsion. The lower valley is characterized by relatively smooth surfaces, an absence of bouldery debris flow snouts, few secondary channels, and relatively old surface ages inferred from the presence of tephra layers. We infer that the lower valley is deposited on by relatively infrequent, large magnitude, low-yield strength debris flows like an event in 1947, which deposited wide, tabular lobes of debris outside of the main channel. Debris flows during the Little Ice Age (LIA) predominantly traveled no further than the upper valley. Stratigraphic evidence suggests that the main Kautz Creek channel was filled during the LIA, enhancing debris flow deposition on the valley surface and perhaps reducing run-out lengths. Diminished areas and gradients in front of glaciers during the LIA also likely contributed to decreased run-out lengths. These findings suggest that changes in debris flow source and depositional zones resulting from temperature and glacier cycles influence the magnitude and run-out distances of debris flows, and the dynamics of deposition in valley bottoms. / Graduation date: 2013