De nos jours, le respect de l'environnement est une obligation dans le cadre de travaux maritimes. Les caissons en béton verticaux à plaques poreuses sont souvent la solution technique privilégiée pour agrandir les ports existants et / ou améliorer l'agitation des bassins. L'empreinte au sol est réduite et la demande en matériaux de carrière est moins importante par rapport aux digues en enrochement classiques. Récemment, des systèmes alternatifs aux caissons en béton verticaux ont été conçus. Les structures amortisseuses de la houle à enrochements verticalisés sont constituées d'une structure métallique entourant des blocs. Ce type de structure offre des avantages environnementaux, une perméabilité aux courants et une bonne performance hydrodynamique. Ce travail de thèse visait à développer un nouveau code 3D-BEM facile à utiliser et intégrant les écoulements en milieux poreux. Des géométries innovantes sont testées comme des gabions espacés avec une chambre d’expansion ou un mélange de plaques poreuses et de milieux poreux.La description des écoulements dans un milieu poreux est une question complexe. Navier-Stokes moyenné au sens de Reynolds (RANS) est le processus mathématique communément utilisé pour modéliser les écoulements en milieux poreux. Ce dernier est considéré comme un milieu continu homogène. L'équation bien connue de Forchheimer étendue décrit les forces volumiques appliquées à l'écoulement par un milieu poreux à travers des coefficients de résistance et d'inertie. Ces recherches ont été l’occasion d’examiner ces coefficients, notamment celui d'inertie mal connu dans le cas d’un milieu poreux. Il joue un rôle majeur pour les écoulements à très faible KC se produisant dans les couches internes des digues poreuses. Une analyse de la littérature sur les écoulements de milieux poreux a été entreprise. Le code numérique appelé Diffra3D a été développé. Il a ensuite été utilisé pour rechercher des coefficients de résistance de milieux poreux à l'aide de données provenant de trois campagnes expérimentales : un test de sloshing sur hexapode et deux études classiques de transmission/réflexion en canal à houle. Celles-ci ont également été l'occasion de tester et de calibrer le code. De nouvelles géométries de structures poreuses ont ensuite été testées expérimentalement et numériquement.Deux nouvelles valeurs de coefficients de résistance d'un milieu poreux sont proposées pour les écoulements à très faibles KC. Certaines caractéristiques intéressantes concernant le coefficient d'inertie CM d'un milieu poreux sont également développées. Ce sujet de recherche mériterait encore d’être approfondi pour tenter de trouver une ou plusieurs loi(s) empirique(s) décrivant l’évolution du coefficient d'inertie dans ce régime spécifique d’écoulement. Le code Diffra3D fonctionne de façon satisfaisante pour modéliser les écoulements en milieu poreux. Il est cependant limité aux vagues à faibles cambrures. Tout l'enjeu est de bien caractériser le milieu poreux étudié. L'utilisation de Diffra3D a permis de montrer que les structures poreuses verticales innovantes, telles que les gabions espacés, ont entièrement leur place en tant que digues amortisseuses de la houle respectueuses de l'environnement. Elles pourraient être couramment utilisées dans le futur. / Nowadays the respect of the environment is an obligation in maritime works. Vertical concrete caissons with porous plates are often the number one technical solution to enlarge existing ports and/or to improve the agitation of the basins. The footprint is reduced and the demand in quarry materials is less important compared to classical riprap breakwaters. Recently, alternative systems to vertical concrete caissons have been designed. Vertical riprap breakwaters are made of a metal framework enclosing blocks. This kind of structure offers environmental benefits, permeability for currents and a good hydrodynamic performance. This thesis work looks at developing a new 3D-BEM code that is easy to use and integrates porous media. Innovative geometries are tested like spaced gabions with damping chamber or a mix of porous plates and porous media.Describing flows in porous media is an complex issue. Volume averaging method is the common mathematical process used to model porous media flows without drawing every grain of a porous medium. The well-known extended Forchheimer equation describes the volumetric forces applied to the flow by a porous medium through resistance and inertial coefficients. These researches were the occasion to look into this coefficients, especially the one of poorly understood inertia in the case of a porous medium. It plays a major role in very low-KC flows currently occurring in porous breakwaters apart from armour layer.First, a literature review on porous media flows was undertaken. In parallel, the numerical code called Diffra3D was produced. It was then used to look for resistance coefficients of porous media through data coming from three experimental campaigns : one sloshing test on hexapode and two classical reflection-transmission studies in a wave tank. These campaigns were also the occasion to test and calibrate the code. New geometries of porous structures were then tested experimentally and numerically. Two new values of resistance coefficients of a porous medium are proposed. Some interesting features concerning the inertia coefficient CM of a porous medium are also developed. In simulations, we observe that the hydrodynamic behaviour of porous structures in low-KC flows is very sensitive to the coefficient of inertia. This research topic would still deserve further studies in order to find empirical law(s) for the inertia coefficient of a porous medium. The code Diffra3D performs well to model porous media flows. However, it is limited to waves with low steepness. The challenge is to properly characterise the porous medium. This research has shown that innovative porous structures like spaced gabions have proven their place as environmentally-friendly damping breakwaters. They may be commonly used in the future.

Thesis (MScEng)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: The Strand is located within False Bay and frequently encounters large quantities of wave overtopping over the coastal defences. This results in the damage of property and infrastructure and causes streets to be flooded. Physical modelling tests were done by a consultancy firm to determine a solution by making use of a recurve structure at the back of the beach. This study is an expansion of the previous physical modelling that was done and focuses on several factors that were not tested by the consultancy firm that could have an influence on the overtopping rate and provide additional information on the recurve design. These tests were: (i) the effectiveness of the proposed recurve wall design in reducing overtopping was compared to a vertical wall; (ii) the influence that modifications to the beach profile such as the beach slope, beach width and beach level have on the overtopping rate; and (iii) the sensitivity of overtopping to changes in wave period was tested. The information gathered from these tests was used to propose a possible solution for the Strand. Numerical modelling was done with Delft3D-Wave to determine the wave height at the back of the beach using a nested grid. The waves at the Strand are depth limited and therefore very sensitive to changes in water level. By altering the beach level the model showed how the significant wave height at the back of the beach changes. First estimates of overtopping were determined using the relevant empirical calculations from the EuroTop Manual 2007 for a vertical seawall. No estimate could be made for the recurve wall since it did not fall in the valid range of the equations. From the physical modelling it was found that the overtopping reduced significantly from a vertical to a recurve seawall by a factor of about 50% depending on the wall height. All the prediction methods tested proved to be accurate in estimating the overtopping when the ratio of freeboard to significant wave height was ≤1.83. For non-breaking wave conditions the beach profiles that were gently sloped (1:50) and wide produced more overtopping than the beach profiles that were steep (1:10) and narrow. Increasing the beach level only decreased the overtopping if the water depth was shallow enough to cause the waves to break before they reached the back of the beach. Overtopping was found to increase with longer wave periods until the wave period became too long and the waves broke offshore which resulted in the overtopping decreasing. Possible solutions to overtopping were proposed based on two beach levels and the implementation of a recurve seawall. Revised crest levels for the wall were made along the length of the beach for both the 1:20 and 1:100 year water levels. / AFRIKAANSE OPSOMMING: Die Strand is in Valsbaai geleë en ervaar dikwels baie oorspoeling deur golwe bo-oor kusverdedigingswerke. Dit lei tot skade aan eiendom sowel as aan infrastruktuur en veroorsaak dat strate oorstroom. Fisiese modelleringstoetse is deur ’n konsultasiefirma gedoen om ’n oplossing te probeer vind deur van ’n teruggebuigde struktuur aan die agterkant van die strand gebruik te maak. Hierdie studie is ’n uitbreiding van die vorige fisiese modellering wat gedoen is en fokus op verskeie faktore wat nie deur die konsultasiefirma getoets is nie. Dit kan moontlik ’n invloed op die oorspoelingstempo hê en verskaf bykomende inligting oor die ontwerp van die terugbuiging. Hierdie toetse is: (i) die doeltreffendheid van die voorgestelde ontwerp van die teruggebuigde strandmuur in die vermindering van oorspoeling word vergelyk met ’n vertikale muur; (ii) die invloed wat veranderinge aan die strandprofiel soos die helling van die strand, die wydte van die strand asook die strandvlak op die oorspoelingstempo het; en (iii) die sensitiwiteit van oorspoeling op veranderinge in golfperiode is getoets. Die inligting wat uit hierdie toetse verkry is, word gebruik om ’n moontlike oplossing vir die Strand voor te stel. Numeriese modellering is met Delft3D-Wave gedoen om die golfhoogte aan die agterkant van die strand vas te stel deur van ’n genestelde ruitenet gebruik te maak. Die golwe by die Strand word deur diepte beperk en is dus baie sensitief vir veranderinge in die watervlak. Deur die strandvlakke te verander het die model getoon hoe die betekenisvolle golfhoogte aan die agterkant van die strand verander. Die eerste beramings van oorspoeling is bepaal deur van die relevante empiriese berekenings uit die EuroTop-handleiding 2007 vir ’n vertikale strandmuur gebruik te maak. Daar kon geen beraming vir die teruggebuigde muur gemaak word nie aangesien dit nie binne die geldige bereik van die vergelykings val nie. Uit die fisiese modellering is daar vasgestel dat oorspoeling noemenswaardig met ’n gemiddeld van ongeveer 50% verminder is, afhangend van die muurhoogte. Al die voorspellingmetodes wat getoets is was akkuraat in die beraming van die oorspoeling wanneer die verhouding van vryboord tot betekenisvolle golfhoogte ≤1.83 was. Vir nie-brekende golftoestande het strandprofiele met ’n lae helling (1:50) en wat wyd was meer oorspoeling tot gevolg gehad as strandprofiele wat steil (1:10) en nou was. ’n Verhoging in die strand se vlakke het die oorspoeling slegs verminder indien die diepte van die water vlak genoeg was om die golwe te laat breek voordat hulle die agterkant van die strand bereik het. Oorspoeling is gevind om te vermeeder met verlengde golflengte tot dat die golflengte só lank geword het dat die golwe in dieper water begin breek wat aanlieding tot verminderde oorspoeling gegee het. Daar word moontlike oplossings vir oorspoeling voorgestel gebaseer op twee strandvlakke en die implementering van ’n teruggebuigde strandmuur. Voorgestelde golfkruinvlakke vir die muur is al langs die lengte van die strand gemaak vir beide die 1:20- en 1:100-jaar watervlakke.

Thesis (MEng)–Stellenbosch University, 2014. / ENGLISH ABSTRACT: The Port of Durban is forecasted to reach its capacity in terms of container handling soon, which necessitates the investigation of an alternative port in the vicinity. The old Durban Airport site has been identified as a potential location to develop a new deep water container harbour. This is driven by a demand for deep water berth capacity as a result of shipping liners preferring the benefits of scale in their operations, leading to the use of larger ships with deeper drafts. To protect the new port from wave energy penetrating inside the basin as well as from sedimentation from the adjacent beaches, the design and construction of breakwaters are required. The proposed main breakwater for this dig-out port is expected to extend 1 200m into the sea, up to depths of 30m at the seaward roundhead. The deeper parts of the breakwater face wave onslaught in a different manner than a conventional breakwater in shallower waters. At these larger depths, the breakwater has to dissipate the energy of non-breaking waves. In this thesis, the wave climate nearshore, adjacent to the proposed breakwater is studied and extreme wave events are simulated with a SWAN numerical model. The results for a range of wave conditions, corresponding to selected events up to a return period of one in 100 years, are presented. A study of deep water breakwaters was undertaken to investigate other examples of similar structures. This indicated a clear distinction between vertical wall type breakwaters and the more traditional rubble-mound type breakwaters. For this thesis, a rubble-mound breakwater was chosen as the breakwater type for testing under conditions of the Durban Dig-Out Port (DDOP). Focussing on a deep water trunk section of the proposed main breakwater, a concept cross-section was designed using deterministic design methods. The formulae incorporated in this method did however not take into account the packing density of the armour layer and only assumed the recommended values. The hypothesis is thus put forward that the breakwater will still be hydraulically stable for packing densities below the recommended values. This would decrease material consumption and save on cost over the entire breakwater. A physical model was designed to experiment with different armour layer configurations of single- and double layer Cubipod arrangements. The unit was chosen for its massive shape and structural integrity even during impact. A physical model study was performed at the facilities of the CSIR in Stellenbosch. It entailed setting up a fixed-bed two-dimensional physical model in a glass wave flume. Measuring wave heights, wave reflection, overtopping, wave transmission and armour damage, the hydraulic stability and operational performance were analysed for several tests. Based on the results of the first few test series, alterations were made to the breakwater geometry and armouring. The results confirmed the hypothesis that lower packing densities were still hydraulically stable under 1 in 100 year return period wave conditions without inhibiting operational performance. A final cross-section is presented as concept design for the deep section of the proposed DDOP main breakwater. / AFRIKAANSE OPSOMMING: Volgens vooruitsigte gaan Durban hawe binnekort sy kapasiteit bereik wat die hantering van skeepshouers betref. Hierdie verwikkeling noodsaak die ondersoek na ‘n alternatiewe hawe in die nabye omgewing. Die voormalige Durban lughawe is intussen geïdentifiseer as ‘n potensiële perseel waar ‘n diep water houervrag hawe ontwikkel kan word. Dit word gedryf deur die aanvraag na diep water kaai kapasiteit as gevolg van skip operateurs wat skaalvoordele verkies, sodat groter skepe met diep rompe meer populêr word. Die ontwerp en konstruksie van breekwaters word dus benodig, om te verhoed dat beide golwe, sowel as sediment van aangrensende strande, die hawe binnedring. Die voorgestelde hoof breekwater vir hierdie hawe sal na verwagting tot 1200m in die see in strek, waar dit tot 30m diep is naby die seewaartse hoof van dié breekwater. Die dieper gedeeltes van só ‘n breekwater sal blootgestel word aan ‘n ander soort golf aanslag as ‘n soortgelyke konvensionele breekwater in vlakker water. In hierdie waterdiepte is die breekwater verantwoordelik vir die energie verbreking van ongebreekte golwe. In hierdie tesis word die golfklimaat langs die kus, naby aan die voorgestelde breekwater bestudeer. Die uiterste golf gebeurtenisse word gesimuleer met ‘n SWAN numeriese model. Die resultate van ‘n reeks golf kondisies, ooreenstemmend met bepaalde gebeurtenissemet herhaal periodes van tot 100 jaar, word aangebied. ‘n Studie van diep water breekwaters is onderneem om voorbeelde van soortgelyke strukture te ondersoek. Die studie toon ‘n definitiewe onderskeid tussen vertikale muur breekwaters en die meer tradisionele “rubble-mound” breekwater tipes. Vir hierdie tesis is die “rubble-mound” breekwater tipe gekies vir toetsing, onderhewig aan die kondisies van die “Durban Dig-Out Port” (DDOP). ‘n Konsep deursnit is ontwerp vir ‘n diep water romp gedeelte van die voorgestelde hoof breekwater, deur van deterministiese metodes gebruik te maak. Die formules soos vervat in hierdie proses maak egter nie voorsiening vir die pakdigtheid van die bewapeningslaag nie, maar aanvaar slegs die voorgestelde waardes. Die hipotese word dus aangevoer dat die breekwater steeds hidrolies stabiel sal wees vir pakdigthede wat laer as die voorgestelde waardes is. Dit sal die verbruik van materiale verlaag en lei tot koste besparings vir die breekwater. ‘n Fisiese model is ontwerp om te eksperimenteer met verskillende opstellings van die bewapeningslaag. Dit sluit enkel- en dubbel laag bewapening met Cubipod eenhede in. Hierdie eenheid is gekies vir sy massiewe vorm en strukturele integriteit, selfs tydens impak. ‘n Fisiese model studie is uitgevoer by die fasiliteite van die WNNR in Stellenbosch. Dit het die opstel van ‘n vaste-bodem, twee-dimensionele fisiese model in ‘n glas golftenk (“wave flume”) behels. Hidroliese stabiliteit en operasionele werksverrigting is geanaliseer deur golf hoogtes, -weerkaatsing, -oorslag, -deurlating, en skade aan die bewapening te meet vir verskeie toetse. Gebasseer op die resultate van die eerste paar toetsreekse, is veranderinge gemaak aan die breekwater se geometrie en bewapening. Die resultate het die hipotese bevestig dat laer pakdigthede steeds hidrolies stabiel is tydens golf kondisies met ‘n 1 in 100 jaar herhaal periode, sonder om die werksverrigting van die breekwater te belemmer. ‘n Finale deursnit word voorgestel as ‘n konsepontwerp vir die diep water deursnit van die DDOP se hoof breekwater.

L’objectif de cette thèse est la détermination des causes du phénomène d'érosion marine, et la production de propositions d'orientation pour la protection de la plage de l’Almanarre. L'étude commence par la collecte et l'analyse de toutes les données disponibles, pour trouver des règles d’évolution du tombolo de Giens. Puis, le logiciel de modélisation MIKE a été appliqué pour confirmer l'hypothèse et à obtenir une meilleure connaissance de la dynamique à l’œuvre au tombolo Ouest. Ensuite, nous avons couplé des facteurs simultanément : les houles, les courants côtiers, et les transports des sédiments pour les différents régimes de houle et de vent dans MIKE 21. Enfin, la possibilité de stabiliser le tombolo Ouest est discutée pour donner des suggestions sur les choix de solutions adaptées. / The objective of this thesis is the determination of the causes of sea erosion phenomenon, and the production of policy proposals for the protection of the beach. The study begins with the collection and analysis of all available data, to find evolution rules of tombolo of Giens. Which aims to explain how the hydrodynamic processes and sediment transport occur. Then MIKE modeling software was applied to confirm the hypothesis and get a better understanding of the dynamics at work in the Western tombolo. Then, we coupled factors simultaneously: waves, coastal currents and sediment transport for different regimes of wave and wind in MIKE 21. Finally, the possibility of stabilizing the Western tombolo is discussed to give suggestions on the choice of solutions.