Ingegneria civile e architettura

Renewable energy resources, such as wind, are available worldwide. Locating areas with high and continual wind sources are crucial in pre-planning of wind farms. Vast offshore areas are characterized with higher and more reliable wind resources in comparison with continental areas. However, offshore wind energy production is in a quite preliminary phase. Elaborating the potential productivity of wind farms over such areas is challenging due to sparse in situ observations. Mediterranean basin is not an exception. The overall aim of this thesis is to perform analysis in model efficiency in estimation of wind energy from regional to local scale. First, we are proposing numerical simulations of near-surface wind fields from regional climate models (RCMs) in order to obtain and fill the gaps in observations over the Mediterranean basin. Four simulations produced with two regional climate models are examined. Remote sensing observations (QuikSCAT satellite) are used to assess the skill of the simulated fields. A technique in estimation the potential energy from the wind fields over the region is introduced locating the three potentially interesting sub-regions for wind farms. Then, we use local-scale model (large-eddy simulation) with implemented parameterization of wind turbine in order to simulate real case flow in theoretical wind farm. Information reported with regional climate model would be used to create inflow conditions for the selected sub-region of the Mediterranean Sea for simulating theoretical offshore wind farm. Finally, we would compare the estimation of wind power potential obtained by regional climate model and power production of theoretical wind farm obtained with large-eddy simulations for chosen sub-region. Within this multi-scale approach, we would present different numerical computational efficiency in application of wind energy and justification in usage of both regional and local scale models. The novelty of this multi-model methodological approach could be considered in offering significant information for wind industry.

In this thesis, a new large-eddy simulation solver, LES-AIR, has been developed, tested and applied to a practical situation of flow and pollutant dispersion in urban environments. The novelty of the present research resides in the application of a high resolution, accurate, CFD technique to the simulation of real-life flows. The code uses a body fitted curvilinear grid to account for the macro geometry such as terrain slopes, and is thus able to reproduce in detail the complex conditions typical of urban areas; by utilizing the technique of immersed boundaries, the code is also able to mimic the presence the micro complexities such as anthropic structures (i.e. buildings). The first part of the thesis presents a detailed description of the mathematical and numerical model on which the code is based. An extensive set of validation tests was performed in flow configurations having an increasing degree of complexity in terms of forcing and geometry. The numerical model thus validated is applied for obtaining flow and pollutant dispersion in the Servola-Valmaura suburban area of the city of Trieste in Italy. The pollutant was introduced into the domain from a line source near the ground, mimicking the emission from vehicular traffic. In spite of the idealizations inherent to the model, LES-AIR is able to predict the flow and dispersion patterns well, and has proven to be a reliable tool for adaptation in urban pollution studies.

Mostriamo un modello allo stato dell’arte, che considera i principali processi fisici che governano il greggio in mare nelle prime ore dopo il rilascio, (Zanier, et al., 2014). Le particelle e i tar sono trattati come particelle lagrangiane, ognuna con la propria densità e il proprio diametro; consideriamo le forze principali che agiscono su di esse ossia: galleggiamento, trascinamento e la forza di Coriolis. Il greggio in forma di film sottile è modellato tramite le equazioni proposte da Nihoul (Nihoul 1983/84). Il modello originale di Nihoul considera le forze principali (ossia gravità, stress indotto da vento e correnti marine) che agiscono sulla macchia e governano il suo trasporto e diffusione, sulla superficie del mare, nelle prime 24 ore dopo il rilascio. Il nostro miglioramento al modello consiste nell’introduzione della forza di Coriolis evitando di utilizzare formulazioni empiriche (Zanier, et al., 2015). Infine i principali processi di weathering che agiscono sulla macchia nelle prime 12-24 ore dopo il rilascio (ossia emulsificazione ed evaporazione) sono considerate in accordo con i modelli presenti in letteratura (Mackay, Peterson, et al., 1980 e Mackay, Buist, et al., 1980, rispettivamente). Per preservare un’accuratezza del secondo ordine del metodo numerico, i termini convettivi, nel modello Euleriano, sono discretizzati usando SMART uno schema numerico upwind del terzo ordine (Gaskell and Lau 1988). Il modello è validato con dei casi test standard. Le correnti marine sono risolte con il modello LES-COAST (IEFLUIDS Università di Trieste), un modello numerico ad alta definizione, adatto per simulare flussi in aree costiere e portuali. Il modello LES-COAST risolve la forma filtrata delle equazioni di Navier-Stokes tridimensionali e non-idrostatiche, assumendo che valga l’approssimazione di Boussinesq; e l’equazione di trasporto degli scalari, salinità e temperatura. Il modello usa l’approccio della large eddy simulation per parametrizzare la turbolenza, le variabili sono filtrate con una funzione filtro, rappresentante la grandezza delle celle. I flussi di sottogriglia (SGS), che appaiono dopo l’operazione di filtraggio delle equazioni, sono parametrizzati con un modello di Smagorinsky anisotropo con due eddy viscosity, per adattare il modello a simulare flussi costieri dove le lunghezze scala orizzontali sono molto più grandi di quelle verticali (Roman et al., 2010 ). Le diffusività di sotto griglia della temperatura e salinità, cioè i numeri di Prandtl e Schmidt, sono imposti come $Pr_{sgs}=Sc_{sgs}=0.8$, assumendo che l’analogia di Reynolds sia valida per entrambi gli scalari. La complessità geometrica che caratterizza le aree costiere, è trattata con una combinazione di griglie curvilinee e il metodo dei contorni immersi (IBM) (Roman, Napoli, et al., 2009). L’azione del vento sulla superficie libera del mare è imposta tramite una formula proposta da Wu (Wu, 1982), nella quale lo stress del vento sul mare è calcolato dalla velocità del vento a 10 m sopra il livello del mare. Allo stress aggiungiamo una varianza del 20% per agevolare la generazione di turbolenza e per tener conto che l’azione del vento non è costante nel tempo e nello spazio. Inoltre vicino agli ostacoli, come moli, navi e frangiflutti, lo stress del vento è ridotto linearmente, per considerare la riduzione del vento che si ha nelle zone di ricircolo. Sui contorni aperti le velocità e le quantità scalari sono ottenute innestando il modello LES-COAST con modelli di larga scala (Petronio, et al., 2013) oppure sono impostati secondo dati rilevati. Vicino ai bordi solidi le velocità sono modellate tramite funzioni parete (Roman, Armenio, et al., 2009). Il modello di rilascio di petrolio e il modello idrodinamico sono stati applicati assieme per simulare degli ipotetici scenari di trasporto e diffusione del greggio in mare nel porto di Barcellona (Mar Mediterraneo Nord-Ovest, Spagna, Galea, et al. 2014) e nella baia di Panzano (Mar Adriatico, Nord, Italia).

In this work the turbulent field developing in case of local erosion around a 45° wing-wall bridge abutment was investigated numerically. Three different scour conditions were considered: beginning of the process, logarithmic phase and equilibrium stage. The flow field was computed using a wall-resolving large eddy simulation (a simulation where the near-wall viscous sub-layer is directly resolved) and the bathymetric data were taken from physical experiments with an equivalent geometry. The dynamics of the coherent structures forming around the obstacle and inside the scour-hole was investigated and its influence on the modeling of the problem and on the erosion process was discussed. The analysis suggested that the full dynamics of the vortex system should be directly solved since simple eddy-viscosity models, as the k-ε model in RANS approach, were found to be not suited for this kind of problem and since high-order statistics were found to be important for the evolution of the local scour. The results of the present study may be helpful to formulate new physical-based local scour models to be used for practical evaluation of the scour depth around bridge abutments.

Computational Fluid Dynamics (CFD) is an established tool for consulting and for basic research in fluid mechanics. CFD is required to provide information where analytical approaches or experiments would be impossible or too expensive. Most of the flows of engineering interest are turbulent. Turbulence is an unresolved problem of classical physics, because of the non linearity of the fluid motion equations. At the moment the only way to face them is numerically. Turbulence is composed of eddies in a broad range of size. To solve numerically the Navier-Stokes equations, the equations set that governs the fluid motion, a very fine grid is necessary in order to catch also the smallest eddies. The computational cost increases as Re3 (Re = ul/ is the Reynolds number with u and l an inertial velocity and length scales and the kinematic viscosity). Real life problems are characterized by very large Reynolds numbers and the consequent computational cost is enormous. So the direct solutions of Navier-Stokes equations (DNS) is not feasible. In many applications it is not necessary to solve all the eddies, it can be sufficient to supply the effects of unresolved scale to the flow. In Large Eddy Simulation (LES) most of the scales of motion are directly solved, in particular all the large energy carrying scales. These scales are influenced by the boundaries and they are strongly anisotropic. The smaller and dissipative scales must be modeled, but these scales loosing memory of the boundary conditions are more isotropic and hence formulating a general model that accounts for their effect is relatively easier. Large Eddy Simulation is a prospective tool for investigation in real life problems, in particular when high detailed analysis is required. This is the case for many industrial and environmental processes. For example, acoustic problems due to hydrodynamic noise are governed over a range of large scales which are easily reproduced by LES solution. However in these types of flows many difficulties arise also for LES. In general these flows are characterized by high Reynolds number. Wall-bounded flow at high Re requires high computational cost because LES is constrained to be DNS-like. Besides complex geometries are often involved. Structured or Unstructured body-fitted grid can be very hard to made, moreover unstructured grid can be expensive and not suited for LES. Scope of this thesis is to develop tools to apply LES to such configurations in order to make numerical simulation more adaptable to real life problems. In particular to deal with complex geometry an Immersed Boundary Methodology has been developed for curvilinear coordinates. The method has been applied to several test cases with good results. Then this methodology has been extended to high Reynolds number flows through the use of a wall model. In order to work on anisotropic grid, typical in sea coastal domain, a modified Smagorisky model has been proposed. Finally particle dispersion has been considered in stratified environmental flow. These tools has been applied to an industrial and to an environmental problem with good results.

A novel high-resolution, eddy-resolving numerical model (LES-COAST) is used to investigate currents, mixing and water renewal in Barcelona harbour and Taranto bay. These environmental sites are of particular importance due to the interplay between touristic and commercial activities, requiring detailed and high-definition studies of water quality within the harbour. We use Large Eddy Simulation (LES) which directly resolves the anisotropic and energetic large scales of motion and parametrizes the small, dissipative, ones. Small-scale turbulence is modelled by the Anisotropic Smagorinsky Model (ASM) which is employed in presence of large cell anisotropy. The complexity of the harbour is modelled using a combination of curvilinear, structured, non-staggered grid and the Immersed Boundary Method (IBM). Both computation grids and harbour structures are purposely constructed for these applications by appropriate programs. Boundary conditions for wind forcing at the free surface and currents at the inlets of the port are obtained from in-situ measurements (for the case of Barcelona harbour) or by nesting this numerical model into a coastal model (Taranto bay). In this dissertation thesis an important modification to LES-COAST is implemented and is proposed as a prototype scheme, namely the possibility to consider the effect of surface waves in coastal semi-closed areas. Particularly, a linear formulation of the free surface boundary condition is considered, which would be able to reproduce the presence of seiches and tides on the dynamics of the area under investigation. The methodology is validated against analytical solution for a stationary oscillating surface wave in a simple computational grid.In both harbours considered, first- and second-order statistics, such as the mean velocity field, turbulent kinetic energy, and horizontal and vertical eddy viscosities are calculated and their spatial distribution is assessed. Water residence time is also considered for the two coastal semi-closed areas examined. Finally, the LES solution is validated against available field data.The study shows the presence of sub-surface elongated rolling structures (with a time scale of a few hours), contributing to the vertical water mixing. The time-averaged velocity field reveals intense upwelling and downwelling zones along the walls of the harbours. The analysis of second-order statistics in these harbours shows strong inhomogeneity of turbulent kinetic energy and horizontal and vertical eddy viscosities in the horizontal plane, with larger values in the regions characterized by stronger currents. The water renewal within the port is quantified for particular sub-domain regions, showing that the complexity of the harbour is such that certain inner basins of Barcelona harbour have a water renewal of over five days, including its yacht marina area, and over seven days for Taranto bay. For the Barcelona simulation, the LES solution compares favourably with available current-meter data; it is also compared with a RANS solution obtained in literature for the same site under the same forcing conditions, the comparison demonstrating a large sensitivity of properties to model resolution and frictional parametrization.

Evaporation and condensation of thin liquid films on solid surfaces are common elements of industrial processes. In many cases they have a significant impact on the physics of the studied case. At the same time, experimental studies can prove to be troublesome, mostly because of the amount of possible setups, complex geometries of interest, numerous materials being used and cost. For that reason it is reasonable to study this phenomena using numerical methods. Having the advantage in speed and cost of performance, computational studies become a valuable tool. For evaporation and condensation process, one has to deal with buoyancy driven fluid flows, conjugated heat transfer between gaseous and solid phases, film thickness modeling, vapor phase behavior, and phase transition of the thin fluid film into vapor phase. The strong conjunction and mutual interaction of mentioned effects is the main focus of presented work. The gas phase behavior is being calculated using incompressible Navier-Stokes equations under Boussinesq approximation. The solutions of the partial differential equations are obtained with numerical methods using Eulerian finite volume discretization (Kundu and Cohen [2002]). Time advancement is being treated with second order implicit discretization. For cases with high Reynolds number, large eddy simulation (LES) techniques are used. Due to the complexity of the geometries of interest a dynamic computation of the Smagorinsky constant is preferred, applying the lagrangian dynamic model proposed by Meneveau et al. [1996]. The liquid film present on the surface of the solids is modeled following Petronio[2010]. Since the film is thin, it is assumed that it can be represented only by its thickness. This also leads to assumption that the heat transfer through the film is instantaneous. The vapor is represented by concentration of this phase in the volume of gas. The concentration is transported by convection and diffusion. The phenomena of evaporation and condensation of the thin films are driven by the presence of concentration gradients next to the surfaces. Phase transition of vapor to fluid, or other way around, acts on the energy balance, id. est latent heat is released into the gas when condensation occurs or the solid is cooled during evaporation. The heat transport is modeled in both solid and fluid domains. The case is split into separate regions with different material properties. These regions are solved one by one in a serial way using numerical techniques consistent with domain decomposition methods described by Quarteroni and Valli [1999]. The energy transport among the regions is performed by applying a heat coupling boundary conditions. The main focus of this work is to provide a reliable model for simulation system with complex physics involving fluid motion, heat transport in multi region domains (fluid-solid), vapor transport, thin film evolution and evaporation and condensation effects on energy balance. Proposed model is validated on simple geometries and later applied to problem of evaporation in vertical channel flow. The reference to the channel case is work of Laaroussi et at. [2009]. Presented study aims in providing comprehensive insights into physical effects that appear when the solid wall is being directly modeled and when latent heat transformations are taken into account. The final test is performed on a vertical channel with forced turbulent flow, directly modeled solid walls and evaporation or condensation happening on the boundary. Having the model working within such complex frame allows for its future usage in elaborate industrial applications.

L'obiettivo principale del progetto di ricerca sviluppato nella presente tesi di dottorato è quello di comprendere meglio le problematiche riguardanti le prestazioni di asciugatura della lavastoviglie, con speciale riferimento alla modellazione dei fenomeni di evaporazione e condensazione che avvengono nella vasca. Tipicamente l'acqua per il risciacquo finale viene portata ad una temperatura di 70°C che scalda le stoviglie permettendo a queste di immagazzinare energia termica. La lavastoviglie si raffredda dall'esterno, cosicché la vasca risulta essere più fredda dei piatti posti all'interno. In questo contesto l'acqua può evaporare dalle superfici delle stoviglie e condensare sulle pareti della vasca stessa. Il sistema fisico può essere descritto come un flusso in presenza di cambiamenti di fase. Tali tipologia di flussi ha un ruolo cruciale in molti processi naturali e tecnologici, in particolare in quelli in cui si hanno asciugatura o formazione di condensa sulle superfici solide. Tuttavia, pur essendo così comuni nelle applicazioni ingegneristiche, la loro comprensione è lontana dall'essere completa. Il complesso problema fisico può essere suddiviso in tre sotto-problemi: la trasmissione del calore tra il corpo bagnato ed il liquido sulla sua superficie; il trasferimento di calore e massa tra la fase liquida e quella gassosa; il flusso della fase gassosa che risulta essere molto influenzato dalle forze di galleggiamento dovute alle variazioni di densità causate dalla diffusione di temperatura e concentrazione di vapor acqueo. Dallo studio della letteratura risulta che tale problema non sia stato ancora investigato completamente. In particolare non è mai stato proposto un modello adatto a scopi ingegneristici, cioè per problemi di larga scala con geometrie complesse, che consideri l'evoluzione del film liquido durante processi di asciugatura. Questo progetto vuole contribuire allo sviluppo della ricerca in questo settore. Il modello matematico del flusso d'aria in presenza di evaporazione e condensazione è stato implementato numericamente nell'ambiente open-source OpenFoam. Il modello consiste nella formulazione delle equazioni di Navier-Stokes per flussi incomprimibili più le equazioni del trasporto per la temperatura e la concentrazione di vapore. Entrambi gli scalari sono considerati attivi e le variazioni di densità sono state incorporate sotto l'approssimazione di Boussinesq. Si assume inoltre l'approssimazione a film sottile, per cui si è inteso che film liquidi, gocce ed, in genere, le zone bagnate di un solido possano essere considerate come un film liquido continuo. Tale film sottile è stato interpretato come una condizione al contorno per il flusso d'aria, prescrivendo una condizione di Dirichlet per la temperatura e per il vapore. Quest'ultimo all'interfaccia del liquido è considerato in condizione di saturazione. Il calcolo della velocità di evaporazione all'interfaccia, imposta anche come condizione al contorno per il campo di velocità, ha permesso la quantificazione del processo di evaporazione/condensazione consentendo il calcolo della massa d'acqua evaporata/condensata. Il modello numerico è stato validato con i dati di letteratura per poi essere applicato nello studio del flusso su un cilindro bagnato, tra due piani paralleli. In questo lavoro è stato evidenziato l'effetto sul flusso di evaporazione attorno al cilindro delle condizioni alle pareti, considerate come bagnate o asciutte ed adiabatiche. Inoltre è stato valutato anche l'effetto della distanza del cilindro stesso dalle pareti. Successivamente il modello è stato applicato ad una geometria 2D della lavastoviglie. I risultati mostrano che il flusso evolve secondo uno schema preciso: le forze di galleggiamento danno luogo ad un moto convettivo che si alza dalle stoviglie più calde ed umide, che poi scende lungo alle pareti più fredde e meno umide. Un'ulteriore analisi è stata fatta simulando il processo fino all'asciugatura completa di un film uniformemente distribuito su tutte le stoviglie. Negli stadi intermedi del processo è stato osservato che, attorno alle porzioni di stoviglie già asciutte, il galleggiamento risulta essere ridotto e la velocità dell'aria minore, per il mancato rilascio di vapor acqueo. Un passo ulteriore verso la modellazione della lavastoviglie è stato condotto considerando una geometria 3D semplificata per testare il modello e verificare le caratteristiche richieste alla griglia computazionale. Anche per questa configurazione è stato osservato l'instaurarsi del moto convettivo e l'effetto dell'asciugatura sul flusso. Infine si è iniziato a studiare il caso della lavastoviglie 3D. La simulazione è potuta durare pochi secondi fisici, nei quali hanno iniziato a svilupparsi sopra le stoviglie i caratteristici plume. Successivamente delle instabilità numeriche hanno dato luogo a valori di pressione non fisici determinando l'interruzione del programma. Tale comportamento è stato spiegato dalla mancata dissipazione turbolenta nel flusso. L'attivazione del modello LES di Smagorinsky con l'analogia di Reynolds per la determinazione delle diffusività turbolente dei due scalari ha dato luogo ad una soluzione numericamente stabile. Tuttavia la eccessiva viscosità di sotto-griglia ha sovrastimato la diffusione degli scalari, inficiando l'accuratezza della simulazione. Per includere nel modello l'accoppiamento termico che caratterizza il processo di asciugatura è stata scelta la tecnica di decomposizione di domini detta di Dirichlet-Neumann in quanto è risultata essere efficace e semplice da implementare. Essa impone la continuità della temperatura e il bilancio dei flussi di calore attraverso le interfacce. Inoltre è stato proposto un modello opportuno per la distribuzione della temperatura nel film liquido, per riprodurre nella maniera corretta il trasferimento di calore attraverso il film stesso. Ciascuna di queste due parti è stata implementata e testata individualmente cosicché la loro inclusione nel modello potrà avvenire in un successivo sviluppo della presente ricerca. Inoltre è in fase di sviluppo l'implementazione del modello di sotto-griglia LES dinamico lagrangiano che permetterà di superare i limiti riscontrati nel utilizzo del modello di Smagorinsky.

Turbulent Rayleigh-Bénard convection (RBC) occurs when a shallow layer of fluid is heated from below. It is a challenging subject in non-linear physics, with many important applications in natural and engineering systems. Because of the complexity of the governing equations, analytical progress in understanding convection has been slow, and laboratory experiments and numerical simulations have assumed increased importance. In regard to numerical work, Large-Eddy Simulation (LES) techniques have proved to be reliable and powerful tool to understand the physics since it provides better coverage for measurements, that are not as easily obtained in physical experiments or the other numerical approaches. This thesis addresses different aspects of Rayleigh-Bénard convection in fully developed turbulent regime through Large Eddy Simulation (LES) to shed light on some important aspect of the geometrical shape of the convection cell. The layout of the thesis is as follows: In Chapter 1, we first introduce Rayleigh-Bénard convection and the equations and parameters that govern it. This is followed by a discussion on different types of boundary conditions used in numerical and theoretical studies of RBC. Subsequently we present various convection states that are observed analytically and experimentally in RBC as a function of Ra and Ʈ. To this end we present a brief survey of the analytical, experimental and numerical works on confined thermal convection. We introduce different regimes and related scaling according to Grossman and Lohse theory. We also present the experimental and numerical results related to the Large Scale Circulation (LSC) within different geometries. In Chapter 2, we present the details of the numerical methods used to solve the governing non-linear equations . In the second part, we provide the details of the solver and the algorithm used to solve the RBC dynamical equations in a Cartesian geometry together with boundary conditions. In Chapter 3, we demonstrate that our numerical method and solver give results consistent with earlier numerical results. Results from the Direct Numerical Simulations (DNS) and Large Eddy Simulation (LES) with constant and dynamic subgrid scale Prandtl number (P_sgs) are presented and compared. We observe close agreement with Lagrangian dynamic approaches. In the first part of Chapter 4 we analyse the local fluctuations of turbulent Rayleigh-Bénard convection in a cubic confinement with aspect ratio one for Prandtl number Pr = 0.7 and Rayleigh numbers (Ra) up to 10^9 by means of LES methodology on coarse grids. Our results reveal that the scaling of the root-mean-square density and velocity fluctuations measured in the cell center are in excellent agreement with the unexpected scaling measured in the laboratory experiments of Daya and Ecke (2001) in their square cross-section cell. Moreover we find that the time-averaged spatial distributions of density fluctuations show a fixed inhomogeneity that maintains its own structure when the flow switches from one diagonal to the other. The largest level of rms density fluctuations corresponds to the diagonal opposite that of the Large Scale Circulation (LSC) where we observed strong counter-rotating vortex structures. In the second part we extended our simulations and Ra up to 1011, in order to identify the time periods in which the orientation of LSC is constant. Surprisingly we find that the LSC switches stochastically from one diagonal to the other. In Chapter 5, we study the effect of 3D-roughness on scaling of Nu(Ra) and consequently on the fluctuations of density. Moreover we present the effect of roughness shape when the tip has a wide angle and the other one is smooth. We study two types of elements, one of which is a pyramid and the other is a sinusoidal function spread over the bottom (heated) and top (cooled) plates, in a cubic confinement. However preliminary results suggest that the effect of roughness appears evident at high Ra numbers when the thermal boundary layer is thin enough to shape around the obstacles.

The subject of modelling flow near wall is still open in turbulent wall bounded flows, since there is no wall layer model which works perfectly. Most of the present models work well in attached flows, specially for very simple geometries like plane channel flows. Weakness of the models appears in complex geometries, and many of them do not capture flow separation accurately in detached flows, specially when the slope of wall changes gradually. In many engineering applications, we deal with complex geometries. A possible way to simulate flows influenced by complex geometry using a structured grid, is to consider the geometry as immersed boundary for the simulation. Current wall layer models for the immersed boundaries are more complex and less accurate than the body-fitted cases (cases without immersed boundaries). In this project the accuracy of wall layer model in high Reynolds number flows is targeted, using LES for attached flows as well as detached flows (flows with separation). In addition to the body fitted cases, wall layer model in the presence of immersed boundaries which is treated totally different also regarded. A single solver LES-COAST (IE-Fluids, University of Trieste) is used for the flow simulations, and the aim is to improve wall layer model in the cases with uniform coarse grid. This is in fact novelty of the thesis to introduce a wall layer model applied on the first off-wall computational node of a uniform coarse grid, and merely use LES on the whole domain. This work for the immersed boundaries is in continuation of the methodology proposed by Roman et al. (2009) in which velocities at the cells next to immersed boundaries are reconstructed analytically from law of the wall. In body-fitted cases, since smaller Smagorinsky constant is required close to the walls than the other points, wall layer model in dynamic Smagorinsky sub-grid scale model using dynamic k (instead of Von Karman constant) is applied to optimize wall function in separated flows. In the presence of immersed boundaries, the present wall layer model is calibrated, and then improved in attached and also detached flows with two different approaches. The results are also compared to experiment and resolved LES. Consequently the optimized wall layer models show an acceptable accuracy, and are more reliable. In the last part of this thesis, LES is applied to model the wave and wind driven sea water circulation in Kaneohe bay, which is a bay with a massive coral reef. This is the first time that LES-COAST is applied on a reef-lagoon system which is very challenging since the bathymetry changes very steeply. For example the water depth differs from less than 1 meter over the reef to more than 10 meters in vicinity of the reef, in lagoon. Since a static grid is implemented, the effect of wave is imposed as the velocity of current over the reef, which is used on the boundary of our computational domain. Two eddies Smagorinsky SGS model is used for this simulation.