NUMERICAL MODELS OF THE ATMOSPHERE

Author: MURILLO CASTARLENAS JAVIER.
Year: 2005.
University: ZARAGOZA [www.unizar.es].
Place of defense: CENTRO POLITÉCNICO SUPERIOR.
Place of preparation: ÁREA DE MECÁNICA DE FLUIDOS.

Summary: In the last years the numerical modelling of shallow water flow in two dimensions in complex geometries involving transient flow and movable boundaries has been a challenge for modellers. Upwind finite volume methods based on Roeâs approximate solver, initially developed for solving problems in gas dynamics, have been accepted as reliable and accurate for the numerical solution of the shallow water equations. The aim of this thesis consists on generating robust and accurate methods to solve realistic hydraulic problems. For that reason, in this work a transient 2D coupled vertically averaged flow/transport model with variable bed elevation surface generated by erosion/deposition processes is presented. The model presented deals with all kind of bed geometries and guarantees global conservation and positive values of both water level and solute concentration in the transient solution. It is based on Roeâs approximate Riemann solver for finite volume schemes. The convenience of considering the fully coupled system of equations is demonstrated along the thesis as it allows a correct upwinding of the source terms ensuring the exact balance of the numerical fluxes, reproducing exactly steady state cases with flow in movement and situations of still water for first and second order approximation. To test the efficiency of higher order resolution methods, the structure of several slope limited second order explicit finite volume schemes is presented as an extension of the first order explicit upwind scheme. One contribution of this work is the importance of including in the slope limiters new conditions to provide the correct balance of the source terms in steady state cases. The sign of the redefined advection velocities must be kept equal to the one obtained for first order approximation. Also the interpolation function for the water depth must be obtained interpolating the water level surface. Another remarkable contribution of this work is the extension of the explicit first order upwind finite volume scheme to values of CFL greater than one for irregular grids, motivated by a minimum numerical diffusion and minimum computational cost. In the presented coupled shallow water-solute transport model the conserved variable is the solute mass. Therefore it is possible to find situations where the final solute concentration oscillates. In particular two cases have been found. An accurate, stable and conservative numerical solution has been provided in this thesis to avoid this problem and has been formulated for all the different construction of finite volume schemes described in this work.The coupled model is defined including solute diffusion. Three numerical discretization techniques are described to discretize this source term. From them, an implicit technique is selected as the best option, providing the less diffusive solution requiring the minimum computational cost. To overcome the generation of negative values of depth and concentration, that can appear as a consequence of existing wetting/drying fronts over variable bed levels, or by the generation of new ones when dry areas appear, a coupled conservative scheme (CCS) with a specific modification of the Riemann solver is proposed. The numerical stability constraints of the explicit model are stated incorporating the influence of the flow velocity, the bed and the possible appearance of dry cells, resulting in a robust and accurate model. It is remarkable the effort that has been required to analyse properly each possible influence, and how if all of them are not integrated in the numerical scheme it is not possible to guarantee a successful solution. The CCS scheme is presented in both fist and second order approximations, leading this second option to new conditions over interpolation functions, in special in wetting/drying fronts. The considerable increas 8 e in the 62b computational effort required for second order does not hold always truly increments in the accuracy of the results, specially in realistic cases. The mathematical modelling of erosion/deposition processes is done under the assumption of low concentrations of sediment load. The flow mixture is then described using the fully coupled system of equations, and the Roeâs approximate Riemann solver for finite volume schemes is applied again. The upwind treatment of the bed elevation variation in time and in space source terms provides an exact balance of the numerical fluxes guaranteeing steady state solutions. Also the performance of the model using first order enlarged in time scheme supplies encouraging results, reducing expectaculary the computational cost of large time scale processes. In the case of transient boundaries the CCS scheme is reformulated including an implicit upwinding formulation of the bed elevation source term.

Author: TRIANA MORENO JORGE EDUARDO.
Year: 2006.
University: POLITÉCNICA DE VALENCIA [www.upv.es].
Place of defense: Universidad Politécnica de Valencia.
Place of preparation: Universidad Politécnica de Valencia.

Summary: At present the evaluation of the safety of dams can be handled in two ways. The first is the traditional approach in which it is assumed that there is a risk of failure of the dam since it was built under some exacting design criteria, while the second is aimed at risk for the risk of failure of the dam, which can be identified, evaluated and managed even if it seems unlikely. In recent years include the security problem starting from the second approach has been motivated by a variety of causes that can be emphasized: the aging and deterioration of existing dams, dams that do not meet existing hydrologic or seismic solicitations and a growing demand social quantify the risk associated with catastrophic events combined with an aversion to risk in a context of apparent increase in extreme weather events. Another important aspect in the estimated impact is the fact that this has been mainly focused on the damage and effects caused by the flooding resulting from the failure of prey in simplified considering the effects on the system of water resources that owns the dam . In this context, to have a structure for the estimation of various kinds of consequences (Persons at risk and loss of life, flood damage, costs of rebuilding the dam, costs or effects on water resource systems) is useful and right, so in this work raises an overall structure for estimating consequences produced by the dam failure in the techniques of risk analysis, which has as its main goals to organize and improve the development of these estimates and both serve as a guide for estimating costs associated with the development of the work.

Author: BERNABEU DIAZ JORDI.
Year: 2006.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: ETSEIB.
Place of preparation: EPSC, EDIFICI C1 Campus BAIX LLOBREGAT.

Summary: The human exploration of the Solar System has gained a lot of interest by space agencies and many efforts are currently under development to solve various problems for human adaptation to the space environment. One of the key points in the hazards of the space environment, especially outside the influence of the Earthâs magnetic field is the radiation environment. The space radiation environment is a mixed field of high energy particles driven by the solar activity. There are two sources of particle radiation: solar particles events (SPE) and galactic cosmic rays (GCR). Radiation protection assessment for this environment involves calculating the protection effects provided by shielding materials covering the habitable parts of a space vehicle or habitat, and the health risks to astronauts due to exposure to this radiation. This work uses the MULASSIS computer simulation code developed for the European Space Agency to calculate the dose received by astronauts under several radiation source conditions, shielding material, and shielding thickness for a simplified geometry. MULASSIS is based on Geant4, a general purpose system for radiationmatter interaction using Monte Carlo algorithms. SPE sources used correspond to a statistical model for worstcase situations, and proton fluence spectra from selected solar events that are typically used as a reference for radiation protection calculations. GCR sources used in our calculations correspond to the CREME96 model, calculated for the worst GCR conditions that happen during solar minimum. Two different scenarios are examined: deep space and the Lunar surface. For the deep space scenario, the geometry of the simulation consists of a set of planar slab layers on which the incoming particles impinge. One slab represents the shielding material, and behind it a reference volume consisting of a 30 cm thick slab of water represents a human body. The dose deposited on the reference volume by both the primary and secondary particles is calculated and compared to currently established limits. Four materials have been considered for shielding: Aluminium, water, polyethylene, and lithium hydroxide. Al is the most common structural material used in the space industry, while the other materials represent different hydrogenrich compounds. Low Z and hydrogenrich materials are found to have better shielding properties for the extremely high energetic particles present in space. The effect of shielding thickness is studied varying the simulation volumes between 2 g cm-2 to 100 g cm-2 areal density; from nominal spacecraft wall thicknesses to considerably thick shielding. The results show that exposure to most SPE can be reduced to acceptable levels with medium shielding (about 20 g cm-2 ), with no clear differences among the materials. The only noticeable difference is that Al shielding is shown to yield the lowest protection compared to the other three materials in all cases. High energy SPE and GCR particles show a higher penetration through the shield, and ultimately the intensity of these particles fluence will determine the limiting conditions for deep space operations. The Lunar surface radiation environment is analyzed, considering the backscattered fluence of secondary particles generated by the interaction of the primary radiation with the regolith that covers the Lunar surface. The effect of this interaction in the proton and neutron fluences within the top 1 m of the Lunar surface is studied for three different regolith materials representative of both Lunar highlands and maria. Protons depth profile is shown to decrease, while there is an increased production of secondary neutrons at depths between 20 cm and 40 cm into the regolith. This is the depth at which most of the spallation processes take place, producing recoil protons, nuclear fragments, and high energy neutrons.