Overshootings and spurious oscillations caused by biharmonic mixingDelhez, Eric ; in Ocean Modelling (2007), 17(3), 183-198 Biharmonic mixing is often used in large scale numerical models of the ocean because of its scale selectivity; it effectively damps small scale noise and leaves the large scale dynamics nearly unaffected ... [more ▼] Biharmonic mixing is often used in large scale numerical models of the ocean because of its scale selectivity; it effectively damps small scale noise and leaves the large scale dynamics nearly unaffected. The biharmonic operator lacks however positiveness and monotonicity and can therefore produce unphysical results exhibiting spurious overshootings and oscillations. This problematic behaviour cannot be avoided by the addition of an ordinary Laplacian diffusion term. It appears in both continuous and discrete approaches/solutions in both unbounded and bounded domains. The overshootings and oscillations are induced by the strong damping of the smaller scale modes and are therefore comparable to the Gibbs' phenomenon. With appropriate boundary conditions, the variance of the field decreases monotonically and the oscillations are expected to remain small. The lack of positiveness is however a severe drawback for (dynamic) tracer studies. (C) 2007 Elsevier Ltd. All rights reserved. [less ▲] Detailed reference viewed: 52 (5 ULg) Unstructured, anisotropic mesh generation for the Northwestern European continental shelf, the continental slope and the neighbouring ocean; ; Delhez, Eric et alin Continental Shelf Research (2007), 27(9), 1344-1356 A new mesh refinement strategy for generating high quality unstructured meshes of the Northwestern European continental shelf, the continental slope and the neighbouring ocean is presented. Our objective ... [more ▼] A new mesh refinement strategy for generating high quality unstructured meshes of the Northwestern European continental shelf, the continental slope and the neighbouring ocean is presented. Our objective is to demonstrate the ability of anisotropic unstructured meshes to adequately address the challenge of simulating the hydrodynamics occurring in these three regions within a unique mesh. The refinement criteria blend several hydrodynamic considerations as the tidal wave propagation on the continental shelf and the hydrostatic consistency condition in steep areas. Several meshes illustrate both the validity and the efficiency of the refinement strategy. The selection of the refinement parameters is discussed. Finally, an attempt is made to take into account tidal ellipses, providing another cause for anisotropy in the mesh. (c) 2007 Elsevier Ltd. All rights reserved. [less ▲] Detailed reference viewed: 46 (9 ULg) Diagnosis of the sediment transport in the Belgian Coastal ZoneMercier, Christophe ; Delhez, Eric ![]() in Estuarine Coastal & Shelf Science (2007), 74(4), 670-683 Estimating the age of particles in marine environment constitutes an invaluable tool to understand the interactions between complex flows and sediment dynamics, particularly in highly energetic coastal ... [more ▼] Estimating the age of particles in marine environment constitutes an invaluable tool to understand the interactions between complex flows and sediment dynamics, particularly in highly energetic coastal areas such as the Belgian Coastal Zone (Southern Bight in the North Sea). To this end, the Constituent Age and Residence time Theory-CART-introduced by Delhez, E.J.M., Campin, J.-M., Hirst, A.C., Deleersnijder, E. [1999a. Toward a general theory of the age in ocean modelling. Ocean Modelling 1, 17-27] for passive water constituents is extended to describe the sediment dynamics. It is then used in combination with a three-dimensional coupled hydrodynamic-sediment transport model to investigate sediment processes in the Belgian Coastal Zone focusing on two complementary aspects of the sediment dynamics: the internal sediment motion and redistribution within the Belgian coast; and the horizontal transport. (C) 2007 Elsevier Ltd. All rights reserved. [less ▲] Detailed reference viewed: 109 (38 ULg) Three dimensional sediment transport model of the Belgian coastal zone: application of the CART theoryMercier, Christophe ; Delhez, Eric ![]() Conference (2006) Suspended sediment processes in the Belgian coastal zone are discussed by presenting the first results given by a 3D coupled hydrodynamic-sediment transport model covering the Southern Bight in North Sea ... [more ▼] Suspended sediment processes in the Belgian coastal zone are discussed by presenting the first results given by a 3D coupled hydrodynamic-sediment transport model covering the Southern Bight in North Sea. The hydrodynamic sub-model is three-dimensional, baroclinic and includes a refined turbulence closure. In the sediment transport sub-model, the dynamic of sediments is described by an evolution equation of the sediment concentration in the water column and an equation for the sediment load on the seabed. Sedimentation is modeled through the addition of a sedimentation velocity in the vertical advection term. For the evaluation of deposition and erosion terms, formu- lae based on the calculation of the bottom stress under the combined effect of prevailing currents and waves are used. For more realism, sediment loads has also been split into different sediment classes with different sedimentation, deposition and erosion characteristics. The equations are solved by a finite volume method based on an Arakawa C grid, using sigma-transform and mode-splitting numerical methods. The horizontal resolution is 500x500 m and 10 vertical sigma-layers are used. Advection is handled using a TVD scheme with superbee flux limiter. Boundary conditions are provided by coupling this model to a 3D coupled hydrodynamic-sediment transport model of the North-Western European Continental Shelf. The results are validated with measurements carried out by Haecon NV and are compared with those obtained with the MU-STM model. A sensitivity analysis of the model to critical parameters of erosion, deposition and sedimentation based on the concept of residence time was carried out. [less ▲] Detailed reference viewed: 45 (6 ULg) The residence time of settling particles in the surface mixed layer; Beckers, Jean-Marie ; Delhez, Eric ![]() in Environmental Fluid Mechanics (2006), 6(1), 25-42 The transport from the upper mixed layer into the pycnocline of particles with negative buoyancy is considered. Assuming the hydrodynamic parameters to be time-independent, an adjoint model is resorted to ... [more ▼] The transport from the upper mixed layer into the pycnocline of particles with negative buoyancy is considered. Assuming the hydrodynamic parameters to be time-independent, an adjoint model is resorted to that provides a general expression of the residence time in the mixed layer of the constituent under study. It is seen that the residence time decreases as the settling velocity increases or the diffusivity decreases. Furthermore, it is demonstrated that the residence time must be larger than z/w and smaller than h/w, where z, h and w denote the distance to the pycnocline, the thickness of the mixed layer and the sinking velocity. In the vicinity of the pycnocline, the residence time is not necessarily zero; its behaviour critically depends on the eddy diffusivity pro. le in this region. Closed-form solutions are obtained for constant and quadratic diffusivity profiles, which allows for an analysis of the sensitivity of the residence time to the Peclet number. Finally, an approximate value is suggested of the depth-averaged value of the residence time. [less ▲] Detailed reference viewed: 45 (11 ULg) On the behaviour of the residence time at the bottom of the mixed layer; Beckers, Jean-Marie ; Delhez, Eric ![]() in Environmental Fluid Mechanics (2006), 6(6), 541-547 To understand why the findings of Deleersnijder et al. [(2006), Environ Fluid Mech 6: 25-42]-the residence time in the mixed layer in not necessarily zero at the pycnocline-are consistent with those of ... [more ▼] To understand why the findings of Deleersnijder et al. [(2006), Environ Fluid Mech 6: 25-42]-the residence time in the mixed layer in not necessarily zero at the pycnocline-are consistent with those of Delhez and Deleersnijder [(2006), Ocean Dyn 56:139-150]-the residence time in a control domain vanishes at the open boundaries of this control domain-, it is necessary to consider a control domain that includes part of the pycnocline, in which the eddy diffusivity is assumed to be zero. Then, depending on the behaviour of the eddy diffusivity near the bottom of the mixed layer, the residence time may be seen to exhibit a discontinuity at the interface between the mixed layer and the pycnocline. If such a discontinuity exists, the residence time is non-zero in the former and zero in the latter. This is illustrated by analytical solutions obtained under the assumption that the eddy diffusivity is constant in the mixed layer. [less ▲] Detailed reference viewed: 20 (8 ULg) The boundary layer of the residence time fieldDelhez, Eric ; in Ocean Dynamics (2006), 56(2), 139-150 The residence time of a tracer in a control domain is usually computed by releasing tracer parcels and registering the time when each of these tracer parcels cross the boundary of the control domain. In ... [more ▼] The residence time of a tracer in a control domain is usually computed by releasing tracer parcels and registering the time when each of these tracer parcels cross the boundary of the control domain. In this Lagrangian procedure, the particles are discarded or omitted as soon as they leave the control domain. In a Eulerian approach, the same approach can be implemented by integrating forward in time the advection-diffusion equation for a tracer. So far, the conditions to be applied at the boundary of the control domain were uncertain. We show here that it is necessary to prescribe that the tracer concentration vanishes at the boundary of the control domain to ensure the compatibility between the Lagrangian and Eulerian approaches. When we use the Constituent oriented Age and Residence time Theory (CART), this amounts to solving the differential equation for the residence time with boundary conditions forcing the residence time to vanish at the open boundaries of the control domain. Such boundary conditions are likely to induce the development of boundary layers (at outflow boundaries for the tracer concentration and at inflow boundaries for the residence time). The thickness of these boundary layers is of the order of the ratio of the diffusivity to the velocity. They can however be partly smoothed by tidal and other oscillating flows. [less ▲] Detailed reference viewed: 48 (6 ULg) Study of the seasonal cycle of the biogeochemical processes in the Ligurian Sea using a ID interdisciplinary modelRaick, Caroline ; Delhez, Eric ; et alin Journal of Marine Systems (2005), 55(3-4), 177-203 A one-dimensional coupled physical-biogeochemical model has been built to study the pelagic food web of the Ligurian Sea (NW Mediterranean Sea). The physical model is the turbulent closure model (version ... [more ▼] A one-dimensional coupled physical-biogeochemical model has been built to study the pelagic food web of the Ligurian Sea (NW Mediterranean Sea). The physical model is the turbulent closure model (version I D) developed at the GeoHydrodynamics and Environmental Laboratory (GHER) of the University of Liege. The ecosystem model contains 19 state variables describing the carbon and nitrogen cycles of the pelagic food web. Phytoplankton and zooplankton are both divided in three size-based compartments and the model includes an explicit representation of the microbial loop including bacteria, dissolved organic matter, nano-, and microzooplankton. The internal carbon/nitrogen ratio is assumed variable for phytoplankton and detritus, and constant for zooplankton and bacteria. Silicate is considered as a potential limiting nutrient of phytoplankton's growth. The aggregation model described by Kriest and Evans in (Proc. Ind. Acad. Sci., Earth Planet. Sci. 109 (4) (2000) 453) is used to evaluate the sinking rate of particulate detritus. The model is forced at the air-sea interface by meteorological data coming from the "Cote d'Azur" Meteorological Buoy. The dynamics of atmospheric fluxes in the Mediterranean Sea (DYFAMED) time-series data obtained during the year 2000 are used to calibrate and validate the biological model. The comparison of model results within in situ DYFAMED data shows that although some processes are not represented by the model, such as horizontal and vertical advections, model results are overall in agreement with observations and differences observed can be explained with environmental conditions. (c) 2004 Elsevier B.V. All rights reserved. [less ▲] Detailed reference viewed: 53 (13 ULg) A note on trust-region radius update; Delhez, Eric ![]() in SIAM Journal on Optimization (2005), 16(2), 548-562 In classical trust-region optimization algorithms, the radius of the trust region is reduced, kept constant, or enlarged after, respectively, unsuccessful, successful, and very successful iterations. We ... [more ▼] In classical trust-region optimization algorithms, the radius of the trust region is reduced, kept constant, or enlarged after, respectively, unsuccessful, successful, and very successful iterations. We propose here to re. ne the empirical rules used for this update by the definition of a new set of iterations that we call "too successful iterations." At such iterations, a large reduction of the objective function is obtained despite a crude local approximation of the objective function; the trust region is thus kept nearly constant instead of being enlarged. The new update rules preserve the strong convergence property of traditional trust-region methods. They can also be generalized to define a self-adaptive trust-region algorithm along the lines introduced by Hei [J. Comput. Math., 21 (2003), pp. 229-236]. Numerical experiments carried out on 70 unconstrained problems from the CUTEr collection demonstrate the positive impact of the modified update strategy on the efficiency and robustness of quasi-Newton variants of a trust-region solver, when BFGS or SR1 updates of the approximation of the Hessian matrix are carried at all iterations. [less ▲] Detailed reference viewed: 35 (6 ULg) Influence of the Aral Sea negative water balance on its seasonal circulation patterns: use of a 3D hydrodynamic modelSirjacobs, Damien ; Grégoire, Marilaure ; Delhez, Eric et alin Journal of Marine Systems (2004), 47(1-4), 51-66 A 3D hydrodynamic model of the Aral Sea was successfully implemented to address the complex hydrodynamic changes induced by the combined effect of hydrologic and climatic change in the Aral region. The ... [more ▼] A 3D hydrodynamic model of the Aral Sea was successfully implemented to address the complex hydrodynamic changes induced by the combined effect of hydrologic and climatic change in the Aral region. The first barotropic numerical experiments allowed us to produce a comparative description of the mean general seasonal circulation patterns corresponding to the original situation (1956-1960) and of the average situation for the period from 1981 to 1985, a very low river flow period. The dominant anticyclonic circulation suggested by our seasonal simulation is in good agreement with previous investigations. In addition. this main anticyclonic gyre was shown to be stable and clearly established from February to September, while winter winds led to another circulation scenario. In winter, the main anticyclonic gyre was considerably limited, and cyclonic circulations appeared in the deep western basin and in the northeast of the shallow basin. In contrast, stronger anticyclonic circulation was observed in the Small Aral Sea during winter. As a consequence of the 10-m sea level drop observed between the two periods considered, the 1981-1985 simulation suggests an intensification of seasonal variability. Total water transport of the main gyre was reduced with sea level drop by a minimum of 30% in May and up to 54% in September. Before 1960, the study of the net flows through Berg and Kokaral Straits allowed us to evaluate the component of water exchange between the Small and the Large Seas linked with the general anticyclonic circulation around Kokaral Island. This exchange was lowest in summer (with a mean anticyclonic exchange of 222 m(3)/s for July and August), highest in fall and winter (with a mean value of 1356 m(3)/s from September to February) and briefly reversed in the spring (mean cyclonic circulation of 316 m(3)/s for April and May). In summer, the water exchange due to local circulation at the scale of each strait was comparatively more important because net flows through the straits were low. After about 20 years of negative water balance, the western Kokaral Strait was dried up and the depth of Berg Strait was reduced from 15 to 5 m. Simulation indicated a quasi-null net transport, except during the seasonal modification of the circulation pattern, in February and October. A limited, but stable, water exchange of about 100 m(3)/s remained throughout the year, as a result of the permanent superposition of opposite currents. (C) 2004 Elsevier B.V. All rights reserved. [less ▲] Detailed reference viewed: 433 (40 ULg) Tracer Methods in Geophysical Fluid DynamicsDelhez, Eric ; Book published by Elsevier Science (2004) Detailed reference viewed: 11 (2 ULg) Structural optimisation of steel frames for industrial applications; Degée, Hervé ; Habraken, Anne et alin Proceedings of the Inverse Problems Design and Optimization Symposium (2004) The paper reports the first results of a work carried out in close collaboration between Astron Buildings SA, a manufacturer of industrial steel buildings, and different groups of civil engineers and ... [more ▼] The paper reports the first results of a work carried out in close collaboration between Astron Buildings SA, a manufacturer of industrial steel buildings, and different groups of civil engineers and mathematicians of the University of Liège to develop an automatic design method for structures with tapered members. This research aims at improving the current method of trialerror followed by experienced engineers to optimize the frames constrained by a chosen national construction code and the technological constraints of the producer. The main benefit of this collaboration arises through the application of a mathematical algorithm based on the sequential quadratic programming method (SQP) in order to reduce, in the first step, the weight of the building, under the great number of constraints. The second step, not yet started, will be devoted to the minimization of the real cost of the frame. This report introduces the first results of this industrial application. [less ▲] Detailed reference viewed: 104 (14 ULg) Variability of shelf-seas hydrodynamic models: lessons from the NOMADS2 projectDelhez, Eric ; ; et alin Journal of Marine Systems (2004), 45(1-2), 39-53 Model simulations at the seasonal time scale are often lacking in any real assessment of the associated error bounds. We use here the results of nine three-dimensional hydrodynamic models covering (at ... [more ▼] Model simulations at the seasonal time scale are often lacking in any real assessment of the associated error bounds. We use here the results of nine three-dimensional hydrodynamic models covering (at least) the Southern and Central North Sea to investigate the range of model variability and model errors. The models are run as they are, i.e. with their usual grid, model domain, equation formulation and numerical details, but in a consistent framework-bathymetry, boundary and initial conditions, meteorological forcing functions interpolated from a common data set-. While the responses of the models are clearly qualitatively similar, large quantitative differences do occur. These differences are often of the same order of magnitude as both the ensemble mean and the sensitivity of the individual results to critical parameters. The direct comparison of the results with measurements from the North Sea Project provides a quantification of the model errors for the salinity and temperature distributions. Using the cost function approach, it is shown that the mean errors (for all the models and all seasons) reach about 70% of the natural variability for the temperature and 90% for the salinity. These errors are larger in summer, when a stratification develops over the Central and Northern North Sea, than in winter. No single model parameter (spatial resolution, turbulence closure scheme, model domain, etc.) can explain the different behaviours of the models. (C) 2003 Elsevier B.V. All rights reserved. [less ▲] Detailed reference viewed: 25 (7 ULg) The age as a diagnostic of the dynamics of marine ecosystem modelsDelhez, Eric ; ; in Ocean Dynamics (2004), 54(2), 221-231 The constituent-oriented age theory (CAT) worked out by Delhez et al. (1999) is a flexible tool that can be applied to diagnose complex models. It is shown here how this can be used to quantify the pace ... [more ▼] The constituent-oriented age theory (CAT) worked out by Delhez et al. (1999) is a flexible tool that can be applied to diagnose complex models. It is shown here how this can be used to quantify the pace at which an ecosystem model works. At the cost of the introduction of one additional evolution equation for each compartment of the ecosystem model, the mean age of the biological material forming these compartments can be computed. The information obtained in this way complements the information provided by the concentration data; while the latter measures the standing stocks, the former provides an integrated assessment of the interaction rates and matter fluxes. The benefits of the method are demonstrated with a simple Lotka-Volterra system and a one-dimensional vertical model of the nitrogen cycle in the Ligurian Sea. The theory can be used to study the biological compartments individually or the ecosystem as a whole. In particular, the age is a valuable tool to quantify the overall cycling rate of nitrogen in the food web. [less ▲] Detailed reference viewed: 21 (6 ULg) Residence time in a semi-enclosed domain from the solution of an adjoint problemDelhez, Eric ; ; in Estuarine Coastal & Shelf Science (2004), 61(4), 691-702 The residence time measures the time spent by a water parcel or a pollutant in a given water body and is therefore a widely used concept in environmental studies. While many previous studies rely on ... [more ▼] The residence time measures the time spent by a water parcel or a pollutant in a given water body and is therefore a widely used concept in environmental studies. While many previous studies rely on severe hypotheses (assuming stationarity of the flow and/or neglecting diffusion) to evaluate the residence time, the paper introduces a general method for computing the residence time and/or the mean residence time without such simplifying hypotheses. The method is based on the resolution of an adjoint advection-diffusion problem and is therefore primarily meant to be used with numerical models. The method and its implications are first introduced using a simplified one-dimensional analytical model. The approach is then applied to the diagnostic of the three-dimensional circulation on the Northwest European Continental Shelf. (C) 2004 Elsevier Ltd. All rights reserved. [less ▲] Detailed reference viewed: 38 (13 ULg) Symmetry and asymmetry of water ages in a one-dimensional flow; Delhez, Eric ![]() in Journal of Marine Systems (2004), 48(1-4), 61-66 Hall and Haine [J. Mar. Syst., in press] briefly addressed the problem of estimating the age of irreducible fluid elements or water parcels in a one-dimensional flow with constant velocity and diffusivity ... [more ▼] Hall and Haine [J. Mar. Syst., in press] briefly addressed the problem of estimating the age of irreducible fluid elements or water parcels in a one-dimensional flow with constant velocity and diffusivity. Herein further developments are achieved on this subject. The age of every water parcel is set to zero at the moment it passes through the point x = 0, where x is an appropriate space coordinate. As time progresses, the age of the water is seen to increase unboundedly upstream of the point x = 0, and tend to the steady-state value \x/u\ downstream of the point x = 0, where a is the water velocity. By contrast, the age of the water parcels that have touched at least once the point x = 0 is symmetric with respect to the point x = 0 is smaller than the water age at any tune and position, and tends to \x/u\ as time progresses. Asymptotic expansions are derived for large times. (C) 2004 Elsevier B.V. All rights reserved. [less ▲] Detailed reference viewed: 26 (12 ULg) A trust region algorithm for automatic identification of elasto-viscoplastic model parameters in metallurgical finite element modelJeunechamps, Pierre-Paul ; ; Duysinx, Pierre et alin Steel Grips 2 (2004), Supplement Metal Forming 2004 Simulation of manufacturing processes, here metal forming, has made major progresses during the last years. The virtual simulation of the manufacturing process being now quite well established, a natural ... [more ▼] Simulation of manufacturing processes, here metal forming, has made major progresses during the last years. The virtual simulation of the manufacturing process being now quite well established, a natural step ahead consists in trying to find automatic procedures to optimize it. Another perspective is also to take care of the influence of the manufacturing constraints upon the design, which is the long term objective of the present research project. It aims at optimizing the design subject to both service constraints (stiffness and strength) and manufacturing constraints. A preliminary stage is to be able to build high fidelity digital models. Complex models require more and more material parameters for behavior laws (e.g. material constitutive and friction laws), which have to be identified numerically from experimental data. In this material parameter identification process, three steps can be identified. At first, an experimental testing is carried out. The second step consists in building a simulation model of the experiment. Finally, the unknown model parameters are determined to match the experimental data. A standard identification procedure consists in minimizing a given norm (here the Euclidean norm) of the error between the model predictions and the experimental results. Even if the identification problem is generally quasi-unconstrained, it has the same complexity as structural optimization problems because of the highly nonlinear and implicit character of the functions, which is especially amplified by the large deformation simulation analysis. In this paper, we use an approach of elastoplastic calculation by finite elements combined with two optimization algorithms: a Levenberg-Marquardt algorithm, which is rather classical in the literature for solving identification problems (see [2]) and a trust-region one (see [1]), which is a rather novel approach at least for structural problems. The results obtained with these two methods are then compared and discussed on two test cases. The first application is an academic test case to validate the identification method. The second one, the compression of a cylinder, takes into account an actual experiment. In this application, the material is assumed to be elasto-viscoplastic and described by a Norton-Hoff behavior law and an isotropic strain hardening law. The parameters to be identified are the coefficients of the Norton-Hoff law. [less ▲] Detailed reference viewed: 35 (6 ULg) Identification of elastoplastic model in large deformation problemsJeunechamps, Pierre-Paul ; Walmag, Jérôme ; Delhez, Eric et alin Cinquini, Carlo (Ed.) Proceedings of the 5th World Congress of Structural and Multidisciplinary Optimization WCSMO5 (2003, May) This paper reports on preliminary results of parameter identification problems in finite element non-linear analyses such as metal forming simulations. Two approaches are compared. The first one is the ... [more ▼] This paper reports on preliminary results of parameter identification problems in finite element non-linear analyses such as metal forming simulations. Two approaches are compared. The first one is the classic Levenberg-Marquardt algorithm. The second one is a trust-region algorithm based on a quadratic model. The two algorithms are compared on two cases, one of them being an actual experiment. [less ▲] Detailed reference viewed: 56 (10 ULg) Global water and salt budget of the Aral Sea from 1960 to 1991.Sirjacobs, Damien ; Grégoire, Marilaure ; Delhez, Eric et alConference (2003, May) Focusing on the global water and salt budget of the Aral sea from 1960 till 1991, a compilation of data available in the litterature is analysed with simple models and compared to the measured evolution ... [more ▼] Focusing on the global water and salt budget of the Aral sea from 1960 till 1991, a compilation of data available in the litterature is analysed with simple models and compared to the measured evolution of the sea. [less ▲] Detailed reference viewed: 13 (3 ULg) Influence of the Aral Sea negative water balance on its seasonal circulation : use of a 3D hydrodynamic modelSirjacobs, Damien ; Grégoire, Marilaure ; Delhez, Eric et alConference (2003, May) Detailed reference viewed: 10 (2 ULg) |
||