mathLab innovating with mathematics
http://mathlab.sissa.it
enReduced Order Methods for Optimal Control Problems: Application in Environmental Marine Sciences and Engineering
http://mathlab.sissa.it/reduced-order-methods-optimal-control-problems-application-environmental-marine-sciences-and
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2017-04-20T11:30:00+02:00">Thursday, 20 April, 2017 - 11:30</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><div><strong>Speaker:</strong> Ms Maria Strazzullo, University of Trieste and SISSA, MSc Mathematics</div>
<div><strong>Room:</strong> A-134</div>
<div><strong>Place:</strong> SISSA campus, via Bonomea 265, 34136 Trieste</div>
<div><strong>Time: </strong>11:30 am</div>
<div> </div>
<div>
<div><strong>Abstract: </strong>The talk aims at introducing reduced order methods in optimal control problems governed by parametrized partial differential equations and at applying them in environmental marine applications. Computationally, optimal control problems are very demanding: a numerical method capable to reduce their dimensionality turns out to be an indispensable tool when several physical and geometrical configurations are involved, as in the case of natural science and, specifically, marine science. For this reason, reduced basis methods have been chosen as resolution strategy: they allow parametric optimal control problems to be solved in a rapid and accurate way. We have focused our analysis on steady optimal control problems characterized by parametric quadratic cost functional constrained to linear parametric partial differential equations. We have recast them in a saddle-point formulation in order to exploit the consolidated knowledge of this kind of structure. A Galerkin-POD algorithm have been applied to several test cases and to some numerical examples in the field of environmental marine sciences and engineering. Two explicative applications are proposed: a large scale climatological application and a small scale pollutant control in the Gulf of Trieste. The first one is inserted in forecasting modeling and data assimilation context, the second deals with the safeguard of Gulf of Trieste and surrounding areas.</div>
</div>
<div> </div>
<div> </div>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Thu, 20 Apr 2017 07:45:40 +0000giovannistabile355 at http://mathlab.sissa.itA multiphase model suitable for the numerical simulation of ice-production in water
http://mathlab.sissa.it/multiphase-model-suitable-numerical-simulation-ice-production-water
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2017-04-12T14:00:00+02:00">Wednesday, 12 April, 2017 - 14:00</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><div>- Speaker: Dr Vanessa Covello, Politecnico di Milano, </div>
<div>- Time 2:00 pm, Wednesday, April 12, 2017</div>
<div>- Place: Room A-133. SISSA main campus, Via Bonomea 265, Trieste</div>
<div> </div>
<div>
<div>
<div>Abstract: Is is well known that ice production is a complex phenomenon, that greatly affects the geophysical, biological and convective processes occurring in the oceans and sea. Moreover, the presence of ice particles in rivers could produce serious damages to hydroelectric facilities, such as the blocking of turbine intakes, or the blockage of hydroelectric reservoirs. Up to now the mathematical and numerical models for the simulation of ice formation</div>
<div>have been mainly focused on the small ice-concentration regimes, according to the Boussinesq approximation. In this talk we present a multiphase model able to describe all the stages of ice production, overcoming the limitation of the Boussinesq approximation. The model is able to reproduce the interaction phenomena occurring between phases in large ice-concentration<br />
regimes, by means of a sophisticated modelling approach, typically adopted for the numerical simulation of multiphase flow of industrial interest. The mixture ice-seawater has been considered as a dense compressible fluid, the behaviour of the liquid phase is indeed modelled by means of an equation of state that links water density to temperature, salinity and pressure. The multiphase equation have been discretized using the finite volume method, and the development of the multiphase solver by means of the OpenFOAM tool is currently ongoing.</div>
<div> </div>
</div>
</div>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Mon, 10 Apr 2017 19:35:34 +0000giovannistabile354 at http://mathlab.sissa.itCurrent research at the Innovative Materials Laboratory at CIRA
http://mathlab.sissa.it/current-research-innovative-materials-laboratory-cira
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2017-03-29T14:00:00+02:00">Wednesday, 29 March, 2017 - 14:00</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><p class="rtejustify"><strong>Speaker</strong>: Andrea Vigliotti, Innovative Materials Laboratory, Italian Aerospace Research Centre</p>
<p class="rtejustify"><strong>Room</strong>: A-134, SISSA Campus, Via Bonomea 265, Trieste</p>
<div>
<div class="rtejustify"><strong>Abstract: </strong>The Italian Aerospace Research Centre is a state funded research centre located in Capua in the south of Italy. The centre was established in 1986 for implementing the National Aerospace Research Program (PRORA). This talk will provide an overview of the current research activities of the Innovative Material Laboratory, with special focus on the following topics:
<ul><li><u>Mechanical Metamaterials</u>. Current additive manufacturing techniques allow the fabrication of components with a prescribed microstructure, which can also include hierarchies and permits to attain performance that cannot be equalled by conventional materials. Specialised modelling tools are necessary for exploiting the potential of these metamaterials, and to understand their limitations. A non-linear numerical homogenization technique will be presented along with recent experimental tests on columns with regular octet microstructure manufactured in Ti-6Al-4V by Electron Beam Melting (EBM);</li>
<li><u>Bayesian inference as a tool for structural mechanics problems.</u> Bayesian inference is a well-known tool for statistical analysis and model testing that has proved to be very successful in many diverse fields. Alike all statistical methods it needs large dataset in order to be effective. Nevertheless, the computational power ordinarily available in modern computers, together with nested sampling techniques, allows the generation of optimized datasets that can be used for the solution of structural mechanics problems, such as the identification of material properties and shape selection. Differently from classical optimization techniques, Bayesian inference permits a thorough analysis of the problem’s phase-space, and provide greater confidence on the representativeness of the chosen solution. An example of the application of these techniques for the identification of the microstructure of a structural component, given limited, noisy, external information will be presented.</li>
</ul></div>
</div>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Wed, 15 Mar 2017 19:13:55 +0000Giovanni Noselli353 at http://mathlab.sissa.itFluid/structure interaction for the flow in the left heart ventricle coupled with natural and prosthetic mitral valves
http://mathlab.sissa.it/fluidstructure-interaction-flow-left-heart-ventricle-coupled-natural-and-prosthetic-mitral-valves
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2017-04-10T15:00:00+02:00">Monday, 10 April, 2017 - 15:00</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><p class="rtejustify"><strong>Speaker</strong>: Valentina Meschini, GSSI-Gran Sasso Science Instutute</p>
<p class="rtejustify"><strong>Room</strong>: A-133, SISSA Campus, Via Bonomea 265, Trieste</p>
<div>
<div class="rtejustify"><strong>Abstract: </strong>The human heart is made of two separate volumetric pumps, the right and the left, the latter being the strongest since it feeds the systemic circulation that brings oxygenated blood to the whole body. Accordingly, the left part has to withstand the largest pressure differences that are in the range 1.6-2.1 x 10^4 Pa (120-160 mmHg). As a consequence, the aortic and mitral valves, that ensure the correct flow direction and prevent blood regurgitation from the aorta to the ventricle and from the ventricle to the atrium, respectively, are the most subjected to damage and impairing. Although several surgical procedures are available to repair and remodel the natural valves, in some cases their replacement is unavoidable and chosing the optimal prosthesis is crucial. Worldwide 280000 valve replacements are performed each year and this number is constantly increasing with a projection of about 800000 by 2050 owing to the increasing age of the population and a growing percentage of it accessing advanced medical care. Aortic and mitral valve replacement are almost equally distributed. However, while the former has already been the topic of extensive medical and scientific research, the latter has been less explored and its post-operative effects on the left ventricle dynamics still need to be analysed in details. In this paper the structure and dynamics of the flow in the left heart ventricle are studied for different pumping efficiencies and mitral valve types (natural, biological and mechanical prosthetic). The problem is investigated by direct numerical simulation of the Navier-Stokes equations, with fluid/structure interaction for the ventricle and mitral valve dynamics. The solver is preliminarily validated by comparisons with ad hoc experiments and then used for production runs. It is found that the left ventricular flow is heavily affected by the specific type of the mitral valve and the effects are more pronounced for ventricles with reduced pumping efficiency. More in details, when the ejection fraction of the ventricle (ratio of the ejected fluid volume and maximum ventricle volume over the cycle) is within the physiological range (EF = 50-70%), regardless of the mitral valve geometry, the mitral jet sweeps the inner ventricle surface up to the apex thus preventing the undesired flow stagnation. In contrast, for pathological ejection fractions (40%) the flow disturbances introduced by the prosthetic devices reduce the penetration capability of the mitral jet and weaken the recirculation in the ventricular apex. This is especially true for the bileaflet mechanical valve, whose disturbances on the mitral flow are the strongest and a region of stagnant fluid is produced. These findings have important clinical implications on the choice of the prosthetic devices in patients that need mitral valve replacement.</div>
<div class="rtejustify"> </div>
<div class="rtejustify"> </div>
</div>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Wed, 15 Mar 2017 09:04:51 +0000Giovanni Noselli352 at http://mathlab.sissa.itThe Geometry and Mechanics of Growing Elastic Sheets
http://mathlab.sissa.it/geometry-and-mechanics-growing-elastic-sheets
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2017-04-13T14:30:00+02:00">Thursday, 13 April, 2017 - 14:30</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><p class="rtejustify"><strong>Speaker</strong>: Prof Eran Sharon,<span style="line-height:1.6em"> Racah Institute of Physics, </span><span style="line-height:1.6em">The Hebrew University of Jerusalem</span></p>
<p class="rtejustify"><strong style="line-height:1.6em">Room</strong><span style="line-height:1.6em">: A-005, SISSA Campus, Via Bonomea 265, Trieste</span></p>
<div>
<div>
<div class="rtejustify"><strong>Abstract: </strong><span style="line-height:1.6em; text-align:justify">Non uniform growth of thin sheets can lead to the formation of elaborate three-dimensional configurations and to induce non trivial shape transformations. In particular, complicated configurations appear in thin sheets when growth leads to geometrical frustration, as often occurs in biological tissues. </span><span style="line-height:1.6em; text-align:justify">I will present examples of different types of systems and discus different types of self shaping principles, together with the theoretical framework of incompatible elasticity which is used to study such systems. Experimental methods for the construction of “programmed” responsive sheets will be reviewed and the connection of the topic to shape selection in chemical and biological systems, as well as to design and art, will be presented.</span></div>
</div>
</div>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Mon, 13 Mar 2017 18:04:25 +0000Giovanni Noselli351 at http://mathlab.sissa.itReduced Basis Approximations for Electromagnetic Applications
http://mathlab.sissa.it/reduced-basis-approximations-electromagnetic-applications
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2016-12-07T11:00:00+01:00">Wednesday, 7 December, 2016 - 11:00</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><p><strong>Speaker:</strong> Martin Hess (Former MPI Magdeburg and now post-doc at SISSA within the AROMA-CFD ERC project)</p>
<p><strong>Venue: </strong>Room 004, ground floor, SISSA main campus, Via Bonomea 265. Trieste</p>
<p><strong>Abstract:</strong></p>
<p>This talk deals with the application of the reduced basis method to microwave systems governed by time-harmonic Maxwell's equations. Several issues are adressed, such as rigorous error estimation, stability preservation and stability constant estimation. Additionally, model reduction results on a dispersive model of the Maxwell-Debye type are presented.</p>
<div> </div>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Tue, 06 Dec 2016 12:12:50 +0000giovannistabile347 at http://mathlab.sissa.itNumerical modelling and simulations of nonlinear elasto-plasticity for geodynamics applications
http://mathlab.sissa.it/numerical-modelling-and-simulations-nonlinear-elasto-plasticity-geodynamics-applications
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2016-12-12T17:30:00+01:00">Monday, 12 December, 2016 - 17:30</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><p><strong>Speaker:</strong> Mr Roberto Porcù, Politecnico di Milano, Department of Mathematics, MOX</p>
<p><strong>Room:</strong> A-005, SISSA Campus, Via Bonomea 265, Trieste</p>
<div><strong>Abstract:</strong></div>
<div class="rtejustify">The aim of this work is the numerical simulation of the Earth's lithosphere extension, with a particular focus on continental rifting that is a tectonical process which has a deep influence on other geodynamical phenomena and on the spatial distribution of some important natural resources. During last decades a large amount of research concerning this topic has been performed by means of analogical and numerical models, named sandboxes. This work is motivated by the observation of several issues inherent to the analogic experiments and by the generally accepted adoption, in the known literature, of purely linear rheologies for numerical tests, which are clearly incoherent with the natural finite-strain deformation regime of the problem at hand. In this work a fully-nonlinear elasto-plastic rheological approach has been developed and implemented. The variables of the problem are parametrized by material coordinates thus a totally-Lagrangian formulation has been adopted together with a quasi-static motion assumption justified by the analysis of the dimensionless quantities of the governing equations. In order to avoid the volumetric locking arising from the local strong enforcement of a nearly incompressible deformation, a three-fields weak variational formulation of the problem has been taken into account, as discussed in [Simo1998]: this method consists in the weak enforcement of the volumetric constraint by introduction of a new independent dilatation variable together with its dual variable that can be interpreted as the Kirchhoff pressure. This leads to an additive decomposition of the Helmoltz free energy into its isochoric and volumetric parts. Coherently to the fully-nonlinear rheological approach the deformation gradient is multiplicatively decomposed into its elastic and plastic parts, according to the Kroner-Lee decomposition. The global nonlinear system is solved by application</div>
<div class="rtejustify">of the Newton-Raphson linearization method. Then the constitutive update, which is computed point-wise on the quadrature nodes, is carried out firstly determining an elastic trial state; if the trial state is nonadmissible, the exponential return map algorithm is performed in order to recover admissibility. The nonlinear return map scheme is solved by means of the Newton-Raphson linearization method. A Drucker-Prager yield criterion and a non-associative flow rule have been considered since they represent really good choices for geological applications, as stated in [BelytschkoLiu2001]. A new C++</div>
<div class="rtejustify">MPI-parallel code has been implemented; it is based on the deal.II finite element library; it handles three-dimensional multi-layered domains. The results are encouraging since they show the natural development of V-shaped shear bands which is a deformation pattern that the most-diffused softwares in the geodynamics field are able to catch only if a weak seed is artificially added into the domain.</div>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Sat, 03 Dec 2016 21:04:33 +0000giovannistabile346 at http://mathlab.sissa.itDynamics and instability of free interface flows: applications to industrial processes
http://mathlab.sissa.it/dynamics-and-instability-free-interface-flows-applications-industrial-processes
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2016-12-12T18:15:00+01:00">Monday, 12 December, 2016 - 18:15</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><p><strong>Speaker</strong>: Dr Michele Girfoglio, University of Naples “Federico II"</p>
<div>
<div><strong>Room</strong>: A-005, SISSA Campus, Via Bonomea 265, Trieste</div>
<div> </div>
<div>
<div><strong>Abstract:</strong></div>
<div>
<div class="rtejustify">Free surface flows occur frequently in many industrial processes such as casting or injection. Within the framework of complex flows with liquid-gas free surfaces, the control of the break up of liquid jets (both plane and annular) plays a key role for the final quality of the product. Vertical liquid curtains are often used in technological processes, classical examples being coating technology and papermaking. In this talk, the unsteady global dynamics of a gravitational liquid sheet interacting with a one-sided adjacent air enclosure (commonly referred to as nappe oscillation configuration) will be addressed, under the assumptions of potential flow and presence of surface tension effects. It will be shown that the unsteady dynamics can be described by means of a typical integro-differential equation of hyperbolic type, exhibiting two distinct characteristic curves. An interesting re-formulation of the nappe global behavior as a driven damped spring-mass oscillator will be addressed. A relevant procedure, which has been employed in order to remove the singularity exhibited by the integro-differential operator for Weber number less than unit, will be presented. The investigation has been carried out by means of a modal (i.e., time asymptotic) linear approach, which is corroborated by direct numerical simulations of the governing equation. In the final part of the seminar it will be briefly illustrated the numerical investigation of a slag-steel interface surface in a continuous casting process.</div>
<div> </div>
</div>
</div>
<div> </div>
<div> </div>
</div>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Sat, 03 Dec 2016 20:59:13 +0000giovannistabile345 at http://mathlab.sissa.itNatural circulation with distributed heat sources: a theoretical and experimental investigation
http://mathlab.sissa.it/natural-circulation-distributed-heat-sources-theoretical-and-experimental-investigation
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2016-11-29T13:00:00+01:00">Tuesday, 29 November, 2016 - 13:00</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><p><strong>Speaker: </strong>Alessandro Pini (Politecnico di Milano, CESNEF )</p>
<p><strong>Venue: </strong>Room 005, ground floor, SISSA main campus, Via Bonomea 265. Trieste</p>
<p><strong>Abstract: </strong></p>
<p>The theme of the research concerns the dynamics of natural circulation loops with distributed heat sources. The dynamic behaviour of this kind of systems is investigated by means of both theoretical (1D/3D modal and non-modal methods) and experimental (ad hoc testing facility) approaches. The study attempts to provide a first systematic analysis of the dynamics of free convection in presence of distributed heating. Attention is paid to the prediction of stable and unstable (oscillating) flow regimes, whose occurrence is determined by several factors (e.g., geometry, heat-exchange features, piping material, and kind of heat source).</p>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Fri, 18 Nov 2016 12:47:04 +0000giovannistabile344 at http://mathlab.sissa.itOn aerodynamic design with a POD surrogate model
http://mathlab.sissa.it/aerodynamic-design-pod-surrogate-model
<div class="section field field-name-field-date-events field-type-datestamp field-label-inline clearfix"><h3 class="field-label">Date: </h3><div class="field-items"><div class="field-item odd"><span class="date-display-single" property="dc:date" datatype="xsd:dateTime" content="2016-11-30T11:30:00+01:00">Wednesday, 30 November, 2016 - 11:30</span></div></div></div><div class="section field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item odd" property="content:encoded"><p><strong>Speaker:</strong> Valentina Dolci (Politecnico di Torino)</p>
<p><strong>Venue: </strong>Room 005, ground floor, SISSA main campus, Via Bonomea 265. Trieste</p>
<p><strong>Abstract:</strong></p>
<p>Since the last few years, surrogate models are becoming a promising research field for engineering applications. A surrogate, or reduced-order model (ROM), is a mathematical tool able to extract the main features of a more computational demanding high-order model, starting from a reduced set of information. Once the surrogate model is built, it can be used to perform faster analyses of the problem. The fields of optimization and database generation in aerodynamics can be best candidates for the application of surrogate models. In the case of aerodynamic optimization, a cost function should be evaluated several times requiring many CFD simulations in order to achieve sufficient information to identify an optimum target. This operation requires a great amount of computational time and effort and the adoption of a surrogate model replacing the CFD high-order model can be attractive. However the construction of an accurate and robust surrogate model is a delicate process. Particular attention should be paid to the type of surrogate model chosen for a specific problem and to the number and position of the initial set of high-order simulations. In this construction phase there is a constant trade-off between required accuracy and computational effort reduction: the building of the surrogate model must be fast and not computational demanding preserving in the meantime the primary characteristics of the problem. In the present work surrogate models are constructed using the proper orthogonal decomposition technique applied in the parameter space. A reduced snapshot set is used, adopting full and fractional factorial planes together with quadtree distribution for the initial positioning of the snapshots. Response surface methodology is employed to compute the POD coefficients. Surrogate models are applied in the analysis of transonic flows, three-dimensional aircraft configurations and shape optimization of vehicles. A posteriori error estimates have been performed and the models showed good agreement with the CFD reference solution.</p>
</div></div></div>
<div class="field field-name-field-category field-type-taxonomy-term-reference field-label-above clearfix">
<p class="field-label">Category: </p>
<ul class="field-items">
<li class="field-item even">
<a href="/events/seminar" typeof="skos:Concept" property="rdfs:label skos:prefLabel" datatype="">Seminar</a> </li>
</ul>
</div>
Fri, 18 Nov 2016 11:17:03 +0000giovannistabile343 at http://mathlab.sissa.it