Lista de las tesis 2012 de ISAE Noviembre de 2012 Title: Unsteady jets and wakes numerical simulation within

aerodynamic shapes design Author : Giner, Pierre Date: 15/05/2012 Research director(s): Aupoix, Bertrand Laboratory: Dpartement d'Arodynamique, Energtique et Propulsion (DAEP) Specialization: Mcanique, Energtique, Gnie civil et Procds (MEGeP) Classification: Physics Abstract: The aerodynamic integration of Ultra-High Bypass Ratio turbofans raises the need for an accurate prediction of the unsteady properties of the jet development using Computational Fluid Dynamics, since the design stages. The ability of numerical methods in predicting these phenomena are assessed in this thesis, using different modeling approaches. A dual-stream jet configuration is investigated, using an associated wind-tunnel test campaign. Reynolds-Averaged Navier-Stokes simulations, using several turbulence models, are shown to correctly reproduce the mean flow especially concerning the shear layers. Shock cells and turbulence levels predictions within the primary and secondary jet flows are however perfectible compared to the test results. Turbulence-accounting approaches based upon experimental data and using different configurations partially overcome this issue. An unsteady method is then applied in order to resolve turbulent phenomena instead of modeling them. The hybrid Detached-Eddy Simulation of the flow demonstrates an at least equivalent accuracy concerning the mean flow and provides additional information on the fluctuating fields and shear layers unsteady properties, in fair agreement with the experimental results. Prospects discuss benefits and consequences of this approach taking into account its cost and the industrial context of this application, as well as its potential use for a more complex, installed jet configuration. Title: Aerothermodynamic and electrical study of a Plasma Synthetic Jet actuator for flow control Author: Hardy, Pierrick Date: 09/05/2012 Research director(s): Cambronne, Jean-Pascal; Millan, Pierre Laboratory: Dpartement d'Arodynamique, Energtique et Propulsion (DAEP) Specialization: Aronautique - Astronautique (AA) Classification: Physics Abstract: Improvement of aerodynamics and environmental performances is a major issue for terrestrial and aeronautical industry. For fulfilling increasing demand, one of the answers is flow control. To achieve flow control, high performance actuators are needed. An innovative technique called Plasma Synthetic Jet actuator consists on applying an electrical discharge in a small cavity. Plasma is created and increases gas temperature and pressure which results on the creation of a micro-jet through cavity opening. The PhD objectives are to develop the PSJ actuator, to describe

actuator mechanisms and to apply it for controlling noise of a high subsonic jet. The first part of the study consists on defining flow control needs and on developing a PSJ actuator prototype. Then, actuator performances are characterized using electrical measurements of the discharge and using aerodynamic measurements. These measurements show that an electrical model of the discharge is needed. Based on the Braginskii model, a simple model is carried out and is applied to the actuator. Efficiency of the PSJ is deduced. The modified Braginskii model is then coupled with an URANS model to achieve frequency modeling of the actuator. Results match aerodynamics measurements. PSJ actuators are applied for controlling jet noise in a second part of the study. Schlieren visualizations are used to show micro-jet interaction with the main jet. Acoustic measurements are then performed and show that the PSJ is a good actuator to control high subsonic jet noise. Title: Multiobjective robust optimization via surrogate models Author: Baudoui, Vincent Date: 07/03/2012 Research director(s): Hiriart-Urruty, Jean-Baptiste; Klotz, Patricia Laboratory: Dpartement de Mathmatiques, Informatique, Automatique (DMIA) Specialization: Aronautique - Astronautique (AA) Classification: Mathmatiques Abstract: This PhD thesis deals with the optimization under uncertainty of expensive functions in the context of aeronautical systems design. First, we develop a multiobjective robust optimization strategy based on surrogate models. Beyond providing a faster representation of the initial functions, these models facilitate the computation of the solutions' robustness with respect to the problem uncertainties. The modeling error is controlled through a new design of experiments enrichment approach that allows improving several models concurrently in the possibly optimal regions of the search space. This strategy is applied to the pollutant emission minimization of a turbomachine combustion chamber whose injectors can clog unpredictably. We subsequently present a heuristic method dedicated to multidisciplinary robust optimization. It relies on local robustness management within disciplines exposed to uncertain parameters, in order to avoid the implementation of a full uncertainty propagation through the system. An applicability criterion is proposed to check the validity of this approach a posteriori using data collected during the optimization. This methodology is applied to an aircraft design case where the surface of the vertical tail is not known accurately. Title: Experimental and numerical study of aircraft engine ignition in high altitude conditions Author: Linassier, Guillaume Date: 03/05/2012 Research director(s): Lavergne, Grard; Lederlin, Thomas Laboratory: Dpartement d'Arodynamique, Energtique et Propulsion (DAEP) Specialization: Mcanique, Energtique, Gnie civil et Procds (MEGeP) Classification: Physics

Abstract: Design of aircraft engines requires tests on engine benches. These tests allow characterizing combustor ignition limits, but are extremely expensive and time consuming. In order to limit their number, it is necessary to develop alternative methods enabling to predict the ignition performances of a combustor prototype, for both ground conditions and high altitude conditions, the latter being particularly critical. The purpose of this thesis is to contribute to the development and validation of numerical tools enabling to predict ignition performances of combustor. Validation will be possible using an experimental data base obtained on the MERCATO test bench, at ONERA. This work results from a close cooperation with TURBOMECA. A numerical model, previously developed to predict the ignition of fuel spray following a spark discharge, has been improved. This model has been combined to a multiphysics CFD code (CEDRE) in order to build ignition map from a mean two-phase flow field, and also to compute the flame propagation stage using RANS approach. In order to validate both methods, experimental characterization of a one-sector combustor has been performed on the MERCATO test rig. A complete data base for validation of CFD code is now available. Comparison of experimental and numerical ignition mapping showed good agreements for the different tests cases, and seems encouraging for an application on an industrial combustor. In the same time, promising results have been obtained with a new experimental three-sector combustor. Its geometry is based on an industrial combustor, allowing a comparison of ignition performances between the simplified and the industrial combustors. Title: Optimization of telecommunication satellite payload test plans Author: Maillet, Caroline Date: 25/04/2012 Research director(s): Verfaillie, Grard; Cabon, Bertrand Laboratory: Dpartement de Mathmatiques, Informatique, Automatique (DMIA) Specialization: Mathmatiques, Informatique et Tlcommunications de Toulouse (MITT) Classification: Informatique Abstract: Telecommunication satellite payload validation requires operations which are expensive in terms of time and manpower. This cost is constantly increasing as the payloads become more and more complex. It is crucial for Astrium to optimise the testing phase to keep these costs under control. The objective of this CIFRE thesis, conducted in collaboration with Astrium and Onera, is to develop a software suite to help generate test plans for the payloads. The problem of generating test plans was modeled using the form of a directed graph with states. The NPcompleteness of this problem was proven. Mathematical models were built using integer linear programming and constraint programming with a view to solving the problems using generic solvers. However, these generic solvers had problems due to insufficient memory on account of the large size of instances to be handled. These problems led us to develop a specialised solver using a tree search, with special mechanisms for choosing variables and values, propagating constraints, computing bounds, backjumping, learning, and restarting. A specialised solver based on local search was developed in parallel. The results obtained by these different solvers with different settings were compared.

Title: Experimental and Numerical Analysis of a Shrouded Contrarotating Coaxial Rotor in Hover Author: Huo, Chao Date: 26/03/2012 Research director(s): Moschetta, Jean-Marc; Gressier, Jrmie Laboratory: Specialization: Mcanique, Energtique, Gnie civil et Procds (MEGeP) Classification: Mathmatiques Abstract: This study aims to analyze the behavior of shrouded, contrarotating coaxial rotor in the reduced MAVs scale in order to exploit its potential to improve the free rotor steady performance. The high hover ability under low operational Reynolds number is therefore, a scientific challenge. Generally, comparing with free rotor, the addition of the shroud decreases the flow contraction and gives the potential to generate an extra thrust. A suitable nozzle can control the mass flow for a given power. The increased mass flow, comparing with free rotor, amplifies the thrust offered by the low pressure formed at the air entrance. To understand the principals of shrouded propulsion system, a simplified theory model was first proposed through the extension of Froude theory for free rotors: the double rotor is initially treated as an actuator disk, generating the flow at varied sections through the shroud passage. A 2D simulation which accounts for an axial flow of viscous effects within the actual shroud profile, confirmed effects of all defined geometrical parameters. It further demonstrated that within the non-stalling region of the different crosssections, shroud shape and inlet shape do not have a significant impact on performance. The experimental study, carried out with coaxial rotor, contributed to the confirmation of the overall performance and the approximation of the flow field through the shroud. Meanwhile, the 3D simulation, developed to better model the actual coaxial rotor in counter rotation, was validated to well solve the steady performance. It was applied to complement the analysis of the flow around the coaxial rotor. Title: Controlling communication latencies in on-board SpaceWire networks Author: Ferrandiz, Thomas Date: 02/03/2012 Research director(s): Fraboul, Christian Laboratory: Dpartement de Mathmatiques, Informatique, Automatique (DMIA) Specialization: Mathmatiques, Informatique et Tlcommunications de Toulouse (MITT) Classification: Informatique Abstract: The SpaceWire network standard is promoted by the ESA and is scheduled to be used as the sole onboard network for future satellites. SpaceWire uses a wormhole routing mechanism to reduce memory consumption and the associated costs. However, wormhole routing can lead to packet blocking in routers which creates large variations in end-to-end delays. As the network will be shared by real-time and non real-time traffic, network designers require a tool to check that temporal constraints are verified for all the critical messages. The metric we chose for this tool is an upper-bound on the worst-case end-to-end delay of a packet traversing a SpaceWire network. This metric is simpler to compute than the exact delay of each packet and provide enough

guarantees to the network designers. During the thesis, we designed three methods to compute this upper-bound. The three methods use different assumptions and have different advantages and drawbacks. On the one hand, the first two methods are very general and do not require strong assumptions on the input traffic. On the other hand, the third method requires more specific assumptions on the input traffic. Thus, it is less general but usually gives tighter bounds than the two other methods. In the thesis, we apply those methods to a case study provided by Thales Alenia Space and compare the results. We also compare the three methods on several smaller networks to study the impact of various parameters on their results. Title: Array processing for GNSS receiver in urban environement [i.e environment] Author: Rougerie, Sbastien Date: 31/01/2012 Research director(s): Vincent, Franois Laboratory: Dpartement d'Electronique, Optronique et Signal (DEOS) Specialization: Mathmatiques, Informatique et Tlcommunications de Toulouse (MITT) Classification: Sciences de l'ingnieur Abstract: In Global Navigation Satellite Systems (GNSS) applications, multipath (MP) errors are still one of the major error sources in conventional receivers. The additional signal replicas due to reflections introduce a bias in Delay Lock Loops (DLL), which finally leads to a positioning error. Several techniques have been developed for multipath mitigation or estimation such as the Narrow Correlator Spacing or the Multipath Estimating Delay-Lock-Loop (MEDLL) algorithm. However, these techniques suffers from high sensitivity to noise, and can not mitigate short delay multipath (<0.1 chip). More recently, the use of antenna array algorithms has been proposed for multipath mitigation. Antenna arrays perform a spatial sampling that makes possible the discrimination of sources in the space domain (azimuth and elevation). However, in conventional receivers, little room remains for antenna integration, and only a small number of antenna elements can be integrated. This study will therefore focus on algorithms for a 2x2 square antenna array. Moreover, the proposed solutions have to be robust against technological defects. Two solutions are investigated to mitigate multipath with an antenna array. The first one tries to filter the multipaths in the space domain in order to "clean" the incoming signal of all the multipaths. However, the results obtained with this solution are quite mitigated. Indeed, the small size of the array implies a low space resolution, and the correlation between the LOS signal and the multipaths strongly degrade the performance of high resolution algorithms. Thus, close spaced multipath are still a problem. In the second approach, a set of parameters (amplitudes, timesdelays, Doppler shifts, elevations and azimuths) of all the incoming sources are estimated. The main difference with the first approach is that, instead of filtering the sources on the space domain only, the different incoming paths are filtered on space, time and frequency domains. To estimate the parameters of all the sources, SAGE algorithm, which is a low-complexity generalization of maximum likelihood theory, has been considered. Moreover, a new implementation of the SAGE algorithm has been investigated in order to reduce the complexity by a factor 500, without loss of estimation performances. The simulations show a real improvement in the multipath mitigation compared to mono antenna algorithms and beamformer approaches.

The impact of technological defects (mutual coupling, RF channel mismatch ), numerical defect (quantization) and SAGE defect (estimation of the number of path) on the estimation performances were also investigated, and several compensation algorithms were proposed. The wide band effects of the RF filter were compensated by FIR equalizer, and mutual coupling can be estimated thanks to the knowledge of the satellites DOA. Simulations show that the estimation performance of the SAGE algorithm after array calibration are very close than the performance in perfect system. Last, we proposed an algorithm to estimate the number of path, and dynamic simulations (by using channel model) show a very good adaptation of the algorithm. Last but not least, the tools developed in this PhD can be also useful in multipath modeling applications for GNSS. The main difference with the first approach is that, instead of filtering the sources on the space domain only, the different incoming paths are filtered on space, time and frequency domains. To estimate the parameters of all the sources, SAGE algorithm, which is a low-complexity generalization of maximum likelihood theory, has been considered. Moreover, a new implementation of the SAGE algorithm has been investigated in order to reduce the complexity by a factor 500, without loss of estimation performances. The simulations show a real improvement in the multipath mitigation compared to mono antenna algorithms and beamformer approaches. The impact of technological defects (mutual coupling, RF channel mismatch ), numerical defect (quantization) and SAGE defect (estimation of the number of path) on the estimation performances were also investigated, and several compensation algorithms were proposed. The wide band effects of the RF filter were compensated by FIR equalizer, and mutual coupling can be estimated thanks to the knowledge of the satellites DOA. Simulations show that the estimation performance of the SAGE algorithm after array calibration are very close than the performance in perfect system. Last, we proposed an algorithm to estimate the number of path, and dynamic simulations (by using channel model) show a very good adaptation of the algorithm. Last but not least, the tools developed in this PhD can be also useful in multipath modeling applications for GNSS. Title: tude des architectures de scurit de systmes autonomes. Formalisation et valuation en Event B Author: Chaudemar, Jean-Charles Date: 27/01/2012 Research director(s): Seguin, Christel; Bensana, Eric Laboratory: Specialization: Systmes (EdSys) Classification: Informatique Abstract: The study of complex system safety requires a rigorous design process. The context of this work is the formal modeling of fault tolerant autonomous control systems. The first objective has been to provide a formal specification of a generic layered architecture that covers all the main activities of control system and implement safety mechanisms. The second objective has been to provide tools and a method to qualitatively assess safety requirements. The formal framework of modeling and assessment relies on Event-B formalism. The proposed Event-B modeling is original because it takes into account exchanges and relations between architecture

layers by means of refinement. Safety requirements are first specified with invariants and theorems. The meeting of these requirements depends on intrinsic properties described with axioms. The proofs that the concept of the proposed architecture meets the specified safety requirements were discharged with the proof tools of the Rodin platform. All the functional properties and the properties relating to fault tolerant mechanisms improve the relevance of the adopted Event-B modeling for safety analysis. Then, this approach is implemented on a study case of ONERA UAV. Title: Real-time distributed simulation architecture Author: Chaudron, Jean-Baptiste Date: 25/01/2012 Research director(s): Siron, Pierre; Adelantado, Martin Laboratory: Dpartement de Mathmatiques, Informatique, Automatique (DMIA) Specialization: Mathmatiques, Informatique et Tlcommunications de Toulouse (MITT) Classification: Informatique Abstract: This work takes place in the global project PRISE (Plate-forme de Recherche pour l'Ingnierie des Systmes Embarqus) in which the focus is to develop an execution platform for embedded software. Embedded software are said criticals and, therefore, are subject to specific design rules. Particularly, these software must meet real time constraints and thus ensure a temporal predictive behaviour in order to always give accurate results with respect to corresponding timing deadlines. The main objective of this thesis is to study the use of distributed simulation techniques (and specifically the HLA standard) to meet the real-time and hybrid simulation needs of the PRISE platform. To comply with these real-time constraints and ensure the predictability of a distributed simulation, we must have a complete view of the whole problem and in particular the different levels of action: application, middleware, software, hardware and also a formal level for validation of the timing behaviour. This work is based on the RTI (Run Time Infrastructure, HLA middleware) from ONERA laboratory called: the CERTI and proposes a methodological approach adapted to take into account these different levels of action. Some case studies, including a flight simulator of an aircraft, have been specified, implemented and tested on the PRISE platform. Title: Aircraft trajectory optimization considering noise for flight management) Author: Le Merrer, Mathieu Date: 18/01/2012 Research director(s): Farges, Jean-Loup; Seren, Cdric Laboratory: Dpartement de Mathmatiques, Informatique, Automatique (DMIA) Specialization: Systmes (EdSys) Classification: Sciences de l'ingnieur Abstract: Forthcoming environmental challenges stimulate the development of trajectory optimization methods by aeronautical actors. This contribution consists in three parts. First, several trajectory optimization techniques are compared. The comparison is based on a simple academic problem. After that, a model is proposed for considering noise nuisance level in the

framework of trajectory optimization. Finally, the optimization problem of an ascent phase of a civil aircraft is solved using a direct approach. The specific issues of the problem are tackled with a general formulation. They consist in the presence of several phases along the trajectory, running state equality constraints and tough numerical integration of the noise model.