Dans un contexte de transition énergétique et numérique, L’Usine du Futur se définit commeéconome en énergie, mais aussi agile grâce à des moyens de production flexibles etreconfigurables. Conférer ces propriétés aux procédés existants est un challenge complexe quiinduit souvent une réorganisation des unités. Dans ce cadre, la méthodologie RREFlex a pourobjectif de proposer des solutions alternatives d’intégration énergétique à la fois viables, robusteset adaptables via le remodelage des réseaux d’échangeurs de chaleur déjà installés.Contrairement à la conception initiale consistant à réaliser de manière conjointe la conception duprocédé et du réseau d’échangeur de chaleur associé, l’exercice de remodelage d’unitésexistantes peut s’avérer sensiblement plus complexe. En effet, il s’inscrit souvent dans unedémarche d’amélioration continue sur des installations ayant déjà connu au cours de leur vie destransformations pour faire face aux évolutions de la demande ou à de nouvelles contraintesenvironnementales. Aujourd’hui, nombre d’analyses de récupération énergétique sont réaliséessur les sites industriels mais celles-ci ne donnent pas nécessairement lieu à des réalisationsconcrètes. Les raisons le plus souvent invoquées sont d’ordre financières mais aussi, pratiques.En effet, les solutions proposées ne sont pas toujours réalistes d’un point de vue opérationnellecar elles tiennent rarement compte de la variabilité du procédé, soit due à des perturbations nonmaitrisées sur les températures et les débits, soit due à la présence de multiples points defonctionnement (changement de campagne, évolution de la charge, etc.). De plus, ces solutionstiennent peu compte des contraintes spécifiques du site étudié (topologie des unités, compatibilitédes courants, sécurité, etc.), la formulation a priori de ces dernières ne pouvant être exhaustive.L’outil RREFlex - outil Robuste pour la synthèse de Réseaux d’Echangeurs Flexibles - a étédéveloppée pour pallier autant que possible ces limites. S’appuyant sur une analyse statistiquedes historiques de mesures remontés en salle de contrôle, un premier module - EDiFy : EnhancedData collection for Flexibility analysis - permet de localiser et de caractériser les différents cas demarche des unités. Notamment, la valeur moyenne et la variance des données caractéristiques duprocédé (températures, flux calorifiques) sont estimées pour chaque cas de marche. Ce jeu dedonnées étant souvent incomplet, il est nécessaire d’avoir recours à un modèle de simulation duprocédé pour le compléter et valider la cohérence des données mesurées pour chaque cas demarche potentiellement identifié. Sur cette base, une seconde étape dédiée au diagnosticénergétique permet d’évaluer la pertinence des échangeurs de récupération déjà installés. Cetteanalyse permet d’identifier et de prioriser un ensemble de scenarii de remodelage considéréscomme prometteurs. Chacun d’eux est défini par la liste des échangeurs à reconsidérer et uncertain nombre de paramètres de configuration. Chaque scenario retenu est alors exploité pour lasynthèse du réseau d’échangeurs associé. Cette étape s’appuie sur un modèle de programmationlinéaire mixte multi-période (PLM) pour déterminer la nouvelle topologie du réseau d’échangeurs.Dans ce cadre, le modèle a évidemment la possibilité d’introduire de nouveaux échangeurs maisaussi de déplacer les échangeurs existants et conservés dans un scénario donné, tant qu’ilscouplent la même paire de courant qu’initialement. Les réseaux obtenus sont donc adaptables auxdifférents cas de marche identifiés à la première étape et reconfigurable grâce à l’implantation deby-pass. Le panel de réseaux proposés est enfin évalué et classé au moyen d’indicateurs deperformance, dont notamment la robustesse vis-à-vis de la variabilité du procédé. L’approche aété validée sur deux sites de dimension industrielle: un procédé de fabrication de MVC et un trainde préchauffe de pétrole brut / In a context of numerical and energy transition, the Factory of the Future is meant to be moreenergy efficient but also smarter and agile through the use of flexible and reconfigurableproduction means. Enabling existing processes to achieve those properties is a difficult challengewhich often induces a reorganization of the units. In this context, RREFlex methodology wasdeveloped to provide several alternatives heat integration solutions both viable, robust andadaptable through the retrofitting of existing heat exchanger networks. Unlike grass-root design,which consists in designing both the process and the heat exchanger network at the same time(and thus, allowing many possibilities), retrofitting existing units can be a lot more complex.Indeed, as part of a continuous improvement process of the production, the plants have oftenalready undergone transformations during their life to cope with changes in demand or newenvironmental constraints. Currently, numerous energy recovery analysis are performed onindustrial sites but do not necessarily involves concrete industrial measures. The main reasons forthe lack of results are mainly financial but also practical. The provided solutions are often nonrealistic in terms of operability because of the lack of accounting for the variability of the process,whether due to external disturbances on temperatures and flowrates or due to multiple operatingconditions (many production campaigns, evolution in process load, etc.). Moreover, thosesolutions also do not take on-site constraints into account (units topology, process streamscompatibility, safety, etc.), as it is difficult to apprehend such constraints. The RREFlex module(Robust software tool for the synthesis of Flexible Heat Exchanger Networks), was developed toassess these issues. Based on a statistical analysis of historical data extracted from on-sitemeasurements, a first module - EDiFy : Enhanced Data collection for Flexibility analysis – enablesthe location and characterization of the multiple steady state regimes. The mean value andvariance of operating conditions characterizing the process (e.g. temperature, heat flow) areestimated for each steady state. As this data set is usually incomplete, it is necessary to use asimulation model of the process to complete and validate the consistency of the measurements ofeach identified steady state.Based upon those data, an energy diagnosis step enables the assessment of each existing heatexchanger liability. This analysis results in the identification and classification of several promisingretrofitting scenarios. Each one is defined by a list of heat exchangers to reconsider and severalconfiguration parameters.Each selected scenario is then used to design the corresponding optimal heat exchanger network.The latter step, which is based on a multi-period mixed linear programming model, aims at thedesign of a new heat exchanger network topology. In this context, the model includes not only thepossibility to add new heat exchangers but also to shift the preserved heat exchangers for a givenscenario, as long as the original pair of streams is kept. The resulting heat exchanger networksare thus adaptable to every operating conditions identified in the first step of the methodology butalso reconfigurable through the use of by-passes. The performances of the resulting networks areevaluated and classified using key performance indicators, especially the robustness which iscrucial to account for the process variability.The approach was validated on two industrial scale case studies: a MVC production process and arefinery heating train.

A recuperação motora da mão é importante para a independência funcional de pacientes com acidente vascular cerebral (AVC). A estimulação magnética transcraniana (EMT) é uma técnica não invasiva que pode ser usada para avaliar a excitabilidade do córtex motor. O objetivo principal deste estudo foi avaliar a correlação entre as medidas de EMT (inibição intracortical, IIC; facilitação intracortical, FIC; período silente ipsilateral, Psil) e o desempenho motor em atividades relevantes para as atividades de vida diária, em indivíduos com AVC (grupo AVC) e boa recuperação motora, assim como em sujeitos sem AVC (grupo controle). As medidas foram realizadas no hemisfério afetado (HA) dos indivíduos com AVC e no hemisfério homólogo dos sujeitos do grupo controle. Os objetivos secundários foram: confirmar que a as medidas de EMT do grupo AVC seriam comparáveis às do grupo controle; investigar a relação entre as medidas de IIC ou FIC e o Psil nos dois grupos. Após avaliação de elegibilidade de 2298 sujeitos, 12 pacientes (seis homens) foram incluídos no grupo AVC com média (+- desvio-padrão) de idade 52,9 (+- 11,8) anos, pontuações na escala de AVC do National Institutes of Health de 2,6 (+- 1,8) e na subescala de avaliação do membro superior de Fugl-Meyer de 57,5 (+- 4,6). No grupo controle foram incluídos 10 indivíduos (seis homens) com média de idade 53,3 (+- 12,0) anos. Nos dois grupos, foram avaliados: desempenho no teste de Jebsen Taylor, força de preensão, limiar motor, IIC, FIC e Psil. O coeficiente de correlação de Spearman foi utilizado para avaliar a correlação entre medidas de desempenho motor e medidas de EMT, assim como para associações entre medidas de excitabilidade intracortical e Psil. Para comparação das medidas entre o grupo AVC e o grupo controle, foi utilizado o teste de Mann-Whitney. Foi encontrada uma correlação significativa (r = 0,68; p = 0,014) entre a força de preensão e a IIC no Grupo AVC, mas não no grupo controle (r= 0,16; p= 0,47). Não foram encontradas correlações estatisticamente significativas entre o desempenho no teste de Jebsen-Taylor e qualquer outra medida de excitabilidade nos dois grupos. Não houve diferenças estatisticamente significativas entre as medidas de EMT nos dois grupos (p > 0,05). Não houve correlação significativa entre as medidas de excitabilidade intracortical e o Psil no grupo AVC, mas houve correlação significativa entre aumento da IIC e maior duração do Psil no grupo controle (r=-0,8; p= 0,008). A FIC não teve correlação significativa com o Psil em quaisquer dos grupos. Esses resultados sugerem que a atividade gabaérgica no córtex motor primário do HA seja relevante para a força da mão parética em indivíduos com AVC. Além disso, confirmam a hipótese de que pacientes com boa recuperação motora apresentem padrões de excitabilidade cortical semelhante aos de indivíduos saudáveis. Finalmente, os resultados não corroboram a hipótese de que um mesmo grupo de interneurônios medeie a IIC do HA e a inibição inter hemisférica do hemisfério afetado para o hemisfério não afetado em indivíduos com AVC e bom desempenho motor, na fase crônica / Recovery of hand motor function is important for the functional independence of patients with stroke. Transcranial magnetic stimulation (TMS) is a non-invasive technique for assessment of motor cortex excitability. The main objective of this study was to evaluate the correlation between TMS measures (short-interval intracortical inhibition, SICI; intracortical facilitation, ICF and ipsilateral silent period, ISP) and motor performance in subjects with stroke and good motor recovery (stroke group) as well as in subjects without stroke (control group). Measurements were performed in the affected hemisphere of stroke subjects and in the homologous hemisphere of subjects in the control group. The secondary objectives were: to confirm that the SICI, ICF and ISP of the stroke group would be comparable to those of the control group; to investigate the relation between SICI or ICF and the ISP in the two groups. After assessment of eligibility in 2298 subjects, 12 patients (six men) were included in the stroke group. Mean (+- standard deviation) age was 52.9 (+- 11.8) years, National Institutes of Health Stroke Scale score was 2.6 (+- 1.8) and Fugl-Meyer upper limb subscale score was 57.5 (+- 4.6) in the stroke group. In the control group, 10 subjects (six men) were included with a mean age of 53.3 (+- 12.0) years. In the two groups, the Jebsen-Taylor test, grip strength, motor threshold, SICI or ICF and ISP were evaluated. The Spearman\'s correlation coefficient was used to assess the correlation between motor performance measures and TMS measures, as well as the associations between intracortical excitability measures and the ISP. The Mann-Whitney test was used to compare the measures between the stroke group and the control group. A significant correlation (r = 0.68, p = 0.014) was found between grip strength and SICI in the stroke group, but not in the control group (r = 0.16, p = 0.47). No statistically significant correlations were found between performance in the Jebsen-Taylor test and any other measures of excitability in the two groups. There were no statistically significant differences between the TMS measurements between the two groups (p > 0.05). There was no significant correlation between the SICI measures and the ISP in the stroke group, but there was a significant correlation between increased SICI and longer duration of the ISP in the control group (r=-0.8; p=0.008). ICF had no significant correlation with the ISP in any of the groups. These results suggest that increased gabergic activity in the motor cortex of the affected hemisphere is relevant for strength of the paretic hand in patients with stroke. Also, they confirm the hypothesis that patients with good motor recovery present patterns of cortical excitability similar to those of healthy subjects. Finally, the results do not support the hypothesis that the same group of interneurons mediate intracortical inhibition and interhemispheric inhibition of the unaffected hemisphere by the affected hemisphere in subjects with stroke and good motor performance in the chronic phase

Ce travail estime le potentiel des plasmas de striction magnétique (Z-pinches) pour le chauffage de cavités radiatives à haute température (>200eV). Des modèles simples sont fournis pour calculer les performances atteignables avec des courants de 5 à 100 MA en 100 ns. La physique monodimensionnelle à l'échelle de l'épaisseur du plasma et les instabilités hydrodynamiques sont étudiées. Puis l'amélioration des performances des cavités avec une double coquille ou l'installation d'un champ magnétique axial est analysée. L'attaque directe par un Z-pinch d'une cible de fusion par confinement inertiel est enfin considérée. Tous les résultats présentés reposent sur une approche théorique et numérique (bidimensionnelle) et sur l'exploitation de résultats expérimentaux obtenus sur le générateur américain "Z". Les annexes rappellent les équations de la MHD radiative et vérifient leur validité pour les plasmas de striction magnétique.

The objectives of this doctoral thesis are to give a basisincluding methods for the development of stationary polymerelectrolyte fuel cell (PEFC) systems for combined heat andpower production. Moreover, the objectives include identifyingprerequisites, requirements and possibilities for PEFC systemsproducing heat and power for buildings in Sweden. The PEFCsystem is still in a pre-commercial state, but low emissionlevels, fast dynamics and high efficiencies are promisingcharacteristics. A thermodynamic model to simulate stationary PEFC systemshas been constructed and pinch technology and exergy analysesare utilised to design and evaluate the system. The finalsystem configuration implies a high total efficiency ofapproximately 98 % (LHV). A flexible test facility was built in connection with theresearch project to experimentally evaluate small-scalestationary PEFC systems at KTH. The research PEFC system hasextensive measurement equipment, a rigorous control system andallows fuel cell systems from approximately 0.2 to 4 kWel insize to be tested. The simulation models of the fuel processorand the fuel cell stack are verified with experimental datataken from the test facility. The initial evaluation andsimulation of the first residential installation of a PEFCsystem in Sweden is also reported. This PEFC system, fuelled bybiogas and hydrogen, is installed in an energy system alsoincluding a photovoltaic array, an electrolyser and hydrogenstorage. Technical aspects of designing a fuel cell system-basedenergy system, including storages and grid connections, whichprovides heat and power to a building are presented in thisthesis. As a basis for the technical and economic evaluations,exemplifying energy systems are constructed and simulated. Fuelcell system installations are predicted to be economicallyunviable for probable near-term conditions in Sweden. The mainfactor in the economic evaluations is the fuel price. However,fuel cell system installations are shown to have a higher fuelutilisation than the conventional method of energy supply. The methods presented in this thesis serve as a collectedbasis for continued research and development in the area. Keywords:Small-scale, stationary, fuel cell system,polymer electrolyte fuel cell, PEFC system, reformer,thermodynamic modelling, pinch technology, exergy analyses,system configuration, test facility, experiments, application,simulation, installation, energy system, energy storage, heatand power demand.

It is relatively straightforward to establish equilibrium inmagnetically confined plasmas, but the plasma is frequentlysucceptible to a variety of instabilities that are driven bythe free energy in the magnetic field or in the pressuregradient. These unstable modes exhibit effects that affect theparticle, momentum and heat confinement properties of theconfiguration. Studies of the dynamics of several of the mostimportant modes are the subject of this thesis. The studies arecarried out on plasmas in the reversed field pinch (RFP)configuration. One phenomenon commonly observed in RFPs is mode walllocking. The localized nature of these phase- and wall lockedstructures results in localized power loads on the wall whichare detrimental for confinement. A detailed study of the walllocked mode phenomenon is performed based on magneticmeasurements from three RFP devices. The two possiblemechanisms for wall locking are investigated. Locking as aresult of tearing modes interacting with a static field errorand locking due to the presence of a non-ideal boundary. Thecharacteristics of the wall locked mode are qualitativelysimilar in a device with a conducting shell system (TPE-RX)compared to a device with a resistive shell (Extrap T2). Atheoretical model is used for evaluating the threshold valuesfor wall locking due to eddy currents in the vacuum vessel inthese devices. A good correlation with experiment is observedfor the conducting shell device. The possibility of succesfully sustaining discharges in aresistive shell RFP is introduced in the recently rebuiltdevice Extrap T2R. Fast spontaneous mode rotation is observed,resulting in low magnetic fluctuations, low loop voltage andimproved confinement. Wall locking is rarely observed. The lowtearingmode amplitudes allow for the theoretically predictedinternal nonresonant on-axis resistive wall modes to beobserved. These modes have not previously been distinguisheddue to the formation of wall locked modes. The internal andexternal nonresonant resistive wall modes grow on the timescale of the shell penetration time. These growth rates dependon the RFP equilibrium. The internal nonresonant resistive wallmodes dominate in Extrap T2R, especially for shallow reverseddischarges. The external nonresonant modes grow solely in deepreversal discharges. <b>Keywords</b>Nuclear fusion, reversed field pinch, resistiveinstabilities, wall locked modes, tearing modes, resistiveshell modes, field errors, EXTRAP-T2, EXTRAP-T2R, TPE-RX

Over the past decade, fusion research has showed the potential of being a main candidate for energy production for future generations. Further advances in improved fusion performance are therefore vital. This thesis focuses on advanced fusion plasma scenarios and their diagnostic requirements. In particular the design of a motional Stark effect (MSE) diagnostic at the MAST spherical tokamak and the analysis of magneto-hydrodynamic mode feedback control and pulsed poloidal current drive (PPCD) at the reversed field pinch (RFP) experiment EXTRAP T2R are discussed. The MSE diagnostic is important for the determination of the plasma current profile, information that is necessary for studies in advanced confinement scenarios like reversed shear profiles or current holes. The MAST MSE system has two channels and selects the spectral components using 1Å FWHM interference filters. The diagnostic has been commissioned during the fall of 2006 and the results show the feasibility of the technique with rms-noise ~0.5° using a time resolution of 1 ms. Investigations of mirror labyrinths for the future ITER MSE diagnostic highlight the need for careful calibration considerations. Feedback control and PPCD are techniques for improved confinement. Feedback control dramatically decreases impurity influx at the end of discharges while transport in the bulk plasma is largely unaffected. During PPCD the transport is seen to decrease and it is demonstrated that PPCD and feedback control can be employed simultaneously. New and innovative techniques for fusion spectroscopy are furthermore described. This includes the use of correlations in line integrated signals to determine ion emission profiles in poloidally symmetric environments. Good agreement with other diagnostic methods is obtained. The assessment of electron temperature profiles using measured differences between Thomson scattering and vacuum ultra-violet spectroscopy is also shown. / QC 20100907

The objective of this thesis is to study two Basins of the Gulf of Guinea (GoG), namely the Dahomey and the Liberian Basins. These Basins are located in the northern part of the GoG, where oil and gas exploration has significantly increased in the last 10 years or so. We proposed geological descriptions of these two Basins. The key characteristics of the two models are the presence of channels and pinch-outs for depths of between 1 km and 2 km (these values are rescaled for our numerical purposes to 600- m and 700-m depths) and normal faults below 3 km (for our numerical purposes we use 1 km instead of 3 km). We showed that these models are consistent with the plate tectonics of the region, and the types of rocks and ages of rocks in these areas. Furthermore, we numerically generated seismic data for these two models and depth-migrated them. We then interpreted the migrated images under the assumption that the geologies are unknown. The conclusions of our interpretations are that we can see clearly the fault systems in both models. However, our results suggest that seismic interpretations of the channels and pinch-outs associated with the geology of the Dahomey and Liberian Basins will generally be difficult to identify. In these particular cases, we missed a number of channels and pinch-outs in our interpretations. The limited resolution of seismic images is the key reason for this misinterpretation.

<p>It is relatively straightforward to establish equilibrium inmagnetically confined plasmas, but the plasma is frequentlysucceptible to a variety of instabilities that are driven bythe free energy in the magnetic field or in the pressuregradient. These unstable modes exhibit effects that affect theparticle, momentum and heat confinement properties of theconfiguration. Studies of the dynamics of several of the mostimportant modes are the subject of this thesis. The studies arecarried out on plasmas in the reversed field pinch (RFP)configuration.</p><p>One phenomenon commonly observed in RFPs is mode walllocking. The localized nature of these phase- and wall lockedstructures results in localized power loads on the wall whichare detrimental for confinement. A detailed study of the walllocked mode phenomenon is performed based on magneticmeasurements from three RFP devices. The two possiblemechanisms for wall locking are investigated. Locking as aresult of tearing modes interacting with a static field errorand locking due to the presence of a non-ideal boundary. Thecharacteristics of the wall locked mode are qualitativelysimilar in a device with a conducting shell system (TPE-RX)compared to a device with a resistive shell (Extrap T2). Atheoretical model is used for evaluating the threshold valuesfor wall locking due to eddy currents in the vacuum vessel inthese devices. A good correlation with experiment is observedfor the conducting shell device.</p><p>The possibility of succesfully sustaining discharges in aresistive shell RFP is introduced in the recently rebuiltdevice Extrap T2R. Fast spontaneous mode rotation is observed,resulting in low magnetic fluctuations, low loop voltage andimproved confinement. Wall locking is rarely observed. The lowtearingmode amplitudes allow for the theoretically predictedinternal nonresonant on-axis resistive wall modes to beobserved. These modes have not previously been distinguisheddue to the formation of wall locked modes. The internal andexternal nonresonant resistive wall modes grow on the timescale of the shell penetration time. These growth rates dependon the RFP equilibrium. The internal nonresonant resistive wallmodes dominate in Extrap T2R, especially for shallow reverseddischarges. The external nonresonant modes grow solely in deepreversal discharges.</p><p><b>Keywords</b>Nuclear fusion, reversed field pinch, resistiveinstabilities, wall locked modes, tearing modes, resistiveshell modes, field errors, EXTRAP-T2, EXTRAP-T2R, TPE-RX</p>

<p>The objectives of this doctoral thesis are to give a basisincluding methods for the development of stationary polymerelectrolyte fuel cell (PEFC) systems for combined heat andpower production. Moreover, the objectives include identifyingprerequisites, requirements and possibilities for PEFC systemsproducing heat and power for buildings in Sweden. The PEFCsystem is still in a pre-commercial state, but low emissionlevels, fast dynamics and high efficiencies are promisingcharacteristics.</p><p>A thermodynamic model to simulate stationary PEFC systemshas been constructed and pinch technology and exergy analysesare utilised to design and evaluate the system. The finalsystem configuration implies a high total efficiency ofapproximately 98 % (LHV).</p><p>A flexible test facility was built in connection with theresearch project to experimentally evaluate small-scalestationary PEFC systems at KTH. The research PEFC system hasextensive measurement equipment, a rigorous control system andallows fuel cell systems from approximately 0.2 to 4 kWel insize to be tested. The simulation models of the fuel processorand the fuel cell stack are verified with experimental datataken from the test facility. The initial evaluation andsimulation of the first residential installation of a PEFCsystem in Sweden is also reported. This PEFC system, fuelled bybiogas and hydrogen, is installed in an energy system alsoincluding a photovoltaic array, an electrolyser and hydrogenstorage.</p><p>Technical aspects of designing a fuel cell system-basedenergy system, including storages and grid connections, whichprovides heat and power to a building are presented in thisthesis. As a basis for the technical and economic evaluations,exemplifying energy systems are constructed and simulated. Fuelcell system installations are predicted to be economicallyunviable for probable near-term conditions in Sweden. The mainfactor in the economic evaluations is the fuel price. However,fuel cell system installations are shown to have a higher fuelutilisation than the conventional method of energy supply.</p><p>The methods presented in this thesis serve as a collectedbasis for continued research and development in the area.</p><p><b>Keywords:</b>Small-scale, stationary, fuel cell system,polymer electrolyte fuel cell, PEFC system, reformer,thermodynamic modelling, pinch technology, exergy analyses,system configuration, test facility, experiments, application,simulation, installation, energy system, energy storage, heatand power demand.</p>

La description magnétohydrodynamique est utilisée pour étudier les plasmas de fusion par confinement magnétique dans deux configurations: tokamak et reversed field pinch. Une méthode de Fourier pseudo-spectrale et une technique de pénalisation en volume sont employées pour résoudre les équations. La méthode de pénalisation permet d'introduire des conditions aux limites de Dirichlet et donc de faire varier facilement la géométrie considérée. Les simulations dans des géométries toroïdales de type tokamak montrent l'apparition spontanée de vitesses. Une importante composante toroïdale se développe si le système est peu dissipatif. Il est aussi montré que la brisure de symétrie dans la forme de la section du tore fait apparaitre un moment angulaire toroïdal. Pour le Reversed Field Pinch on montre l'émergence de structures hélicoïdales. La forme de ces structures varie en fonction des coefficients de transport ainsi que du paramètre de pincement du champ magnétique imposé. Pour compléter l'étude on compare les résultats du tore aux calculs dans un cylindre périodique. Les différences dans la dynamique des deux cas sont mises en avant. Finalement les simulations sont confrontées à des expériences et un meilleur accord est observé entre simulation et expérience pour la géométrie toroïdale que pour la géométrie cylindrique.