Modèle en cascade

Résumé

Le modèle en cascade, ou « waterfall » en anglais, est une organisation des activités d'un projet sous forme de phases linéaires et séquentielles, où chaque phase correspond à une spécialisation des tâches et dépend des résultats de la phase précédente. Il comprend les phases d'exigences, de conception, de mise en œuvre et de mise en service. Le modèle en cascade est un cycle de vie de projet issu des industries manufacturières et du secteur de la construction, où une conception préalable est nécessaire, compte tenu des fortes contraintes matérielles et des coûts élevés afférents aux changements de la conception en cours de réalisation. Il est utilisé notamment dans les domaines de l'ingénierie et du développement de logiciels. Historique La première présentation décrivant un modèle de phases pour le développement de logiciels, est celle de Herbert D. Benington au « Symposium sur les méthodes de programmation avancées pour les calculateurs numériques » le . L'article ava

Source officielle

https://fr.wikipedia.org/wiki/Mod%C3%A8le_en_cascade

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Publications associées (15)

Publications associées

Chargement

Personnes associées

Chargement

Unités associées

Chargement

Concepts associés (28)

Concepts associés

Chargement

Cours associés (4)

Cours associés

Chargement

Séances de cours associées (7)

Séances de cours associées

Chargement

Currently, most Spent Nuclear Fuel (SNF) is kept safely in storage either at on-site facilities or at centralized interim storage sites. Moreover, many countries face delays in implementing their waste management programmes for SNF and high-level waste disposal. As storage pools approach their capacity, an increasing demand for interim dry-storage solutions is foreseen in the near future and, is associated with emerging regulatory issues. Storage prolongation in interim facilities for time periods far beyond than what originally envisaged, requires license renewal of the facility, as well as of transport and storage casks. At the same time, the safety of SNF operation activities (transportation and handling) at surface encapsulation facilities must be ensured. In order to evaluate the SNF response under normal or accident scenarios, in different stages of the back-end of the nuclear fuel cycle, it is essential to assess the mechanical properties of fuel rods, as well as their evolution during and after irradiation. Several previous studies have identified mechanisms that could affect such properties during dry-storage. However, data generated from experiments with irradiated SNF are extremely rare. This PhD complements previous studies by generating new data for the structural performance of SNF, in areas where technical gaps have been identified. In particular, the mechanical properties of SNF as function of burnup have been investigated both experimentally and numerically, under quasi-static and dynamic bending loads. The experimental activities were conducted at the hot-cell facilities of JRC-Karlsruhe and include three-point bending and gravitational impact tests performed at room temperature on pressurized samples from commercially used PWR UO2 rods. Their dynamic response was studied by successfully applying a new Image Analysis methodology developed in this work. Preliminary tests were conducted on surrogate rodlets consisting of fresh and hydrogenated Zircaloy-4 cladding tubes filled with alumina pellets. The results showed that the sampleâs ductility decreases with increasing hydrogen concentration in the cladding. Five SNF rods, with a wide range of burnup, were selected for the tests on irradiated fuel. High burnup samples showed higher toughness due to irradiation damage on the cladding, and less ductile behavior. Overall, strain rate had marginal influence on ductility for high-burnup samples, whereas for low burnup ones, ductility decreased significantly with strain rate. The flexural mechanical properties of the claddings as function of burnup were derived from the 3-point bending tests using the Euler-Bernoulli beam theory for beam elements with hollow-circular profile. A series of post-test examinations concerning the rod failure processes showed that fuel release in case of rod fracture is very limited, amounting to less than the mass of one pellet. Finite Element Analysis was performed in order to simulate the experiments and the rodsâ bending behavior under the examined load conditions. An extensive sensitivity analysis was performed to minimize modeling uncertainties and the final model was calibrated against experimental data from the 3-point bending tests. Good agreement was observed between numerically predicted and experimentally derived mechanical properties, and the model can be used to simulate the mechanical behavior of SNF rods under different loading configurations.

EPFL2021

Chargement

Learning from demonstration using products of experts: Applications to manipulation and task prioritization

Sylvain Calinon, Emmanuel Pignat

Probability distributions are key components of many learning from demonstration (LfD) approaches, with the spaces chosen to represent tasks playing a central role. Although the robot configuration is defined by its joint angles, end-effector poses are often best explained within several task spaces. In many approaches, distributions within relevant task spaces are learned independently and only combined at the control level. This simplification implies several problems that are addressed in this work. We show that the fusion of models in different task spaces can be expressed as products of experts (PoE), where the probabilities of the models are multiplied and renormalized so that it becomes a proper distribution of joint angles. Multiple experiments are presented to show that learning the different models jointly in the PoE framework significantly improves the quality of the final model. The proposed approach particularly stands out when the robot has to learn hierarchical objectives that arise when a task requires the prioritization of several sub-tasks (e.g. in a humanoid robot, keeping balance has a higher priority than reaching for an object). Since training the model jointly usually relies on contrastive divergence, which requires costly approximations that can affect performance, we propose an alternative strategy using variational inference and mixture model approximations. In particular, we show that the proposed approach can be extended to PoE with a nullspace structure (PoENS), where the model is able to recover secondary tasks that are masked by the resolution of tasks of higher-importance.

SAGE PUBLICATIONS LTD2021

Chargement

An MPEG RVC AVC Baseline Encoder Based on a Novel Iterative Methodology

Hussein Aman-Allah, Ihab Mostafa Amin Mohamed Amer, Ehab Hanna, Karim Maarouf

With the emergence of new generations of multicore architectures, the need for efficient multimedia algorithm implementations has become critical. This paper describes a new methodology for efficient implementations of algorithms targeting reconfigurable architectures. The Reconfigurable Video Coding (RVC) standard aims to provide a framework allowing a dynamic development,implementation and adoption of standardized video coding solutions with features of higher flexibility and reusability. RVC-CAL actor language is a dataflow language that makes better use of the multicore and parallel architectures. The proposed design flow methodology follows an iteration-based implementation model rather than the traditional sequential model. Analysis, design, development, simulation, testing and adaptation are performed with every iteration ending up with a functional “version” of the algorithm. A case study is conducted to illustrate the productivity of the proposed methodology in which the implementation of an AVC baseline encoder on a Xilinx Virtex 5 XC5VLX50T FPGA demonstrated for intra prediction architecture search space co-exploration.

2009

Chargement

EPFL2021