21549 research outputs found
Sort by
Investigating waste plastic pyrolysis kinetic parameters by genetic algorithm coupled with thermogravimetric analysis
Get PDFPyrolysis of waste plastic (WP) is a promising method to solve the plastic pollution issue. WP is mainly composed of polyethylene (PE). Moreover, the products of waste polyethylene (WPE) pyrolysis could serve as high quality fuels and the feedstocks of petrochemicals. Therefore, it is essential to investigate the WPE and WP pyrolysis process. This study evaluates pure PE, WPE and WP pyrolysis kinetic parameters by the use of genetic algorithm (GA) and isoconversional methods coupled with thermogravimetric analysis (TGA), respectively. Additionally, three representative reaction models, i.e. reaction-order, extended Prout–Tompkins and Sestak–Berggren models, are investigated for obtaining the most suitable model, which could describe the PE, WPE and WP pyrolysis process more accurately. Consequently, the reaction-order model turns out to be the optimal method for appropriately describing PE, WPE and WP pyrolysis processes. Hence, the pyrolysis parameters optimized by GA are proven to be accurate and reliable, in comparison of calculated values of activation energy by isoconversional methods and experimental data. Moreover, it might be applicable of GA coupled with TGA with reaction-order model to the future industrial WPE and WP pyrolysis circumstances that have variable heating rate
How to introduce expert feedback in one-class support vector machines for anomaly detection?
Get PDFAnomaly detection consists of detecting elements of a database that are different from the majority of normal data. The majority of anomaly detection algorithms considers unlabeled datasets. However, in some applications, labels associated with a subset of the database (coming for instance from expert feed- back) are available providing useful information to design the anomaly detector. This paper studies a semi-supervised anomaly detector based on support vector machines, which takes the best of existing supervised and unsupervised support vector machines algorithms. The proposed algorithm allows the maximum proportion of vectors detected as anomalies and the maximum proportion of errors in the supervised data to be controlled, through two hyperparameters defining these proportions. Simulations conducted on various benchmark datasets show the interest of the proposed semi-supervised anomaly detection method
Modelling of three-dimensional particle rebound from an anisotropic rough wall
Get PDFIn gas-solid confined flows, particle rebound on a rough wall may have a great influence on the macroscopic properties of both phases. Existing simulation methods use simplified approaches where surface roughness is modelled as two-dimensional or isotropic. To improve the accuracy of such simulations, an anisotropic virtual rough wall statistical approach and the corresponding Lagrangian stochastic particle-wall collision model are proposed. From deterministic simulations of the impact of particles on the anisotropic virtual wall given as correlated bi-Gaussian surface, wall-normal vector angle statistics are used for model development. Rebound angle statistics computed from deterministic simulations and using the proposed stochastic model are in very good agreement. Incident particles in vertical planes not collinearwith one of the principal directions of the anisotropic rough surface are bouncing with a mean transverse angle towards the principal direction of lower roughness
Loss assessment of a counter rotating open rotor using URANS/LES with phase-lagged assumption
Get PDFThis paper presents the study of the losses generated in a Counter Rotating Open Rotor (CROR) configura- tion at three different operating conditions (approach, cutback and sideline). Unsteady Reynolds Averaged Navier-Stokes (URANS) and Large-Eddy Simulation (LES) approaches are used and compared to describe the flow field and the mechanisms of loss. Since no common circumferential periodicity occurs in the two blade rows of the configuration (11 blades for the front rotor and 9 for the rear rotor), a full 360 ◦simu- lation would be required. In order to reduce the related computational cost, a phase-lagged assumption approach is used. This method enables to perform unsteady simulations on multi-stage propulsive con- figurations including multiple frequency flows with a computational domain reduced to one single blade passage for each row. The phase-lagged approach requires a large data storage reduced in the study by a data compression method. The data compression method is based on a Proper Orthogonal Decomposition (POD) replacing the traditional Fourier Series Decomposition (FSD). The inherent limitation of the phase- shifted periodicity assumption remains with the POD data storage but this compression method alleviates some issues associated with the FSD, especially spectrum content issues. The analysis of the losses gen- erated in the configuration is based on an entropy formulation. In particular, the losses are split between boundary layer contributions and the remaining domain where wakes and secondary flows occur. The study shows the influence of the leading edge vortex on the suction side boundary layer transition of the front and rear rotor blades at high rotational speed (cutback and sideline). The main source of losses is associated with the suction side boundary layer over the front and rear rotor blades with a main peak of loss production at around 75% of the blade chord
Mental Workload Estimation Based on Physiological Features for Pilot-UAV Teaming Applications
Get PDFManned-Unmanned Teaming (MUM-T) can be defined as the teaming of aerial robots (artificial agents) along with a human pilot (natural agent), in which the human agent is not an authoritative controller but rather a cooperative team player. To our knowledge, no study has yet evaluated the impact of MUM-T scenarios on operators' mental workload (MW) using a neuroergonomic approach (i.e., using physiological measures), nor provided a MW estimation through classification applied on those measures. Moreover, the impact of the non-stationarity of the physiological signal is seldom taken into account in classification pipelines, particularly regarding the validation design. Therefore this study was designed with two goals: (i) to characterize and estimate MW in a MUM-T setting based on physiological signals; (ii) to assess the impact of the validation procedure on classification accuracy. In this context, a search and rescue (S&R) scenario was developed in which 14 participants played the role of a pilot cooperating with three UAVs (Unmanned Aerial Vehicles). Missions were designed to induce high and low MW levels, which were evaluated using self-reported, behavioral and physiological measures (i.e., cerebral, cardiac, and oculomotor features). Supervised classification pipelines based on various combinations of these physiological features were benchmarked, and two validation procedures were compared (i.e., a traditional one that does not take time into account vs. an ecological one that does). The main results are: (i) a significant impact of MW on all measures, (ii) a higher intra-subject classification accuracy (75%) reached using ECG features alone or in combination with EEG and ET ones with the Adaboost, Linear Discriminant Analysis or the Support Vector Machine classifiers. However this was only true with the traditional validation. There was a significant drop in classification accuracy using the ecological one. Interestingly, inter-subject classification with ecological validation (59.8%) surpassed both intra-subject with ecological and inter-subject with traditional validation. These results highlight the need for further developments to perform MW monitoring in such operational contexts
Etude des propriétés intrinsèques et tribologiques des composites à matrice PEEK chargés de renforts micro/nanométriques.
Get PDFCe projet veut répondre à un enjeu de l'industrie : prévenir l'usure des pièces et proposer des matériaux composites plus performants. Ainsi, cette thèse porte sur l’étude des mécanismes qui se mettent en place lors du frottement des composites à matrice PEEK. L’ambition est d’établir un lien entre la morphologie cristalline du polymère, la nature des charges, leur taille et leur concentration et, les mécanismes phénoménologiques lors du frottement. Pour cela, des composites à matrice PEEK à charges micrométriques et nanométriques sont élaborés par voie fondue. Les charges sont des composés lamellaires: le nitrure de bore hexagonal et le graphite ; et des particules céramiques : l'alumine et le carbure de silicium, de concentration comprise entre 2,5 et 10 % en volume. Les matériaux sont caractérisés par des mesures de densité, DRX, DSC, spectroscopie Raman et rhéométrie pour connaître la microstructure du polymère et des charges. Des essais tribologiques de translation réciproque en configuration bille/plan sont réalisés à température ambiante avec une force normale de 25 N et une fréquence de 5 Hz de 5 à 25 minutes pour suivre l’évolution des mécanismes au cours du temps. Les charges particulaires renforcent mécaniquement le composite, la déformation est plus faible. Cependant, les PEEK faiblement chargés en graphite apparaissent comme les plus prometteurs pour retarder l’usure du composite. Dans ces cas-là, le troisième corps protège la contreface métallique et le composite. Grâce à leur conductivité thermique plus haute, l’énergie créée lors du frottement est dissipée plus rapidement. De plus, une diminution locale de la cristallinité dans les traces d’usure a été identifiée par spectroscopie Raman. Les analyses morphologiques et structurales menées à la suite des essais tribologiques établissent un lien entre la microstructure et la morphologie de la trace d’usur
Hybrid direct and interactive solvers for sparse indefinite and overdetermined systems on future exascale architectures
Get PDFIn scientific computing, the numerical simulation of systems is crucial to get a deep understanding of the physics underlying real world applications. The models used in simulation are often based on partial differential equations (PDE) which, after fine discretisation, give rise to huge sparse systems of equations to solve. Historically, 2 classes of methods were designed for the solution of such systems: direct methods, robust but expensive in both computations and memory; and iterative methods, cheap but with a very problem-dependent convergence properties. In the context of high performance computing, hybrid direct-iterative methods were then introduced inorder to combine the advantages of both methods, while using efficiently the increasingly largeand fast supercomputing facilities. In this thesis, we focus on the latter type of methods with two complementary research axis.In the first chapter, we detail the mechanisms behind the efficient implementation of multigrid methods. The latter makes use of several levels of increasingly refined grids to solve linear systems with a combination of fine grid smoothing and coarse grid corrections. The efficient parallel implementation of such a scheme is a difficult task. We focus on the solution of the problem on the coarse grid whose scalability is often observed as limiting at very large scales. We propose an agglomeration technique to gather the data of the coarse grid problem on a subset ofthe computing resources in order to minimise the execution time of a direct solver. Combined with a relaxation of the solution accuracy, we demonstrate an increased overall scalability of the multigrid scheme when using our approach compared to classical iterative methods, when the problem is numerically difficult. At extreme scale, this study is carried in the HHG framework(Hierarchical Hybrid Grids) for the solution of a Stokes problem with jumping coefficients, inspired from Earth's mantle convection simulation. The direct solver used on the coarse grid is MUMPS,combined with block low-rank approximation and single precision arithmetic.In the following chapters, we study some hybrid methods derived from the classical row-projection method block Cimmino, and interpreted as domain decomposition methods. These methods are based on the partitioning of the matrix into blocks of rows. Due to its known slow convergence, the original iterative scheme is accelerated with a stabilised block version of the conjugate gradient algorithm. While an optimal choice of block size improves the efficiency of this approach, the convergence stays problem dependent. An alternative solution is then introduced which enforces a convergence in one iteration by embedding the linear system into a carefully augmented space.These two approaches are extended in order to compute the minimum norm solution of in definite systems and the solution of least-squares problems. The latter problems require a partitioning in blocks of columns. We show how to improve the numerical properties of the iterative and pseudo-direct methods with scaling, partitioning and better augmentation methods. Both methods are implemented in the parallel solver ABCD-Solver (Augmented Block Cimmino Distributed solver)whose parallelisation we improve through a combination of load balancing and communication minimising techniques.Finally, for the solution of discretised PDE problems, we propose a new approach which augments the linear system using a coarse representation of the space. The size of the augmentation is controlled by the choice of a more or less refined mesh. We obtain an iterative method with fast linear convergence demonstrated on Helmholtz and Convection-Diffusion problems. The central point of the approach is the iterative construction and solution of a Schur complemen
Numerical simulations of the agitation generated by coarse-grained bubbles moving at large Reynolds number
Get PDFWe present a numerical method for simulating the flow induced by bubbles rising at large Reynolds number. This method is useful to simulate configurations of large dimensions involving a great number of bubbles. The action that each bubble exerts on the liquid is modelled as a volume source of momentum distributed over a few mesh-grid elements. The flow in the vicinity of the bubbles is thus not finely resolved. The bubbles are treated as Lagrangian particles that move under the influence of the hydrodynamic force exerted by the liquid. The determination of this force on a given bubble requires knowledge of the liquid flow that is undisturbed by this bubble. A model is developed to accurately estimate this disturbance for large-Reynolds-number objects and get rid of any spurious self-induced effect. Thanks to that, a homogeneous swarm of rising bubbles is simulated. Comparisons with experiments show a good agreement with the flow scales larger than the bubbles, which turn out to be controlled by the interactions between bubble wakes and rather independent of unresolved smaller scales. This method can be used to study the coupling between bubble-induced agitation and large-scale motions, such as those produced in industrial bubble columns
Electro-mechanical Resonant Ice Protection Systems: Energetic and Power Considerations
Get PDFThis article focuses on resonant ice protection systems and proposes key performance indicators to analyze the performances of such systems with respect to levels of energy, force, and power required for de-icing. The principle of these systems is to apply vibrations or ultrasonic waves onto the structure that create high-level stresses greater than those required to crack and delaminate to remove the ice accumulated on the structure. The computation of the indicators requires two values: the ice adhesion strength and the critical strain energy release rate. Computations are performed assuming three stages of a de-icing mechanism: firstly an initiation of cohesive fractures by tensile stress at the top surface of the ice layer, secondly, a propagation of cohesive fractures within the ice and thirdly a propagation of adhesive fractures at the ice/substrate interface starting from the base of the cohesive fractures previously created. The proposed key performance indicators provide guidance on the use of flexural and extensional modes in resonant ice protection systems and on the frequency range to favor when looking at fractures initiation and propagation. Calculations based on the key performance indicators show a potential power reduction by ten with resonant electro-mechanical de-icing systems compared to electro-thermal systems
A Nonmonotone Matrix-Free Algorithm for Nonlinear Equality-Constrained Least-Squares Problems
Get PDFLeast squares form one of the most prominent classes of optimization problems, with numerous applications in scientific computing and data fitting. When such formulations aim at modeling complex systems, the optimization process must account for nonlinear dynamics by incorporating constraints. In addition, these systems often incorporate a large number of variables, which increases the difficulty of the problem, and motivates the need for efficient algorithms amenable to large-scale implementations.
In this paper, we propose and analyze a Levenberg-Marquardt algorithm for nonlinear least squares subject to nonlinear equality constraints. Our algorithm is based on inexact solves of linear least-squares problems, that only require Jacobian-vector products. Global convergence is guaranteed by the combination of a composite step approach and a nonmonotone step acceptance rule. We illustrate the performance of our method on several test cases from data assimilation and inverse problems: our algorithm is able to reach the vicinity of a solution from an arbitrary starting point, and can outperform the most natural alternatives for these classes of problems