The recent advent of unstructured high order methods holds the promise of transforming industrial computational fluid dynamics practices. These novel methods combine high precision to high geometrical flexibility, but also provide excellent serial and parallel efficiency up to very large core counts. They also provide many possibilities for adaptive resolution, including accurate non-conformal connections and order adaptation. Due to the combination of these features, these methods are technically better suited for industrial scale resolving computations than the second order finite volume codes that pervade industrial CFD. As this type of computations will be required in the near future, and industrial simulation practices have not yet been firmly established, there is a unique window of opportunity to introduce these novel high-potential methods. Of these methods, the discontinuous Galerkin method is arguably the most mature, in terms of the theoretical framework, its practical deployment and applications. The main objective of this work is the development of this method for tackling scale-resolving simulations of turbomachinery flows. This thesis treats important practical and theoretical steps towards this goal. The first contribution is the study of the stability of the interior penalty method on hybrid meshes. Through the sharp computation of a trace inverse inequality, already available for simplex meshes, for all types of elements, and the generalisation of the coercivity analysis, optimal penalty parameters are found. In particular a specific variant for high aspect ratio meshes, typical for CFD, has been found. A further development concerns the definition of optimal data structures and assembly routines, based on the evaluation of computational performance of algebraic primitives. It has been shown how to define an extremely flexible matrix structure, whilst obtaining very high efficiency. By the reorganisation of assembly routines the increase of computational complexity with interpolation order is largely compensated higher computational efficiency. Both matrix-free GMRES and p-multigrid iterative strategies are furthermore implemented and compared. For the p-multigrid strategy, a generic framework is presented that provides high-quality transfer operators for h- and p-multigrid. Finally, a number of conceptual extensions to the DGM are proposed, namely the practical implementation of a non-conformal method for incompressible flows, as well as the application to frequential formulations, in particular of the Linearised Euler Equations including Perfectly Matched Layers. A first industrial CFD application is shown, which concerns the direct numerical simulation of transitional flow in a low pressure turbine cascade.

(FSA 3)  – UCL, 2013

The new contra-rotating turbofan architecture leads to a substantial reduction of the pollutants and the noise emission of civil aircrafts. This architecture is based on two shorter fans, rotating in opposite direction at lower speed. Due to the mechanical link between the fan and the low-pressure compressor, its speed is also reduced. In order to keep the pressure ratio constant, without weight increase, the low-pressure compressor stages are designed with an unusually high loading coefficient and flow coefficient, which leads to a reduced stability margin. The purpose of this thesis is to present a complete characterization of the first stage of the low-pressure compressor for the new contrarotating turbofan architecture in terms of global performance, unsteady flow field and instability phenomena. This work will be the basis for future research in this field, considering the low number of publications on the low-pressure compressors and the absence of references concerning the low-pressure compressor of contra-rotating turbofan. The characterization is performed on a half scale model on the VKI R4 test facility. This facility is a closed loop, high speed compressor rig, allowing the investigation of the influence of the Reynolds number on the performance. The effect of a circumferential groove casing treatment on the performance, the stability and the unsteady flow field is studied. The measurements are compared to numerical simulations from various CFD codes using various turbulence models.

(FSA 3)  – UCL, 2012

The dam-break flow is an extreme example of a severe transient flow, often associate with catastrophic flood events. In a real case it is a very complex flow, involving not only many different time and space scales but also a complex fluid-bed interaction that will affect the propagating wave celerity. The need for studies on this kind of flow arises mainly from the risk assessment needed for existing, or to-be build, dams. The definition of the flooded areas and their evolution with time constitutes a valuable information for the authorities' evacuation plans conception. Numerical models are often used to obtain the required information about the dam-break flow propagation. However for an improved modelling it is important to know the physics of the flow. This can be acquired by means of small scale models in the laboratory, where the effect of the different variables can be isolated and analysed. The information thus obtained can be used to gain a better comprehension of the dam-break flow, but also to obtain an experimental results data base that can be used to calibrate and validate existing numerical and analytical models. The present study is devoted to the experimental analysis of the dam-break flow, with special emphasis on the initial stages of the dam-break flow. Rather than focus only in the macro-scale variables, i.e., water level evolution and wave celerity, the present study deals with the point-wise velocity measurements within the dam-break flow. For velocity measurements different techniques are used. Two of them based on imaging techniques (PIV and PTV), and a third one based on the acoustic Doppler Effect (UVP). These techniques allowed for a space and time description of the dam-break flow velocity field. The measurements were carried considering different flow heights and bed types. The dam-break flow over beds of sediments is here analysed and the velocity field of both water layer and sediment layer measured simultaneously. A two-layer numerical model is than used and tested against the obtained experimental results. The transient character of the dam-break flow motivated a statistical analysis of the dam-break flow. It consisted in a) repeating several times, for the same initial conditions, the dam-break experiments; b) obtaining images of the flow with an high acquisition rate (1000 Hz). An ensemble average procedure is presented and the results discussed. For fixed smooth beds, the obtained spatial and time resolved velocity profiles made it possible, under some conditions, to determine the shear stress in the bed. %This was made by identifying in some conditions the appearance of a log-layer. Finally the different measurement techniques are compared and discussed.

(FSA 3)  – UCL, 2013

In this Doctoral Thesis we are concerned with the theoretical and numerical analysis of hydrostatics and low-Mach number flows of fluid saturated granular materials. In the first part, we generalize the concept of low-Mach number approximation to multi-phase flows and apply it to derive the low-Mach number equations for the mixtures of interest. In the second part, we develop a projection-type, successive-over-relaxation method for the computation of hydrostatics of such mixtures. We also prove that this method is formally convergent. The second part concludes with the presentation and discussion of representative numerical results. In the third part, we develop an existence theory for hydrostatics of two-phase granular mixtures. By employing weak-convergence methods and by choosing suitable non-linear test functions, we prove the existence of at least one weak solution, in the sense of distributions. This gives a positive answer to the open question of the well-posedness of continuum models for granular materials at equilibrium. In the fourth part, we present an algorithm for the calculation of constant density flows of the mixtures of interest. This algorithm combines a predictor-corrector time-integration scheme with a projection method for the computation of the pressure of each phase. It also incorporates a detection and treatment methodology for the material interfaces that develop in such flows. The efficiency and robustness of the proposed algorithm are tested via a series of numerical tests. Finally, this algorithm is employed to perform the first ever simulations of unsteady shear flows of granular mixtures, the characteristics of which are discussed in detail.

(FSA 3)  – UCL, 2012

The present thesis focuses on the development of numerical models capable of predicting reliably, with constant material parameters, the failure of metal-to-metal adhesive joints in a large range of configurations. Two finite element-based models are considered. The first one, referred to as the AACZ model (Adherend + Adhesive + Cohesive Zone), consists of embedding in the adhesive a single row of a cohesive zone elements with zero thickness which are meant to represent and condense the damage mechanisms responsible for fracture. The second one, referred to as the MPS model (Maximum Principal Stress), considers that the conditions for steady-state crack propagation are met when the maximum principal stress reaches a particular value at a critical distance ahead of the crack tip, this particular value being needed to debond, cleave or cavitate the second-phase particles present in the adhesive. The numerical results first show that the MPS model is very promising since it reproduces accurately, over a large range of adhesive layer thicknesses and test conditions, the failure of three commercial structural adhesives exhibiting different levels of resistance to fracture or adhesive fracture energies. The numerical predictions obtained with both models are further exploited to gain a better understanding of: (i) the physical mechanisms responsible for fracture in these adhesives, (ii) the different energy terms contributing to the adhesive fracture energy, and (iii) the effects of constraint. Hence, it has been possible to determine how the adhesive fracture energy is affected as a function of the geometry of the test specimen and the type of test.

(FSA 3)  – UCL, 2012

The collapse of a dam induces a fast transient flow that may interact with the erodible bed and banks of the downstream river, entraining large amounts of sediments and modifying the valley morphology. It may constitute a danger for the surrounding inhabitants and infrastructures. To reduce the catastrophic consequences of such an event, the decision-makers need tools to assess the risk posed by a dam and to prepare emergency plans. This thesis presents a numerical model aimed at simulating such transient flows over mobile beds. The two-dimensional horizontal (2DH) two-layer shallow-water model considers an upper clear-water layer and a lower bed-load transport layer flowing on a motionless sediment bed. The depth-averaged velocities in both layers are assumed distinct in norm and direction. Distinct granular concentrations are assumed in the bed-load layer and in the motionless bed. The mass and horizontal momentum balances are written in each layer. Through the source terms, vertical mass and momentum transfers may occur between the layers following erosion / deposition. The conservative part of the governing equations is shown to be hyperbolic. The set of equations is solved on unstructured triangular meshes by a finite-volume scheme with a LHLL solver. Several boundary conditions can be handled. A bank-failure operator is inserted in the model. It compares local bank slopes with threshold values and tilts the unstable elements until they reach a repose angle. The model is validated against theoretical and experimental results of dam-break flows over sand beds, e.g. in an enlargement, on a floodplain and in a prismatic flume.

(FSA 3)  – UCL, 2012

One countermeasure used during long duration spaceflight to maintain bone and muscle mass is a treadmill equipped with a subject loading system (SLS) that simulates gravity. To date, little is known about the biomechanics of running in weightlessness on such a treadmill-SLS system. We have designed an instrumented treadmill/force-plate to compare the biomechanics of running in weightlessness to running on Earth. Gravity was simulated by two pneumatic pistons pulling downwards on a subject’s harness, with a force approximately equal to body weight on Earth. Four transducers, mounted under the treadmill, measured the three components of the reaction force exerted by the tread-belt under the foot. A high-speed video camera recorded the movements of limb segments while electromyography of four lower limb muscles was registered. Experiments in weightlessness where conducted during European Space Agency parabolic flight campaigns. Control experiments were performed on the same subjects on Earth. When running on the treadmill with a SLS, the bouncing mechanism of running is preserved. Depending on the speed of progression, the ground reaction forces, contact and aerial times, muscular work and bone stress differed by a maximum of ± 5-15% during running on the treadmill with a SLS, as compared to on Earth. The movements of the lower limb segments and the EMG patterns of the lower limb muscles were also comparable. Thus the biomechanics of running on Earth can reasonably be duplicated in weightlessness using a treadmill with a SLS that generates a pull-down force close to body weight on Earth.

(FSA 3)  – UCL, 2011

Surfaces are more and more used to add specific properties to materials towards multifunctional systems. These properties are introduced using thin coatings deposited on the bulk material. Reliability is of prime importance for these systems and adhesion is one of the most important parameter to control and optimize. Understanding the influence of the deposition parameters on the adhesion is a very challenging issue. Methods to properly measure the adhesion are not widely used in the industry due to experimental and conceptual difficulties. Fracture approaches to measuring, describing or predicting thin film adhesion essentially remain an academic field of research. The objective of this thesis has been to conciliate these two different viewpoints. A simple and versatile adhesion testing method, the wedge opening test, has been developed with a robust data reduction scheme. A plate identical to the substrate is bonded on top of the film to allow the insertion of the wedge. The adhesion is determined using geometrical parameters only without the need to a force measurement. The most difficult parameter to measure, in the case of non transparent substrate, is the crack length. An original method based of the out-of-plane displacement field of the specimen has been validated and used during this thesis. A steady-state finite element code is used to simulate the test and to study the contribution of each layer to the total fracture energy. The simulation code allows the determination of the contribution of each layer to the total fracture energy. Specific attention is paid to the contribution of the adhesive layer which is extrinsic to the system. The dissipation in the adhesive layer can be as high as 2/3 of the total energy dissipated in the specimen during crack propagation, in the case of film thicknesses lower than one micron.

(FSA 3)  – UCL, 2010