1,720,965 research outputs found
Goal-oriented mesh adaptation using mesh sensitivities as an indicator
No full textAerospace EngineeringAerodynamics, Wind Energy & PropulsionAerodynamic
Aerodynamic analysis of a platoon of bluff bodies subjected to cross wind, a numerical investigation on the effect of drag reduction devices
No full textDue to the awareness of climate change, more sustainable and efficient ways of transport are needed. The most polluting road transport mode in the European Union are the long-haul tractor-trailer combinations. This study will investigate the aerodynamic characteristics of three vehicles in a platoon subjected to cross wind conditions. The effect of front- and rear drag reduction devices is investigated numerically. The frontal edge radius is considered the front drag reduction device while the rear drag reduction device is represented by a boat tail. ANSYS Fluent was used to solve the RANS equations that were closed with the SST k-omega turbulence model. The GETS model was used to analyse the aerodynamic characteristics of a simplified heavy-duty vehicle. The geometry was meshed and a mesh sensitivity study showed the asymptotic behaviour of the drag as a function of mesh size. It was shown that the frontal edge radius influences the flow behaviour largely. For the radii of 0.54 m and 0.27 m, a large thrust force on the frontal edges decreased the drag of the front part of the vehicle significantly such that the drag contribution of the front part was lower than 15% of the total drag while the pressure drag at the rear delivered around$70% of the drag. Halving the frontal edge radius to 0.135 m, caused the drag contributions of the front and rear part of the vehicle to be about 45% of the total drag. Adding an inward deflected tail increased drag drastically by increasing the base pressure. Adding a cross wind component to the incoming flow increased the drag and gave a side force. The inter-vehicle distance had a large influence on the drag of the individual vehicles. At a very short distance, the lead vehicle experienced a very large drag reduction because of the presence of the high pressure region in front of the middle vehicle. This lead to a large reduction in pressure drag. When the distance is increased, the drag goes asymptotically to the value in isolation. The trailing vehicle showed opposite behaviour. The drag of the trailing vehicle was increased when the inter-vehicle distance decreased due to the fact that the streamlines were deflected inward after the middle vehicle and they experienced a deceleration due to the concave trajectory of the flow arriving at the frontal surfaces of the trailing vehicle. This lead to a decreased thrust force on the frontal edges of the trailing vehicle. For the lowest inter-vehicle distance, the trailing vehicle was positioned inside the near-wake of the middle vehicle and the drag decreased again. The middle vehicle experienced a combination of the effects on the lead and trailing vehicle, its drag remains fairly constant except for the closest distance where the negative effect on the frontal part started to decrease. These trends are similar for all platoons, the amount of drag increase or decrease however was determined by the geometric variables and the cross wind condition. The drag reductions of the different vehicles in the platoon were mainly determined by the frontal edge radius. The relative drag reductions on the lead vehicle were higher when the frontal edge radius was larger, since then the pressure drag at the rear represents a larger portion of the total drag. The drag decrease in drag counts however only slightly changed. For the large radii, the trailing vehicle experienced a drag that increased up to a value above that in isolation when the inter-vehicle distance was decreased. A trailing vehicle with the smallest frontal edge radius, 0.135 m, experienced a drag reduction. The combined effects of the lead and trailing vehicle caused the drag reductions of the middle vehicle to increase for a lower frontal edge radius. The drag reductions decreased when tails were added or deflected inwards. A tail on a specific vehicle mainly influences the vehicle driving behind the tail. Deflecting a tail decreased the drag reductions caused by the slower incoming flow field, i.e. the lower dynamic pressure experience by the following vehicles. Deflecting a tail on the lead vehicle inward has the same effect on the middle vehicle as increasing the inter-vehicle distance, it decreases the influence of the lead vehicle's wake on the middle vehicle. For a frontal edge radius of 0.27 m, the total drag reduction of a following vehicle was decreased when the tail was deflected from 0° to 6°, but it was increased when deflecting the tail from 6° to 12°. Decreasing the frontal edge radius to 0.135 m gave different results. The total drag of the following vehicle was increased when the tails of the vehicle in front were deflected more inward. Overall, the drag reductions obtained in a platoon were not much affected by the applied cross wind conditions. It is shown that the lead vehicle of a platoon redirects the flow such that the middle and trailing vehicle experience a significantly lower side force. The side force of the lead vehicle itself is not changed significantly. When inward deflected tails are applied to the lead and middle vehicle, the side force reduction on the following vehicles is diminished but still significant. The final results showed that the platoon with the worst performing vehicles in isolation, the `nnn' configuration with R = 0.135 m, experienced the largest drag reduction, namely 29%, corresponding to 1223 drag counts. If this is extrapolated to a vehicle with an initial drag coefficient as high as a real heavy-duty vehicle, this could lead to a fuel saving of 2.48 L/100km. Considering the absolute drag coefficients, the platoon with the best performing vehicles was still the optimal platoon with the lowest drag coefficient, namely the `t12t12t12' platoon with the largest frontal edge radius. This aerodynamic analysis investigated the effect of the drag reduction devices in a platoon of bluff bodies subjected to cross wind conditions. It was shown that the frontal edge radius can determine whether the following vehicles experience a drag reduction or a drag increase. Tails reduced the drag of the individual vehicles but did also decrease the drag reductions experienced in the platoon. The drag reductions were still present under cross wind conditions. Considering the set-up of platoon formations, vehicles with the most streamlined front should be placed as lead vehicle. Vehicles with large tail angles should be used as trailing vehicle.AerodynamicsAerodynamics & Wind EnergyAerospace Engineerin
Robust Aerodynamic Optimization through Conjugate Gradient Method with Taguchi's Theory
No full textThe main objective of this thesis project is to establish an optimization framework that can carry out robust aerodynamic design tasks. The work is based on the single-point optimization module of the SU2 code, which contains the partial differential equation solver for flow evaluation and gradient calculation based on the adjoint method. The research work can be divided into two parts: the first one is the establishment of the optimization structure and the corresponding implementation. And the second part is the aerodynamic design examples. During creating the framework of the robust optimization process, the conjugate gradient (CG) algorithm is used to establish the main structure (outer loop). With the conjugate search directions provided by the CG method, line searches are implemented with the application of the strong Wolfe condition (inner loop). In order to carry out robust optimizations within the uncertain operating conditions, the format of the objective function should be defined properly. The Taguchi’s robust design theory is used to create the objective function that takes both the performance expectation and the variance into account simultaneously. As the CG algorithm cannot directly deal with constraints, they should be converted to the penalty terms in the objective function. To check the validity of the established robust optimization process, two examples are tested concerning the wave drag reduction of the NACA0012 airfoil under subsonic condition, with lift and thickness constraints. The first problem is to reduce the drag under the uncertain Mach number and angle of attack which obey certain kind of normal distributions separately. The continuous probabilities of the two uncertainties are firstly discretized into 9 operating conditions and the joint probability is calculated. After that optimizations are carried out under these sampled conditions. The results show that the process can indeed provide robust drag reduction. Compared to the results of the two single-point optimizations under different conditions, the drag value is effectively reduced especially under higher Mach numbers and larger angles of attack. The change of the weight factor distribution for the drag expectation and the variance has noticeable influence on the drag value under the most critical condition. The second problem is to reduce the drag within a certain range of the Mach number while keep a constant lift. The Mach number is the only uncertainty source and it is discretized at 3 sampled points. The results show that the shock wave can be eliminated under all 3 conditions. As a result, both the drag expectation and the variance are significantly reduced. The robust optimizations with different weight factor distribution have similar results except for the one that only focuses on the drag variance reduction. The latter optimization provides a result with nearly no drag variance at the cost of higher drag values throughout the whole tested domain. And the drag increases more quickly than other robust optimization results with the increment of Mach number. Actually, for this problem, a single-point optimization under the highest Mach number could also provide robust drag reduction performance. The test examples preliminarily proved the validity of the established robust optimization process. However, it is recommended that the effectiveness should be further tested with more complicated problems in the future.Flight Performance and PropulsionAerospace Engineerin
Discrete-Time Nonlinear Reduced-Order Models for Aeroelastic Analysis: Linear System Idenfitifcation Methods and a Polynomial Nonlinear Model Investigation
No full textIn computational aeroelasticity, unsteady aerodynamics is computationally expensive compared to the structural calculation. This problem becomes more severe for prediction of flutter or Limit-Cycle oscillation (LCO), which requires multiple runs of simulations at various flow conditions. The former can be captured by linear computational-aerodynamic-fluid (CFD) solver, while the prediction of the latter requires nonlinear CFD solver. Reduced-order modelling (ROM) techniques for unsteady aerodynamics have been investigated extensively. Among these ROMs, Auto-Regression with eXogeneous variables (ARX) has been applied to predict the flutter behaviour successfully for relatively complicated test cases. Motivated by the nice performance of ARX, whether other linear system identification methods, such as Auto-Regression Moving Average with eXogeneous inputs (ARMAX), Output-Error (OE) and Box-Jenkins (BJ), are effective as linear ROMs is investigated. When it comes to the nonlinear ROM, the linear information is expected to use. A polynomial-based state-space model, extended from linear system identification, is defined to combine the linear part with nonlinear functions of the state and input. A direct requirement for the linear method is the representation in state-space form. The nonlinear functions compose of polynomials of degree equal or greater than . The coefficients of this nonlinear model are obtained by solving an optimisation problem with a Levenberg-Marquardt (LM) algorithm. Furthermore, we assume that the linear method is capable of capturing the flutter and keep the linear matrices constant during the nonlinear optimisation to reduce computational cost. As for the test case for LCO, an analytical Van-der-pol (VDP) oscillator is selected. The nonlinear ROM is built to replace the nonlinear term in VDP. Coefficients of the nonlinear ROM are obtained by solving an optimisation problem and the validation is carried out by reproducing the VDP oscillation. Another important factor for a fairly accurate ROM is the training signal. Different training signals are examined to reproduce the VDP oscillation. After checking the theoretical representation and the numerical solution algorithm, ARX and ARMAX are selected as linear ROMs for two reasons: the construction of state-space representation is explicit; ARMAX model is a variation from ARX by adding averaged error terms without changing the stability of the system. The comparison is carried out in two test cases. In the analytical test case, ARX and ARMAX can reproduce the training signal very well after an order selection and capture the flutter boundary. For the second test case, using CFD data, ARX and ARMAX match training signals well, but for test signals, ARMAX shows a lower fitness caused by overestimation of error contribution. For the nonlinear training, the optimiser can follow the nonlinear behaviour of the reference output, which demonstrates significant error reduction compared to the linear model. The validation of the ROM is examined for different training signals: chirps, random phase multi-sine, sequential sinusoids and multi-chirps. Chirps and sequential sinusoids fail to reproduce VDP. For most cases, random phase multi-sine is unable to follow VDP oscillation. Although the bounded oscillation is predicted for certain cases, repetitive tests are not consistent. For the multi-chirps, the optimiser is changed to accustom the cost function. With a couple of tests, this signal can deterministically reproduce the bounded oscillation, but the accuracy is not high. Another assumption that typical frequencies at flutter and LCO are not far apart is considered. Sequential sinusoids with narrow frequency band and wide amplitude range are applied to construct the nonlinear ROM in both random and deterministic cases. The frequencies and amplitudes are randomly chosen with the frequencies predetermined in the random case, while frequencies and amplitudes in the deterministic case are scattered uniformly. The LCO behaviour is predicted quite well in terms of amplitude and frequency in the bounded phase but not in transitional stage for both the cases. Repetitive tests are required for the random case. This study applied ARX and ARMAX as the linear ROM for flutter prediction and built the polynomial-based nonlinear ROM for predicting LCO. The observation that narrowing frequency range and widening amplitude range increase the chance of capturing LCO can be used for nonlinear training signal design.Aerospace EngineeringAerodynamics, Wind Energy & Propulsio
High-Order Numerical Schemes for Compressible Flows
No full textHigh-order numerical methods for Computational Fluid Dynamics have undergone significant fundamental developments over the last two decades owing to combined efforts from the applied mathematics and engineering communities. Even though low-order numerical methods are still the standard in industry, the increased requirements of engineering applications have led to significant scientific interest in developing efficient and robust numerical methods. Applications that would benefit from high-order numerical methods include Direct Numerical Simulations (DNS), Large Eddy simulations (LES), Computational Aero-Acoustics (CAA) and vortex dominated flows. The objective of this thesis is to successfully implement and validate a fifth order traditional WENO scheme in a finite volume framework, for a solver currently being developed in the Aerodynamics group of TU Delft. A detailed literature study of classical numerical schemes has been performed along with a study of the traditional WENO schemes. The quality of results using the fifth order scheme is studied for a variety of test cases to study the shock capturing ability of the scheme. Implementing the finite volume WENO schemes includes the calculation of numerical flux at cell faces using Gaussian quadrature formulas. The effect of varying the number of Gaussian quadrature points while calculating the numerical flux is investigated. Also, the effect of the approximate Riemann solvers on the quality of results is studied by implementing four different Riemann solvers and studying the results for different test cases using these Riemann solvers. The test cases are governed by the inviscid Euler equations and deals with flow in the compressible regime. They involve shocks, other discontinuities and often also complicated structures in the smooth part of the solution which tests the design of the schemes to be non-oscillatory at the discontinuities and still gives a high order of accuracy in the smooth parts of the flow. Convergence tests of the error for test cases using the linear advection equation is used to study the order of accuracy of the scheme using different number of Gaussian quadrature points. The tests clearly show that the order of accuracy remains the same irrespective of the number of quadrature points used. This result is important as it allows simulation run with just one quadrature point which is less expensive, and saves memory. This result is highly relevant while running test cases for LES where very fine grids have to be used. WENO schemes have been considered to be too dissipative for LES in their traditional form. This is indeed true as seen by Kelvin-Helmholtz type small scale vortices (which are characteristic of high Reynolds number flows), even in the test cases using the inviscid Euler equations, due to the inherent dissipation in the schemes. However, this could be seen as motivation for using the WENO schemes for Implicit LES where no explicit sub-grid scale models are used to represent the unresolved scales. The different Riemann solvers exhibit different levels of dissipation and recommendations are made for the choice of Riemann solvers according to the application.Aerospace EngineeringFlight Performance and Propulsio
Time-Supersampling 3D-PIV Measurements by Vortex-in-Cell Simulation
No full textMeasurement rate limitations of time-resolved 3D-3C velocity measurements by tomographic PIV limit application of the technique to small measurement volumes and low speed flows (~10 m/s). To reduce the challenging repetition rate requirements historically set by the Nyquist criterion, in the present thesis work a novel method is proposed, combining PIV measurements with numerical simulation of the vorticity transport equation using a hybrid vortex particle discretization. The principle of the time-supersampling method is that the spatial information available by the measurements can be leveraged to increase the temporal-resolution. The solution of the governing equations is based on the Vortex-in-Cell (VIC) method and the unsteady numerical simulation of the temporal evolution of the measured flow is applied within the 3D measurement domain. Both forward and backward time-integration is performed between pairs of consecutive measurements. The accuracy of the proposed time-supersampling method is studied with two experimental datasets obtained from time-resolved tomographic PIV measurements: a turbulent wake, and a circular jet. The results are compared to linear interpolation, advection-based supersampling, and measurement data at high sampling rate. In both flows the ability to reconstruct detailed temporal dynamics from data sampled at a rate far below the Nyquist frequency is demonstrated. The study demonstrates that measurement rate requirements can be strongly reduced when the measurements are super-sampled with the proposed time-supersampling method, thereby extending the range of application of tomographic PIV. In addition, an alternative application of the approach in the field of noise reduction and application to instantaneous measurements is illustrated. The latter can on the one hand allow for a significantly improved predictor for fluid trajectory correlation methods and on the other hand when validated can pose a radically simplified approach for calculation of the instantaneous and unsteady pressure field from single tomographic PIV snapshots, in comparison to multi-pulse systems.AerodynamicsAerospace Engineerin
Aerodynamic analysis of cowling misalignment on a two-man bobsleigh
No full textBobsleighing is one of the fastest winter sports and races are decided within seconds over total times up to four minutes. Therefore marginal gains in performance of the bobsleigh and its crew can determine the race outcome. In order to improve the performance of a bobsleigh the aerodynamic drag can be investigated. A bobsleigh consists of two parts which are connected pivotally: the nose and the rear part of the sleigh. In a track curve the bobsleigh nose will be misaligned with the rear cowling. This phenomena is believed to influence the aerodynamic drag of a bobsleigh. Therefore the objective of this research is to investigate the effect of cowling misalignment between front and rear cowling on the aerodynamic drag of a two-man bobsleigh. To analyse this effect on the aerodynamic drag two types of drag measuring campaigns are performed: wind tunnel experiments and numerical simulations. Flowvisualisation (bothwoollen tuft measurements and particle image velocimetry) is performed and force measurements are executed on a simplified bobsleigh model in the wind tunnel. As bobsleigh features such as bumpers, skies and frame are influencing the flow behaviour these are not integrated in the model such that the measurements solely focus on the effect of nose rotation. Next to these wind tunnel measurements numerical simulations are performed on the samemodel to predict and visualise the internal and external bobsleigh flow behaviour. A bobsleigh is a bluff body which means that the drag is mostly influenced by the geometry and flow separation causing a pressure difference between the flow in front of and the flow trailing the body. This pressure difference causes a suction force on the bobsleigh which accounts for the major drag component. As the air flows over the bobsleigh it separates behind the head of the pilot and at the cowling edges. A low pressure region is therefore seen behind the pilot and brakeman which causes the the external streamlines to be sucked into the cavity. This causes a low pressure wake with rotational flows. It is shown that the aerodynamic drag increases with nose rotation. This increase is partly due to the fact that the frontal area increases with nose rotation. The flow stagnates on one side of the body whereas an overlap is found on the opposite site of the bobsleigh. Behind this overlap a separated region develops where air enters the nose due to backflow in the separation bubble behind the nose on the rear body. This separated flow reattaches to the rear body such that an effective frontal area increase increases the aerodynamic drag. However, it is found that this effective frontal area increase is not the only contributor to the total drag. The air entering the nose influences the internal flow behaviour causing a loss inmomentumdue to the internal blockage imposed by both the crew and the equipment. As the distance between the nose and rear cowling increases the aerodynamic drag increases as well. In order to minimise the influence of nose rotation on the aerodynamic drag a shape optimisation of the nose is recommended. This can lead to better aerodynamic performance of the bobsleigh both on a straight track segment as in a curve.Aerospace EngineeringAerodynamics, Wind Energy & Propulsio
Data driven modeling of junction flows
No full textA wall resolved LES simulation of the Anti-Fairing wing/body junction introduced by Belligoli et al. [6] to reduce interference drag is performed. The LES mesh is composed of 61.7 million cells with a C-fitted grid around the wing . The simulation is performed using the pimple solver of Open- Foam 4 with a time and space varying inlet boundary condition obtained thanks to a precursor. This simulation will be used to assess the impact of the Anti-Fairing by comparing the result to the wall resolved baseline case of Alberts [2] and to serve as a training data for data driven techniques applied to junctions flows. Using the wall resolved LES we apply the data driven algorithm method Sparse Regression of Turbulent Stress Anisotropy (SpaRTA) developed by Schmelzer et al. [38] in the case of junction flows. It is shown that the first step of the method, namely the k-corrective frozen RANS, is able to produce corrective fields to the Reynolds tensor and the turbulent kinetic energy equations in this case. The corrective fields once added in a k-omega SST simulation make it possible to obtain the exact location, strength and shape of the main horseshoe vortex. The upstream boundary layer is also subject to corrections indicating RANS-LES mismatch in the inflow. Mutual Information (MI) is calculated to identify the relevant tensors, physical features and invariants that correlate with the junction flow data. Finally, algebraic models for the corrective fields are obtained. They are compared to the true values of these fields. It is possible to see that the performance of SpARTA models is good upstream of the wing. However, models found and tested in the vicinity of the wing, where the separation and horseshoe vortex are located, are not fully able to capture the relevant corrections. Additional constraints or steps to the ones performed in the time of this study may be necessary in order to use SpaRTA to generate models giving improved predictions compared to classic RANS turbulence models.Aerospace Engineerin
Goal oriented adaptation of unstructured meshes: Application to finite volume methods
No full textIn the present thesis report the author synthetizes 9 months of work at the DSNA department of ONERA in Chatillon, France. The topic of the thesis is goal oriented mesh adaptation with particular application to unstructured grids and to _nite volume methods. The motivation of the present work is the application to unstructured meshes of a novel indicator for mesh adaptation, based on the total derivative of the goal function with respect to mesh nodes coordinates, introduced by Peter et al. in [6], [7] and that has been tested until now on structured grids only. In chapter 1 a brief literature survey is presented, with the aim of introducing the theoretical background of the work and the state of the art of goal oriented mesh adaptation techniques. In chapter 2 the author gives a description of the gradient computation module of the CFD software elsA, developed by ONERA, that has been the main tool used for ow simulations and mesh adaptation and the core of the code development work.Aerospace Engineering | Aerodynamics and Wind Energ
- …
