Figure 9 - uploaded by Ney Rafael Secco
Content may be subject to copyright.
Trailing edge of wing-fuselage junction after each optimization problem. Red regions indicate reversed flow. The redesigned fairings eliminate the recirculation bubble in both optimizations.
Source publication
Mesh generation for high-fidelity computational fluid dynamics simulations and aerodynamic shape optimization is a time-consuming task. Complex geometries can be accurately modeled using overset meshes, whereby multiple high-quality structured meshes corresponding to different aircraft components overlap to model the full aircraft configuration. Ho...
Similar publications
It is the tradition for the fluid community to study fluid dynamics problems via numerical simulations such as finite-element, finite-difference and finite-volume methods. These approaches use various mesh techniques to discretize a complicated geometry and eventually convert governing equations into finite-dimensional algebraic systems. To date, m...
Citations
... Implementing adjoint methods is instrumental in streamlining the optimization process, offering a solution to the intricacies of achieving optimal designs in aerospace engineering. Adjoint methods, originally pioneered in fluid dynamics 4 and control theory, 5 have found extensive application in aerodynamic design optimization, [6][7][8][9][10][11][12][13] as well as in diverse engineering domains such as hydrodynamics, 14,15 heat transfer, 16,17 and structural optimization. 18,19 Various methodologies model adjoint equations for continuous and discrete Partial Differential Equation (PDE) solvers. ...
The adjoint method is a powerful tool in high-fidelity aerodynamic shape optimization, providing an efficient means to compute derivatives of a target function with respect to various design variables. This paper delves into the discrete adjoint method. It offers a theoretical exploration of its implementation as an innovative tool for calculating partial derivatives (sensitivities) related to objective functions and design variables, specifically applied to a subsonic NACA0012 airfoil. The study conducts a qualitative evaluation using a designated test case, considering specified Mach number and Reynolds number values of 0.297 and 6,667 million, respectively. The Spalart-Allmaras turbulence model is employed to enhance computational cost efficiency. The results affirm the efficacy of the introduced tool, DAFoam, showcasing its ability to generate optimal geometries. The achieved performance optimization is evidenced by minimizing the drag coefficient value to an impressive 0.0131. While this research does not delve into the post-processing of sensitivity calculations, it acknowledges the potential for future investigations. The primary objective and novelty of this study is to provide the elementary background of the state of the art test case (NACA0012) within the subsonic regime, introducing the pioneer discrete adjoint aerodynamic optimization methodology (DAFoam) with the potential to explore its higher order capabilities in other aerodynamic related studies. Furthermore, it caters the educational needs of both graduate students and engineers in this exciting field. By presenting this cutting-edge methodology, it contributes to future advancements for the aerodynamicists in terms of optimal solutions.
... MACH-Aero is a gradient-based framework that integrates the CFD solver, geometry parametrization, mesh perturbation, optimizer, and some other base classes relevant to ASO. The tool suite shows good performance in the ADODG standard cases [24][25][26] and other state-of-art ASO research [27][28][29]. As mentioned above, the geometry of interest is the CRM model. ...
... RANS models have been heavily integrated into a number of optimization research. Numerous RANS-based design optimization applications have been conducted to the design of airplanes [8][9][10][11][12][13][14][15], wind turbines [16][17][18][19], and turbomachinery [20][21][22]. According to these studies, multidisciplinary design optimization employing a CFD solver based on RANS yields reliable results to discover crucial aspects of the design. ...
... Three-dimensional RANS equations are used to model the aerodynamics of the coupled aeropropulsive problem. In ADflow, an implicit hole cutting scheme is used to compute the overset mesh connectivities [8]. Martins [33] points out that the CFD solver must be capable of solving a wide range of shapes during the optimization to prevent optimization failure. ...
Over-wing nacelle (OWN) configurations have two advantages over the traditional under-wing nacelle (UWN) configurations: 1. they make it easier to integrate high bypass ratio (BPR) and ultra-high BPR engines by removing ground clearance concerns, and 2. they significantly reduce noise because the wing blocks the noise of the fan and the jet. Despite their benefits, OWN technology is not widely used in real aircraft designs because of the challenges and lack of research in aeropropulsive integration of the propulsion system. In this work, we present multipoint aero-propulsive design optimizations of an OWN configuration to address several scenarios. To effectively design the propulsor in OWN, we consider on-design and off-design formulations of engine design concepts. In this work, we present one single-point optimization and three multipoint optimizations. In the multipoint optimizations, the objective is to minimize the shaft power at cruise conditions subject to constraints. To compare the performance degradation of single-point optimization, we performed several sub-optimizations to reflect the multipoint cases. The single-point optimal design required additional 0.4% shaft power across the cruise conditions when compared to the multipoint optimization with 3 cruise points. In addition, distortion-constrained and distortion-unconstrained optimizations were performed at the takeoff scenario. By modifying the inlet shape of the nacelle, the distortion-constrained optimization effectively reduces the distortion in the takeoff scenario at the outermost radial ring over the fan by penalizing the cruise performance. The multipoint optimizations in this work highlight a number of significant design effects. An effective design is produced by including the required flight conditions and specifying the proper objective and constraints in the OWN multipoint problem. These design optimization capabilities will be crucial in determining the viability of novel propulsion concepts for next-generation OWN aircraft. Nomenclature A = area ASO = aerodynamic shape optimization 1 BPR = bypass ratio C D = drag coefficient C L = lift coefficient C p = coefficient of pressure D = drag force d = fan diameter FPR = fan pressure ratio f = force vector F fan = force applied by the fan F net = total net thrust F pressure = force due to pressure F propulsor net = propulsor net thrust g geo = geometric constraints h = enthalpy LE = leading edgė m = mass flow rate OWN = over-wing nacelle M = Mach number MDO = multidisciplinary design optimization n cruise = number of cruise points n RTO = number of takeoff or rotation points p = pressure P flow = flow-power added by the fan P heat = heat added by the fan P shaft = total shaft fan power R = residual function T = temperature t = thickness TE = trailing edge UHBPR = ultra high-bypass ratio UWN = under-wing nacelle V = velocity vector X = design variables XDSM = eXtended Design Structure Matrix x shape = shape design variables α = angle of attack κ = KS-aggregated distortion metric (·) * = target value in optimization problems (·) aero = value in the aerodynamic model (·) prop = value in the propulsion model (·) cruise = quantity at cruise conditions (·) ff = quantity at the fan face (·) fe = quantity at the fan exit (·) t = total quantity (·) s = static quantity 2
... Many optimization studies have made extensive use of RANS models [3]. Numerous applications of RANS-based design optimization to the design of aircraft [4][5][6][7][8], wind turbines [9][10][11][12], aeropropulsive optimization [13][14][15][16][17][18] and turbomachinery [19][20][21] have been performed. These investigations have demonstrated that using MDO in conjunction with a CFD solver based on RANS yields reliable outcomes for identifying critical design elements. ...
Multidisciplinary design optimization (MDO) plays a vital role in designing aircraft. Thus, the modularization of various disciplines is necessary to efficiently couple them in MDO or multidisciplinary design analysis (MDA). In this paper, we demonstrate how recent library extensions [1] lead to combining both the HPC ecosystem FlowSimulator, developed by DLR and partners, and the MDA/MDO framework, OpenMDAO, for a gradient-based MDO scenario of a generic static aeroelastic problem for a transport aircraft. In the MDO problem, the software architecture enables the optimal integration of high-fidelity workflows with central data management for high performance computing (HPC). To this end, we demonstrate how the FlowSimulator Data Manager (FSDM) enables the associated HPC simulation components to exchange massively parallel data in memory. Here, the algorithms provided by OpenMDAO-linear and nonlinear solution methods, and forward and backward processing of sensitivity derivatives-operate on the infrastructure of the FlowSimulator HPC ecosystem. Based on the defined scenario, different solution strategies can be applied. In this work we define a nested aeroelastic coupling setup within a trim equilibrium constraint definition. In particular, we highlight how an efficient modularization of the individual disciplines supports an effective gradient assembly and thus makes them available to the individual solution hierarchies and strategies. With this approach, we show how it is possible to transform a baseline constrained optimization problem into an unconstrained optimization definition, while still ensuring the trim equilibrium through an implicit solution based on the Schur complement method.
... We use the open-source pyhyp (https://github.com/mdolab/ pyhyp) to generate a simulation mesh [33]. The cell number of the calculation mesh is 7:0 × 10 4 , as shown in Figure 2. Figure 3 compares the simulation results and experimental data. ...
The joined-wing configuration has great technical appeal for the development of next-generation SensorCraft. Research based on the simplified tandem airfoil system can improve understanding of the joined-wing configuration’s aerodynamic characteristics. We combine the adjoint-based aerodynamic shape optimization and self-organizing map- (SOM-) based data mining technology to reveal the flow interactions of tandem airfoils and aerodynamic characteristics from the perspective of the entire aerodynamic design space. The SOM is used to explore the correlation between relative position parameters and aerodynamic force coefficients of tandem airfoil systems. Results show that the drag coefficient at the defined range of lift coefficients has obviously positive linear correlation and greatly dependents on the value of decalage. The tandem airfoils with negative decalage around -2.7° have the smallest drag coefficients. Due to variations in the aerodynamic interaction strength, the drag coefficient of each airfoil changes from a linear law to a nonlinear law as airfoils approach each other. We then perform single-point aerodynamic shape optimization based on two sets of relative position parameters with different aerodynamic interaction strengths, and 1.8% and 1.28% drag reductions are obtained, respectively. Based on optimized airfoils, the SOM is used to reveal the distribution of drag variation in the design space constructed by relative position parameters. Results illustrate that the aerodynamic interference strength between the front and rear airfoils significantly affects the drag reduction mechanism, which results in the different distribution patterns of drag variation in design space.
... The 3D modelling of the NASA STARC-ABL aircraft [22] relies on openVSP [190] which is a 3D modular modelling software to generate and optimise geometries [21,169,170,173,174] that are meshed volumetrically via mesh morphing [191]. Nonetheless, although being acknowledged for accurately predicting the physics, structured meshes are computationally expensive and rather complicated to generate. ...
The aviation sector is experiencing increasing pressure to reduce emissions via long-term strategies for a ceaselessly growing number of flight passengers. Aircraft currently in operation have typically been designed by considering the airframe somewhat separately from the propulsion system. In doing so, conventional aero-engine architectures are approaching their limits in terms of propulsive efficiency, with technological advancements yielding diminishing returns. A promising alternative architecture for improving the overall performance of the next generation of commercial aircraft relies upon boundary layer ingestion (BLI). This technology aerodynamically couples the airframe with a strategically positioned propulsion system to purposely ingest the airframe’s boundary layer flow. Nonetheless, there is a lack in consensus surrounding the interpretation and quantification of BLI benefits. This is primarily because conventional performance accounting methods breakdown in scenarios of strong aerodynamic coupling. Subsequently, there is a major challenge in defining appropriate performance metrics to provide a consistent measurement and comparison of the potential benefits. This review examines the various accounting methods and metrics that have been applied in evaluating BLI performance. These are discussed and critiqued in the context of both numerical and experimental models. Numerically, the geometric, aerodynamic and propulsive models are sorted by their orders of fidelity along with the plenitude of methods used for flow feature identification enabling a phenomenological understanding of BLI. Particular attention is then given to experimental BLI models with their different set-ups, methods and associated limitations and uncertainties. Finally, the numerous unconventional BLI aircraft concepts are categorised, compared and critiqued with reference to their associated design exploration and optimisation studies.
... To undertake coupled aeropropulsive design optimization, a computational fluid dynamic solver (CFD) solver with the requisite fidelity, capturing viscous and shock effects, such as Reynoldsaveraged Navier-Stokes (RANS) equations, is necessary. Numerous applications of RANS-based design optimization have been performed [8][9][10][11][12][13][14][15][16][17]. Specifically, several BLI aeropropulsive optimization were carried out using RANS [18][19][20][21][22]. Furthermore, Lamkin et al. [23] performed fully coupled high-bypass engine design optimization using a RANS solver. ...
... We will use 3-dimensional Reynolds-averaged Navier-Stokes (RANS) equations to solve the coupled aeropropulsive problem. The overset mesh connectivities are calculated using the implicit hole cutting scheme [8] in ADflow. Fig 4 shows a mesh-slice at the center of nacelle that has around 2.6 million cells. ...
Optimal nacelle placement is critical to commercial or transport aircraft designs that use engines mounted with nacelles. Over-wing nacelle (OWN) configuration has a high potential to improve upon the conventional under-wing nacelle (UWN) configuration. OWN configurations have two critical benefits when compared to the conventional UWN configuration: 1. they ease the integration of high BPR and ultra-high BPR engines by alleviating ground clearance issues, and 2. they provide a significant amount of noise reduction because the wing blocks the noise of the fan and the jet. Despite their advantages, the OWN technology is not used in practical aircraft designs due to the difficulties in the aeropropulsive integration of the propulsion system. In this work, we propose using a coupled aeropropulsive design optimization framework to study the aeropropulsive integration for OWN configurations. The coupling behavior between the aerodynamics of the wing and the propulsion system is extremely important. Especially in OWN configurations, the propulsion system is highly influenced by the aerodynamic performance of the wing. In this paper, we address the design problem of coupling between the aerodynamic and the propulsion system. We also explore wide range of design space to achieve the best practical design. The changes in wing shape are insignificant when the OWN configuration is optimized for different FPR. Wing in OWN performs better when the propulsor is close to the wing's root; however, the overall drag increases when the propulsor is close to the root. This increases the shaft power need. In addition, nacelle placement at the aft of the trailing edge shows better performance than forward of the trailing edge of the wing. These advancements in design optimization capabilities will be critical in the future design of OWN configurations to achieve more environmentally sustainable aircraft designs.
... More recently, researchers have developed boundary element methods for analyses and designs of hydrofoils and propellers [23][24][25][26][27][28] Despite that the low-fidelity methods and boundary element methods can capture the trends with low computational cost, high-fidelity simulations are needed to more accurately capture critical physics, such as cavitation and separation [29][30][31]. In the numerical design optimization context, Reynolds-averaged Navier-Stokes (RANS) model is commonly used because it is currently the highest-fidelity approach that is still computationally tractable [32][33][34][35][36][37]. However, high-fidelity models are computationally prohibitive for design problems requiring many simulations, particularly for complex multipoint loading. ...
... For a geometry with intersecting components, the surface mesh deformation near the junction is challenging because the CFD solver requires the surface mesh nodes to conform with the changed design outer mold lines and maintain the watertight property [43]. Secco et al. [35], Secco and Martins [44] developed a robust algorithm to handle this mesh deformation with overset meshes and demonstrated the advantages of using a wing-body configuration and a strut-braced wing. ...
... To achieve this goal, we use the surface mesh deformation method developed by Yildirim et al. [43], which is based on the pySurf module developed by Secco et al. [35]. To parameterize the geometry of the configuration that includes multiple components that intersect, we first use separate FFD volumes to parameterize the design of the foil and the strut separately. ...
Hydrodynamic lifting surfaces usually include junctions. High-fidelity simulations are necessary to capture critical physics near these regions, such as separation, junction vortices, and cavitation. We present RANS-based hydrodynamic optimizations of a T-shaped hydrofoil, including changes in the junction geometry. The optimized hydrofoils avoid separation and delay cavitation compared to the baseline. The full optimization design with planform, cross-section, and junction geometry variables yields a total drag reduction of 6.4%. The optimized results show that the relative locations of the maximum foil thickness and the maximum strut thickness significantly impact the junction cavitation. Including the translation between the strut and the foil and more strut geometric variables as design variables will provide further improvement. The comparison between optimized designs demonstrates that optimizing planform and detailed junction geometry provides further improvement in addition to designing the cross-sectional geometry. The hydrostructural analyses show that the optimized T-foils have lower stress at the junction than the baseline because of the resultant junction fairing. However, these hydrodynamic-only optimized T-foils have higher deformation and maximum stress, which could result in accelerated fatigue, highlighting the need for hydrostructural responses in design optimization. The results demonstrate that the developed methodology is useful for designing next-generation complex hydrodynamic lifting surfaces.
... ADflow solves the RANS equations with the Spalart-Allmaras turbulence model and has shown to be rather robust even the aerodynamic shape is abnormal [34]. We choose ADflow as the CFD solver because it has an efficient adjoint solver [35,36], based on which many state-of-the-art adjoint-based aerodynamic shape design works [37][38][39][40] are performed. We use the adjoint suite to verify the efficacy of the proposed adjoint-free optimization method in high-dimensional wing design, which can be more reliable and persuasive. ...
Gradient-based optimization is a calculus-based point-by-point technique that relies on the gradient (derivative) information of the objective function with respect to a number of independent variables. The nature in which gradient-based methods (GBM) operate make them well suited to finding locally optimal solutions but may struggle to find the global optimal. With gradient-based algorithms an understanding of the design space is assumed, as an appropriately pre-conceived starting design point must be given. If large changes in topology are expected lower fidelity panel codes can facilitate useful optimization procedures.
... Secco et al. [38] developed the most flexible and advanced approach in this field using overset meshes and later used this method to optimize the design of a strut-braced wing [39]. In this approach, the surface collar mesh that conforms to the intersection of the two components is re-computed at every design iteration. ...
... These surface mesh nodes should be sufficiently close to the actual design OML so that the distances of these points to the CAD surface are within the meshing tolerance. The proposed method is implemented using the pySurf module developed by Secco et al. [38], which contains a wide range of geometric operations defined on triangulated surfaces and piecewise linear curves. Furthermore, the operations in this module are differentiated using AD, and we use these differentiated routines to obtain the derivatives of the resulting surface mesh coordinates with respect to the design variables. ...
... The baseline design has a large area of flow separation on the upper skin of the wing near the fuselage at the cruise condition. There are several drag computation studies for this configuration [46,57,58], along with efforts to design a wing-fairing to reduce the amount of separation near the fuselage intersection [35,36,38,59]. ...
Aerodynamic shape optimization based on computational fluid dynamics (CFD) requires three steps: updating the geometry based on the design variables, updating the CFD surface mesh for the new geometry, and updating the CFD volume mesh based on the new surface mesh. While there are many tools available for the first and third steps, the methods available for the second step are insufficient for geometries that have intersecting components. For these geometries, the CFD surface mesh needs to be updated near component intersections to conform to the component geometries and the updated intersection curves. To address this need, we introduce a method that can deform the CFD surface mesh nodes near component intersections. The method can handle arbitrary design changes for each intersecting component as long as the geometric topology is unchanged. Furthermore, the method is suitable for gradient-based optimization because it smoothly deforms every CFD surface node without introducing topological changes in the CFD surface mesh. In this paper, we detail each step of the proposed method and visualize the range of design changes that can be achieved with this approach. Finally, we use the proposed method in an aerodynamic shape optimization problem to optimize the wing-body intersection of the DLR-F6 configuration. These results demonstrate the effectiveness of the proposed method in a high-fidelity design optimization framework. The method applies to both structured and unstructured CFD meshes and makes it possible to use computer-aided design and conceptual design geometry tools within high-fidelity design optimization.
