The goal of "greener" aviation is one of the main challenges in aircraft design. The target of Europeans "Flightpath 2050'' and IATA is to reduced the net aviation CO2 emission by 75% relative to 2000 and 50% relative to 2005, respectively. Novel unconventional aircraft claim to increase the efficiency and reduce the environmental impact. Designs differing from the conventional tube-low-wing concept are investigated regarding their performance benefit. The employment of a high aspect ratio wing is an effective way to increase the aerodynamic efficiency. However, the long and slender wing structure is more flexible and thus more prone to aeroelastic effects. Critical phenomena, such as flutter and limit-cycle oscillation are more likely to drive the design. Therefore it is important to assess the interdependence of aerodynamic and structural forces. The effects of the wings flexibility can affect the design and off-design performance, possibly jeopardizing the intended efficiency benefit. To evaluate the different disciplines involved in aircraft design, a multi-disciplinary design optimization environment offers the required tools. While computationally demanding, the obtained solution is more efficient if the disciplines are assessed simultaneously. Equipped with low- and high-fidelity assessments, aircraft performance can be predicted at the preliminary design stage, while mitigating some computational expenses. To further reduce the computational burden, adaptive surrogate modelling approaches can be employed, requiring less computational evaluations while efficiently guiding the optimization process towards improved designs. Considering surrogate models for expensive physics based objective and constraint functions bears the disadvantage of more uncertainty in the models. Thus, a new technique is proposed to utilizing the probability of feasibility for the constraints in combination with a transformed normalized objective function to address the uncertainty consideration. The approach is assessed via mathematical test functions and an engineering application and compared against established methods. The results suggests an applicability of the method, with further improvements to be examined. Limitations are revealed regarding local optima and convergence. Further, the degree of maturity does not yet suffice for industrial applications. In a multi-disciplinary design optimization of a high aspect ratio wing aircraft and a strut braced wing aircraft a more classical EGO approach was therefore the choice of approach. The configurations were optimized towards a multi-objective, blending manufacturing and operational costs. Towards cost efficient evaluations, investigations were performed to incorporate high-fidelity assessments, yet limiting their number by reducing active constraints. Driven by aero-structural and aeroelastic constraints, the novel designs could improve the performance satisfactory. / Graduate

<p dir="ltr">Neural networks (NNs) effectively address the challenges of engineering design optimization by using data-driven models, thus reducing computational demands. However, their effectiveness depends heavily on hyperparameter optimization (HPO), which is a global optimization problem. While traditional HPO methods, such as manual, grid, and random search, are simple, they often fail to navigate the vast hyperparameter (HP) space efficiently. This work examines the effectiveness of integrating Bayesian optimization (BO) with multi-armed bandit (MAB) optimization for HPO in NNs. The thesis initially addresses HPO in one-shot sampling, where NNs are trained using datasets of varying sample sizes. It compares the performance of NNs optimized through traditional HPO techniques and a combination of BO and MAB optimization on the analytical Branin function and aerodynamic shape optimization (ASO) of an airfoil in transonic flow. Findings from the optimization of the Branin function indicate that the combined BO and MAB optimization approach leads to simpler NNs and reduces the sample size by approximately 10 to 20 compared to traditional HPO methods, all within half the time. This efficiency improvement is even more pronounced in ASO, where the BO and MAB optimization use about 100 fewer samples than the traditional methods to achieve the optimized airfoil design. The thesis then expands on adaptive HPOs within the framework of efficient global optimization (EGO) using a NN-based prediction and uncertainty (EGONN) algorithm. It employs the BO and MAB optimization for tuning HPs during sequential sampling, either every iteration (HPO-1itr) or every five iterations (HPO-5itr). These strategies are evaluated against the EGO as a benchmark method. Through experimentation with the analytical three-dimensional Hartmann function and ASO, assessing both comprehensive and selective tunable HP sets, the thesis contrasts adaptive HPO approaches with a static HPO method (HPO-static), which uses the initial HP settings throughout. Initially, a comprehensive set of the HPs is optimized and evaluated, followed by an examination of selectively chosen HPs. For the optimization of the three-dimensional Hartmann function, the adaptive HPO strategies surpass HPO-static in performance in both cases, achieving optimal convergence and sample efficiency comparable to EGO. In ASO, applying the adaptive HPO strategies reduces the baseline airfoil's drag coefficient to 123 drag counts (d.c.) for HPO-1itr and 120 d.c. for HPO-5itr when tuning the full set of the HPs. For a selected subset of the HPs, 123 d.c. and 121 d.c. are achieved by HPO-1itr and HPO-5itr, respectively, which are comparable to the minimum achieved by EGO. While the HPO-static method reduces the drag coefficient to 127 d.c. by tuning a subset of the HPs, which is a 15 d.c. reduction from its full set case, it falls short of the minimum of adaptive HPO strategies. Focusing on a subset of the HPs reduces time costs and enhances the convergence rate without sacrificing optimization efficiency. The time reduction is more significant with higher HPO frequencies as HPO-1itr cuts time by 66%, HPO-5itr by 38%, and HPO-static by 2%. However, HPO-5itr still requires 31% of the time needed by HPO-1itr for the full HP tuning and 56% for the subset HP tuning.</p>

The primary objectives of this work are two-fold. First, additive manufacturing (AM) and rapid prototype (RP) testing are evaluated for use in production of a wind tunnel (WT) models. Second, an approach was developed to incorporate stability and control (SandC) WT data into aircraft conceptual design multidisciplinary design optimization (MDO). Both objectives are evaluated in terms of data quality, time, and cost. FDM(TM) and PolyJet AM processes were used for model production at low cost and time. Several models from a representative tailless configuration, ICE 101, were printed and evaluated for strength, cost and time of production. Furthermore, a NACA 0012 model with 20% chord flap was manufactured. Both models were tested in the Virginia Tech (VT) Open-Jet WT for force and moment acquisition. A 1/15th scale ICE 101 model was prepared for manufacturing, but limits of FDM(TM) technology were identified for production. An approach using WT data was adapted from traditional surrogate-based optimization (SBO), which uses computational fluid dynamics (CFD) for data generation. Split-plot experimental designs were developed for analysis of the WT SBO strategy using historical data and for WT testing of the NACA 0012. Limitations of the VT Open-Jet WT resulted in a process that was not fully effective for a MDO environment. However, resolution of ICE 101 AM challenges and higher quality data from a closed-section WT should result in a fully effective approach to incorporate AM and RP testing in an aircraft conceptual design MDO. / Master of Science

Thesis (MEng)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: We investigate the simultaneous analysis and design (SAND) formulation of the topology optimisation problem. The characteristics of the formulation are presented considering the simple compliance/weight constrained problem and the more complex local stress constrained case. The problems are solved in an efficient sparse sequential approximate optimisation (SAO) framework with the SAND formulation showing an significant reduction in computational requirements compared to the traditional and inherently expensive nested analysis and design (NAND) approach. In SAND the state equations are included in the optimisation problem as a set of equality constraints and not solved exactly in each iteration, as would be the case in NAND. Decision and state variables are thus independent, resulting in an immensely sparse optimisation problem. The availability of simple exact analytic expressions for all the constraint functions (via the finite element method) allows for the construction of accurate approximate subproblems with little computational effort. Furthermore, material can be removed completely from the design domain with few complications, resulting in a decrease in subproblem size as the algorithm progresses, further reducing computation time. The inclusion of void material in the design domain leads to the formulation of stress constraints as so-called ‘vanishing’ constraints. Furthermore, the SAND formulation provides a new perspective on the infamous singularity problem. Amongst other results, we present some test cases that seem to scale linearly in computational requirements for a specific range of problem sizes. / AFRIKAANSE OPSOMMING: Die formulering van die topologie optimerings probleem as ’n gelyktydige analise en ontwerp (simultaneous analysis and design (SAND)) formulering word ondersoek. Die eienskappe van die formulering word bespreek in die konteks van die eenvoudig begrensde styfheid/gewig geval en die meer komplekse plaaslike spanning begrensde geval. Die probleme word opgelos in ’n sekwenti¨ele benaderde optimering (SBO; sequential approximate optimisation (SAO)) raamwerk met die SAND formulering, wat lei tot ’n wesenlike vermindering in berekenings vereistes benodig in vergelyking met die tradisionele en inherente duur geneste analise en ontwerp (nested analysis and design (NAND)) geval. In SAND word die vergelykings wat die respons van die struktuur beskryf met gelykheidsbegrensings in die optimerings probleem verteenwoordig. Die respons van die struktuur word dus nie presies opgelos in elke iterasie nie, soos in die geval van NAND wel gebeur. Alle optimerings veranderlikes is dus onafhanklik en lei tot ’n baie yl optimerings probleem. Deur middel van die eindige element metode is die analitiese vorm van alle begrensings beskikbaar en kan dit gebruik word om akkurate benaderde subprobleme op te stel sonder ekstra berekenings koste. Verder kan materiaal heeltemal verwyder uit van die ontwerpsgebied met weinig komplikasies. Dit lei tot ’n verkleining van subprobleme soos die algoritme vordering maak wat berekenings tyd nog meer verminder. Die feit dat materiaal heeltemal verwyder kan word van die ontwerp gebied lei tot die formulering van spannings begrensings as sogenaamde ‘verdwynende’ begrensings. Verder gee die SAND formulering ’n nuwe uitsig op die bekende singulariteitsprobleem. Met verskeie ander resultate word daar ook gewys dat dit voorkom of ’n spesifieke stel toetsprobleme lineˆer skaal in berekenings tyd.

This thesis consists of three chapters, covering topics in both the design and modeling aspects of computer experiments as well as their engineering applications. The first chapter systematically develops a new class of space-filling designs for computer experiments by splitting two-level factorial designs into multiple layers. The new design is easy to generate, and our numerical study shows that it can have better space-filling properties than the optimal Latin hypercube design. The second chapter proposes a novel modeling approach for approximating computationally expensive functions that are not second-order stationary. The new model is a composite of two Gaussian processes, where the first one captures the smooth global trend and the second one models local details. The new predictor also incorporates a flexible variance model, which makes it more capable of approximating surfaces with varying volatility. The third chapter is devoted to a two-stage sequential strategy which integrates analytical models with finite element simulations for a micromachining process.

No description available.

Cette thèse présente un framework d’optimisation pour la conception hydrodynamique de forme d’hydrofoils. L’optimisation d’hydrofoil par simulation implique des objectifs d’optimisation divergents et impose des compromis contraignants en raison du coût des simulations numériques et des budgets limités généralement alloués à la conception des navires. Le framework fait appel à l’échantillonnage séquentiel et aux modèles de substitution. Un modèle prédictif est construit en utilisant la Régression par Processus Gaussien (RPG) à partir des données issues de simulations fluides effectuées sur différentes géométries d’hydrofoils. Le modèle est ensuite combiné à d’autres critères dans une fonction d’acquisition qui est évaluée sur l’espace de conception afin de définir une nouvelle géométrie qui est testée et dont les paramètres et la réponse sont ajoutés au jeu de données, améliorant ainsi le modèle. Une nouvelle fonction d’acquisition a été développée, basée sur la variance RPG et la validation croisée des données. Un modeleur géométrique a également été développé afin de créer automatiquement les géométries d’hydrofoil a partir des paramètres déterminés par l’optimiseur. Pour compléter la boucle d’optimisation,FINE/Marine, un solveur fluide RANS, a été intégré dans le framework pour exécuter les simulations fluides. Les capacités d’optimisation ont été testées sur des cas tests analytiques montrant que la nouvelle fonction d’acquisition offre plus de robustesse que d’autres fonctions d’acquisition existantes. L’ensemble du framework a ensuite été testé sur des optimisations de sections 2Dd’hydrofoil ainsi que d’hydrofoil 3D avec surface libre. Dans les deux cas, le processus d’optimisation fonctionne, permettant d’optimiser les géométries d’hydrofoils et confirmant les performances obtenues sur les cas test analytiques. Les optima semblent cependant être assez sensibles aux conditions opérationnelles. / This thesis presents a practical hydrodynamic optimization framework for hydrofoil shape design. Automated simulation based optimization of hydrofoil is a challenging process. It may involve conflicting optimization objectives, but also impose a trade-off between the cost of numerical simulations and the limited budgets available for ship design. The optimization frameworkis based on sequential sampling and surrogate modeling. Gaussian Process Regression (GPR) is used to build a predictive model based on data issued from fluid simulations of selected hydrofoil geometries. The GPR model is then combined with other criteria into an acquisition function that isevaluated over the design space, to define new querypoints that are added to the data set in order to improve the model. A custom acquisition function is developed, based on GPR variance and cross validation of the data.A hydrofoil geometric modeler is also developed to automatically create the hydrofoil shapes based on the parameters determined by the optimizer. To complete the optimization loop, FINE/Marine, a RANS flow solver, is embedded into the framework to perform the fluid simulations. Optimization capabilities are tested on analytical test cases. The results show that the custom function is more robust than other existing acquisition functions when tested on difficult functions. The entire optimization framework is then tested on 2D hydrofoil sections and 3D hydrofoil optimization cases with free surface. In both cases, the optimization process performs well, resulting in optimized hydrofoil shapes and confirming the results obtained from the analytical test cases. However, the optimum is shown to be sensitive to operating conditions.

Traditionally design engineers have used the Factor of Safety method for ensuring that designs do not fail in the field. Access to advanced computational tools and resources have made this process obsolete and new methods to introduce higher levels of reliability in an engineering systems are currently being investigated. However, even though high computational resources are available the computational resources required by reliability analysis procedures leave much to be desired. Furthermore, the regression based surrogate modeling techniques fail when there is discontinuity in the design space, caused by failure mechanisms, when the design is required to perform under severe externalities. Hence, in this research we propose efficient Semi-Supervised Learning based surrogate modeling techniques that will enable accurate estimation of a system's response, even under discontinuity. These methods combine the available set of labeled dataset and unlabeled dataset and provide better models than using labeled data alone. Labeled data is expensive to obtain since the responses have to be evaluated whereas unlabeled data is available in plenty, during reliability estimation, since the PDF information of uncertain variables is assumed to be known. This superior performance is gained by combining the efficiency of Probabilistic Neural Networks (PNN) for classification and Expectation-Maximization (EM) algorithm for treating the unlabeled data as labeled data with hidden labels.

Rising fuel cost has motivated increased fuel efficiency for freight trains. At cruising speed,the largest contributing factor to the fuel consumption is aerodynamic drag. As a result of stagnationand flow separation on and around lead and trailing cars, the first and last railcars experiencegreater drag than intermediate cars. Accordingly, this work focused on reducing drag on lead locomotivesby designing and optimizing an add-on nose fairing that is feasible for industrial operation.The fairing shape design was performed via computational fluid dynamic (CFD) software.The simulations consisted of two in-line freight locomotives, a stretch of rails on a raised subgrade,a computational domain, and a unique fairing geometry that was attached to the lead locomotive ineach non-baseline case. Relative motion was simulated by fixing the train and translating the rails,subgrade, and ground at a constant velocity. An equivalent uniform inlet velocity was applied atzero degree yaw to simulate relative motion between the air and the train.Five fairing families-Fairing Families A-E (FFA-FFE)-are presented in this thesis.Multidimensional regressions are created for each family to approximate drag as a function ofthe design variables. Thus, railroad companies may choose an alternative fairing if the recommendedfairing does not meet their needs and still have a performance estimate. The regression forFFE is used as a surrogate model in a surrogate based optimization. Results from a wind tunneltest and from CFD are reported on an FFE geometry to validate the CFD model. The wind tunneltest predicts a nominal drag reduction of 16%, and the CFD model predicts a reduction of 17%.A qualitative analysis is performed on the simulations containing the baseline locomotive, the optimalfairings from FFA-FFC, and the hybrid child and parent geometries from FFA & FFC. Theanalysis reveals that optimal performance is achieved for a narrow geometry from FFC becausesuction behind the fairing is greatly reduced. Similarly, the analysis reveals that concave geometriesboost the flow over the top leading edge of the locomotive, thus eliminating a vortex upstreamof the windshields. As a result, concave geometries yield greater reductions in drag.The design variable definitions for each family were strategically selected to improve manufacturability,operational safety, and aerodynamic performance relative to the previous families.As a result, the optimal geometry from FFE is believed to most completely satisfy the constraintsof the design problem and should be given the most consideration for application in the railroadindustry. The CFD solution for this particular geometry suggests a nominal drag reduction of 17%on the lead locomotive in an industrial freight train.