Aerodynamic shape optimization of the DLR-F6 wing by using openfoam as CFD solver integrated with rsm

Abstract

Aerodynamic shape optimization plays a critical role in aerospace engineering as it allows designers to enhance aerodynamic performance by altering the shape of a body. The ability to optimize the shape of structures like aircraft wings, wind turbine blades, and rockets can lead to increased efficiency, reduced fuel consumption, and minimized emissions. Given the pressing need to address climate change and the exponentially escalating global crisis, it's essential to prioritize sustainable solutions in every aspect of design, including minimizing the impact of aircraft emissions. This requires a primary focus on decreasing the drag and increasing the lift on the airplane, which is one of the most significant factors affecting aerodynamic performance and range. As air traffic continues to grow, the importance of aerodynamic shape optimization in reducing emissions and increasing fuel efficiency becomes increasingly clear. This thesis on the aerodynamic shape optimization of the DLR-F6 wing demonstrates an effective and a comprehensive way of an optimization process that can contribute to the ongoing research in this field. The DLR-F6 wing is a common benchmark for aerodynamic research due to its complex geometry and challenging flow characteristics. By optimizing the shape of the wing, it is aimed to improve its performance and contribute to ongoing research in this field. To ensure that the structural strength of the wing is not compromised, the optimization process also involves some considerations on various design constraints like modal frequency and mass of the wing. In the optimization process, different chord slices at various locations along the span has been taken and twisted some angles by taking their aerodynamic centers as reference. The work focuses on to determine the best improved angle sets which let the better performance on L/D value without sacrificing its structural integrity. The optimization model tree was constructed using ModeFRONTIER software, which integrated different software tools and automated the optimization process. The construction involved four main stages. Firstly, Pointwise software was used to create a new wing database by twisting at the six sliced chords with angles determined by the software. The software automatically executed all the steps to twist, create a new database, make surface mesh, and create a pre-meshed geometry for Abaqus, with the help of a journal. In the second stage, volume meshes were prepared using ANSYS Fluent Meshing, which automatically executed all the meshing processes controlled by optimization software. Thirdly, CFD analysis was conducted using OpenFOAM as the CFD solver to simulate the flow around the wing. The volume mesh created by Fluent Meshing was used as the solution cells. With the help of a function, OpenFOAM can convert a fluent mesh to foam format. To make the optimization process faster and well talent based, the HPC (High Performance Computer) was used to run in OpenFOAM. ModeFRONTIER makes an automatic connection with HPC systems based on SSH protocol and with a Linux bash script, aerodynamic analysis had been conducted. After the simulation was done, a file storing all the forces at each iteration was transferred to the host computer, and lift, drag and L/D values were computed using an inner MATLAB stage. The L/D value and lift value were set as design objectives. The purpose was the maximize these values. Since the twisting rotation angles were set as the input values, the optimization tool organizes and selects the best input values to succeed the design objectives. Lastly, Abaqus software was used to perform structural analysis to ensure the strength of the wing, with the pre-meshed geometry file directly transferred after the first stage. With an Abaqus journal, for each different design mass and modal frequency are calculated and processed with another MATLAB stage. These were selected as another design objective, with the goal being to minimize the mass and maximize the frequency value. However, 'Lift' and 'L/D' objectives were prioritized to make the optimization processes easy to go on. The optimization process was streamlined through these main stages, with simultaneous file transfers and evaluation of results made in intermediate steps. About 220 DoE (Design of Experiments) were created and evaluated. Due to the expensive CFD simulations, direct optimizations were not feasible to proceed. Therefore, RSM (Response Surface Methodology) was used to reproduce more experiments in an inexpensive way. RSM, also known as Surrogate Models, are a collection of statistical and mathematical techniques used to create, model, and analyze the relationships between input variables and output responses. By using RSM, the number of experiments needed can be reduced to obtain optimal results, saving time and resources in the optimization process. The direct optimization results were used to train the data set to build a good quality RSM. After using good RSM strategies, 1000 more experiments were created. After selecting the best design, a real analysis was conducted and showed that the RSM predicted the design output very well. The results of the optimization process were evaluated using a set of performance metrics, including the L/D ratio, maximum lift, minimum drag. Since the mass and modal frequency objectives were required to assure the structural integrity of the new design, they were not included as optimization performance metrics. The modified DLR-F6 aircraft was then compared with the original DLR-F6 aircraft using these performance metrics. The results show that the modified aircraft had a 13.15% improvement in L/D ratio and 2.24% improvement in lift compared to the original aircraft. A 3DOF flight simulation was done using MATLAB Simulink tool, with two aerodynamic databases created by sweeping 2 Mach numbers and 13 angle of attacks. One was for the airplane with the original wing, and the other was for the airplane with the optimized wing. Overall, the thesis demonstrates the effectiveness of using a multi-disciplinary optimization approach to improve the performance of complex aerodynamic shapes such as the DLR-F6 wing. The optimized wing not only shows a significant improvement in its aerodynamic performance but also maintains its structural strength. Although the flow chart may seem complex, it helps make the optimization process comprehensive and efficient. The optimization process and methodology used in this research can be applied to other complex aerodynamic shapes to improve their performance as well.