Dr. Max Sagebaum, Chair for Scientific Computing (SciComp), TU Kaiserslautern

Title:
CoDiPack 2.0: Lessons learned and new ideas

Abstract:

CoDiPack has initially been released 5 years ago. Since then, additions like primal value taping and several helper structures have been added to the code base. Recent developments for SIMD, tasking, OpenMP and DSL support have shown that the initial design of CoDiPack makes it difficult to add such special interest tapes in a maintainable way and therefore CoDiPack is redesigned for the 2.0 release. The focus for the redesign is mainly on the maintainability aspects. In addition, there are also changes for the ease of development and applying modern coding guidelines. The talk will cover the lessons learned, new design choices and a performance analysis of the new implementation.

How to join

The talk is held online via Jitsi. You can join with the link https://jitsi.uni-kl.de/SciCompSeminar_01. Please follow the rules below:

• Use a chrome based browser (One member with a different browser can crash the whole meeting).
• Mute your microphone and disable your camera.
• If you have a question, raise your hand.

More information is available at https://www.rhrk.uni-kl.de/dienstleistungen/netz-telefonie/konferenzdienste/jitsi/.

Steffen Schotthöfer, Institute for Applied and Numerical Mathematics, Karlsruher Institut für Technologie

Title:
Windowing Regularization Techniques for Unsteady Aerodynamic Shape Optimization

Abstract:

Unsteady Aerodynamic Shape Optimization presents new challenges in terms of sensitivity analysis of time-dependent objective functions. In this work, we consider periodic unsteady flows governed by the URANS equations. Hence, the resulting output functions acting as objective or constraint functions of the optimization are themselves periodic with unknown period length, that may depend on the design parameter of said optimization. Sensitivity Analysis on the time-average of a function with these properties turns out to be difficult. Therefore, we explore methods to regularize the time average of such a function with the so called windowing-approach. Furthermore, we embed these regularizers into the discrete adjoint solver for the URANS equations of the multi-physics and optimization software SU2. Finally, we exhibit a comparison study between the classical non regularized optimization procedure and the ones enhanced with regularizers of different smoothness and show that the latter result in a more robust optimization.

How to join

The talk is held online via Jitsi. You can join with the link https://jitsi.uni-kl.de/SciCompSeminar_02. Please follow the rules below:

• Use a chrome based browser (One member with a different browser can crash the whole meeting).
• Mute your microphone and disable your camera.
• If you have a question, raise your hand.

More information is available at https://www.rhrk.uni-kl.de/dienstleistungen/netz-telefonie/konferenzdienste/jitsi/.

Dr. Beckett Y. Zhou, Chair for Scientific Computing (SciComp), TU Kaiserslautern

Title: Simulation and Sensitivity Analysis of a Wing-Tip Mounted Propeller Configuration from the Workshop for Integrated Propeller Prediction (WIPP)

Abstract:

Propeller-driven engines have garnered renewed interest in the aircraft industry due to the emergence of electric and distributed propulsion systems. In addition, propellers mounted on the wing tips have been shown to significantly reduce the wing’s induced drag by attenuating the wing-tip vortex by the propeller slipstream. An undesirable factor associated with propeller-driven engines is their noise level. In this talk, we present the latest simulation results for a wing-tip mounted propeller configuration using URANS, DDES and sliding mesh capabilities of SU2. The simulation reveals a number of important noise sources, namely the interaction noise caused by the turbulent wake and tip vortices from the propeller blades impinging on the wing and the nacelle surfaces transmitting unsteady aerodynamic loads.

How to join

The talk is held online via Jitsi. You can join with the link https://jitsi.uni-kl.de/SciCompSeminar_03. Please follow the rules below:

• Use a chrome based browser (One member with a different browser can crash the whole meeting).
• Mute your microphone and disable your camera.
• If you have a question, raise your hand.

More information is available at https://www.rhrk.uni-kl.de/dienstleistungen/netz-telefonie/konferenzdienste/jitsi/.

Thomas Dick, Chair for Scientific Computing (SciComp), TU Kaiserslautern

Title:
Development of a gradient smoothing method for shape optimization in SU2

Abstract:

For gradient based shape optimization the application of second order derivatives could greatly improve performance. Many possible ways to approximate a Hessian have been investigated in the past. In this talk we will discuss the Sobolev gradient method for incorporating second order derivative information. This talk will give an overview of the mathematical background and show a sequential quadratic programming method for this case. Furthermore, the implementation of this method into the existing shape optimization capabilities of the SU2 multiphysics and design software will be discussed

How to join

The talk is held online via Jitsi. You can join with the link https://jitsi.uni-kl.de/SciCompSeminar_04. Please follow the rules below:

• Use a chrome based browser (One member with a different browser can crash the whole meeting).
• Mute your microphone and disable your camera.
• If you have a question, raise your hand.

More information is available at https://www.rhrk.uni-kl.de/dienstleistungen/netz-telefonie/konferenzdienste/jitsi/.

Myles C. Morelli, Department of Aerospace Engineering, Politecnico di Milano

Title:
Radial basis function mesh deformation for icing and rotorcraft simulations in SU2

Abstract:

This work presents a detailed evaluation of radial basis function mesh deformation for complex iced geometries and rotorcraft simulations. Given the high computational cost of mesh deformation, state-of-the-art radial basis function techniques are used for data reduction. The principle procedures adopted are multi-level greedy surface point selection and volume point reduction. The multi-level greedy surface point selection reduces the control point list to increase the efficiency of the interpolation operation and the volume point reduction improves the computational cost of the volume mesh update operation which is important for large data sets. The study demonstrates the mesh deformation capabilities on two in-flight icing test cases as well as for a AH-1H cobra helicopter main rotor in forward flight.

How to join

The talk is held online via Jitsi. You can join with the link https://jitsi.uni-kl.de/SciCompSeminar_05. Please follow the rules below:

• Use a chrome based browser (One member with a different browser can crash the whole meeting).
• Mute your microphone and disable your camera.
• If you have a question, raise your hand.

More information is available at https://www.rhrk.uni-kl.de/dienstleistungen/netz-telefonie/konferenzdienste/jitsi/.

Payam Dehpanah, Scientific Computing Group, TU Kaiserslautern

Title:  

Integrating Airfoil Optimal Design within the Blended-Wing-Body Conceptual Design Framework

Abstract:

The blended-wing-body configuration is introduced as the future airliner in recent decades. A quick computational airfoil design can assess design individuals in the design space. This study determines the possibility of integrating an optimal airfoil design within the blended-wing-body conceptual aircraft design framework. The continuous adjoint method for the Euler equations is developed within the conceptual design framework. The framework evaluates the configuration before and after the airfoil optimal design process. The airfoil aerodynamic design is studied for one symmetrical and two S-shaped airfoils with adequate thickness-to-chord ratio for the centerbody profile. Three sequential objective functions including lift, drag, and pitching moment are subjected to the design constraints. Projected aerodynamic performances of the optimized airfoils, including lift-to-drag ratio and static margin are acquired appropriately. In addition, off-design performance, including the RANS solutions, are studied. The result of the blended-wing-body conceptual design framework with optimized airfoils are evaluated.

How to join

The talk is held online via Jitsi. You can join with the link https://jitsi.uni-kl.de/SciCompSeminar_06. Please follow the rules below:

• Use a chrome based browser (One member with a different browser can crash the whole meeting).
• Mute your microphone and disable your camera.
• If you have a question, raise your hand.

More information is available at https://www.rhrk.uni-kl.de/dienstleistungen/netz-telefonie/konferenzdienste/jitsi/.

PD Dr. Stephan Schmidt, Department of Mathematics, HU Berlin,
Melanie Gräßer, Chair for Particle Technology, University of Paderborn

Title:
Computation of Capillary Bridges via Shape Optimization Techniques

Abstract:

The interaction of very limited amounts of fluids with solid bodies differs considerably from the fully immersed case. Instead of dealing with an uninterrupted continuum of fluid, one faces a problem, where the properties are determined by the behavior of droplets of fluids, in particular so called capillary brides, i.e., droplets of fluids that form between solid bodies and mutually connect them. The resulting capillary forces govern the macroscopic behavior of e.g. moist granulates. Sandcastles provide an intuitive access to these phenomena: It is neither possible to build a sandcastle under water, nor when using completely dry sand.

The presentation is split into two parts. The first half will present the challenges when using classical CFD simulation tools as most simulation approaches for free surface flows do not capture the various interfaces between liquids, air and solids to a satisfactory degree. The second half of the presentation focuses on tackling the problem from a purely geometric perspective. Mathematically, the formation of a capillary bridge can be considered a generalized plateau-problem, i.e. constructing a surface of zero mean curvature, which, in this case, also follows a variety of thickness, interface and contact angle constraints. We present a numerical scheme based on shape calculus and finite elements on surfaces to address the problem numerically for arbitrarily shaped liquid-solid contact surfaces. Thickness and contact constraints are dealt with in a hybrid level-set and resolved mesh approach.

Extension to quadratically convergent shape-Newton methods to this problem are straight forward but subject to future research.

How to join

The talk is held online via Jitsi. You can join with the link https://jitsi.uni-kl.de/SciCompSeminar_07. Please follow the rules below:

• Use a chrome based browser (One member with a different browser can crash the whole meeting).
• Mute your microphone and disable your camera.
• If you have a question, raise your hand.

More information is available at https://www.rhrk.uni-kl.de/dienstleistungen/netz-telefonie/konferenzdienste/jitsi/.

Payam Dehpanah, Scientific Computing Group, TU Kaiserslautern

Title:  

Introduction to an Entrepreneurial Solution: HyParter

Abstract:

Every year new technologies are introduced for design, development, and operation of subsonic, transonic, supersonic and hypersonic air transport as aeronautics enters a new era in which an energy efficient clean sustainable air transport is substantially important. Electric and autonomous urban mobility emerges as a very practical solution to shape the future. Private companies have already started their own VTOL projects. Meanwhile, computation outstandingly helps all these new technologies thrive to shape a better future along with experiment, but still the long lasting gap between them exists. Extensive endeavors have been being made ever since by many researchers and developers to empower and improve their computational system to be able to address new technologies. Digitization and virtualization become important and the role of AI and ML in this transition known as re-invention of aviation remains obscure, hard to perceive and challenging. World-class scientific leaders across the globe support development of technologies to enhance re-invention of aviation with scalable and profitable ideas. HyParter stands for hyper partition and is an entrepreneurial solution with an open system that offers insightful solutions for different technologies at a systematic level.

In the second part of the talk meshing techniques with Pointwise for several cases will be demonstrated including:

1. Two dimensional normal extrusion for supercritical and multi-element airfoils.
2. Structured grid for 3D Backward Facing Step (3DBFS).
3. Structured multi-block grid for wall-mounted cylinder.
4. Hybrid and Voxel grid for wall-mounted cylinder and unstructured solver.
5. Introduction to Glyph Scripting.

How to join

The talk is held online via Jitsi. You can join with the link https://jitsi.uni-kl.de/SciCompSeminar_08. Please follow the rules below:

• Use a chrome based browser (One member with a different browser can crash the whole meeting).
• Mute your microphone and disable your camera.
• If you have a question, raise your hand.

More information is available at https://www.rhrk.uni-kl.de/dienstleistungen/netz-telefonie/konferenzdienste/jitsi/.