This presentation was made at the NAFEMS European Conference on Simulation-Based Optimisation held on the 15th of October in London.
Optimisation has become a key ingredient in many engineering disciplines and has experienced rapid growth in recent years due to innovations in optimisation algorithms and techniques, coupled with developments in computer hardware and software capabilities. The growing popularity of optimisation in engineering applications is driven by ever-increasing competition pressure, where optimised products and processes can offer improved performance and cost-effectiveness which would not be possible using traditional design approaches. However, there are still many hurdles to be overcome before optimisation is used routinely for engineering applications.
The NAFEMS European Conference on Simulation-Based Optimisation brings together practitioners and academics from all relevant disciplines to share their knowledge and experience, and discuss problems and challenges, in order to facilitate further improvements in optimisation techniques.
Modern design approaches include optimized designs early in the development process. The goal is to minimize the number of design iterations necessary to reach a final product , therefore saving time and cost. Often times methods like topology or shape optimization are used to improve the design basis. Since many of these optimization approaches discretize either the design space or the component, interpreting and creating functional CAD models from such FE based results can be difficult and has been a focus of research.
Meanwhile, engineering software developers such as Siemens NX, Dassault, etc. have started to integrate CAD and CAE capabilities into a single software suite. Designing and simulating in such a software package easily allows for changes in the geometry, while keeping the boundary conditions associatively connected. In combination with the software’s API, this allows for the creation of complex workflows, where CAD and FE models are automatically altered and solved according to given design parameters. Since optimization algorithms require many target evaluations, such an automated workflow is required to efficiently solve an optimization problem. Moreover, the CAD model has to be properly parameterized for optimization purposes.
The parametrization of the CAD model forms its design space for the optimization and has a significant impact on the results. A lot of research has been done on how to model and parametrize structures effectively for optimization and how to choose optimization parameters efficiently. The goal is to use few parameters, while allowing a wide range of different designs. The use of B-splines is a common approach in many different fields to describe and optimize complex shapes with few parameters.
Structural optimization applications that consider flexibility, hitting specific stiffness targets in different directions and decoupling connected components can lead to unwanted results as described in . The goal of decoupling different parts of an assembly by flexible components
with specific stiffness requirements is to not transfer unwanted disturbances. An alternative solution to this problem, using the introduced automated workflow, is presented.
To illustrate this method a connection ring to a roller bearing is used as an example (Figure 1). The objective of the optimization is to minimize the misalignment of the upper and lower bearing rings under load, while minimizing weight, complying to maximum stress constraints and maintaining a minimal radial flexibility. Therefore, an overall stiffening of the structure as in a typical minimize compliance problem is not possible. The difficulty lies in the contradicting objectives to minimize the tilt of the bearing and keeping the connection arm radially flexible to decouple the connected components on each side.
B-splines were used to model the overall shape of the connection ring and a permutating B-spline to efficiently vary the rings thickness radially. Different CAD parametrization strategies are compared regarding performance and efficiency. Different design variations have been investigated, varying stress and compliance constraints and altered design requirements such as connection points to neighbouring components, Topological changes to the structure would have been possible as described in , but were not performed in this example.
The goal is to illustrate the capability of a fully automated CAD geometry based approach, giving design engineers another tool to improve initial designs in a well-known software environment. Significant improvements regarding weight and performance have been achieved and the results are directly accessible in a fully functional CAD model.
|Date||15th October 2019|