Work description

The concept of the digital twin in product development enables new application scenarios in testing. In this process, a simulation is coupled with a test bench and a bidirectional data exchange, known as twinning, is established. This continuous validation creates reliable simulation models. These can now be used to optimize products.

In the future, electrohydrostatic actuators (EHAs) will be increasingly used in aviation. The individual components are subject to strong interactions. For example, changing a parameter such as the cross-sectional area of the hydraulic cylinder has a very strong effect on the system behavior of the EHA's. Many of these parameters can be changed as design targets for the next generation of products. Due to the large number, these interactions cannot be fully understood by the product developer. Optimizing the system is challenging for this reason.

Task:

The goal of this work is the AI-based optimization of an EHA by using a digital twin. Through a sensitivity analysis of an existing simulation of an EHA, the relevant design target variables are to be determined first. These are to be used as parameters or variables of the objective function for the optimization. The boundary conditions from aviation are to be taken into account. Depending on the optimization problem, suitable methods such as Particle Swarm Optimization or Multi-Objective Genetic Algorithm shall be used. The result shall be recommendations for the development of the next product generation.

Profile:

• Field of study: mechanical engineering, mechatronics, industrial engineering or related subjects.
• Interest and first experience with MATLAB
• Knowledge in optimization methods (AI methods)