Quickly changing trends in market demand of mechatronic systems require engineering companies to reduce production costs and time-to-market and to increase the production
flexibility and precision of product design. Virtual prototyping aims at fulfilling these requirements, by utilizing virtual prototypes of mechatronic systems, in order to test system design and performance in any phase of the engineering development process. The main objective of this thesis is to develop a generic evaluation framework, that allows product designers and engineering domain experts to efficiently and meaningfully evaluate multiple criteria of virtual prototypes for a variety of model parameters in simulation. Furthermore, two exemplary applications involving the Franka Emika Panda
robot arm are implemented, in order to demonstrate the capabilities of the developed framework. In the first scenario, different trajectories for robotic pick and place use cases are evaluated for their energy efficiency. The second scenario models a bin picking process of the Panda robot in an industrial plant and aims at identifying valid and objective-specific optimal robot positions inside the plant.
The evaluation framework was designed and implemented by following a top-down approach. Evaluating a mechatronic system was separated into three distinct steps and correspondingly the evaluation framework comprises three main components: (1) Defining the evaluation parameters and objective values of the evaluation from an existing simulation model, i.e. choosing the design alternatives to be tested and the corresponding decision criteria to be evaluated, is implemented in Unity in the C# programming language. (2) The evaluation framework’s second component is implemented in C# as a Microsoft .NET Core solution and uses .NET parallelization methods to efficiently evaluate the executable of the Unity simulation model for all combinations of evaluation parameters. (3) Using Python’s tkinter library, the third component visualizes the resulting point clouds of all decision criteria for all design alternatives in an interactive graphical user interface (GUI).
The evaluation results of the pick and place application show that the energy-optimal robot trajectories in the simulated scenario correspond to those, where the robot moves as few joints as possible, showing that the amount of distance
covered in joint space constitutes the major fraction of electric energy consumption on a trajectory.
Evaluating bin picking with the Panda robot in an industrial plant reveals, there is only very little scope available for positioning the robot such that it is able to successfully perform bin picking. This lies in the fact that the Panda robot’s range of motion is relatively small compared to its workspace in the industrial plant. Furthermore, the results show, that robot positions resulting in trajectories with low execution time concurrently result in short joint distance.
In both exemplary scenarios, the complete evaluation framework was applied successfully. The applications showcase the capabilities of the evaluation framework and
demonstrate its strength in performing basic, extensive evaluation of virtual prototypes.
For future work, the generic nature of the evaluation framework makes it possible to apply it in many different application areas and easily extend it to combine multiple
simulation software and integrate multi-objective optimization algorithms.
«
Quickly changing trends in market demand of mechatronic systems require engineering companies to reduce production costs and time-to-market and to increase the production
flexibility and precision of product design. Virtual prototyping aims at fulfilling these requirements, by utilizing virtual prototypes of mechatronic systems, in order to test system design and performance in any phase of the engineering development process. The main objective of this thesis is to develop a generic evaluation...
»
Login
BibTeX