Advertised ADPs / ARPs, Bachelor and. Master theses

As part of the KI4ETA project, for example, load peak smoothing on hybrid storage should be designed intelligently. For this purpose, strategies such as deep reinforcement learning (DRL) and mathematical optimization (MILP) have already been implemented in the eta-utility package. However, conventional controllers such as PID also have advantages and should therefore be used for comparison.

As part of the KI4ETA project, for example, load peak smoothing on hybrid storage should be designed intelligently.For this purpose, strategies such as deep reinforcement learning (DRL) and mathematical optimization (MILP) have already been implemented in the eta-utility package. However, in this simulative setup, only the data actually preceding the test data is currently used as a forecast, which would correspond to an ideal forecast and therefore cannot work in real applications

As part of the KI4ETA project, for example, load peak smoothing on hybrid storage should be designed intelligently. For this purpose, strategies such as Deep Reinforcement Learning (DRL) using PPO have already been implemented in the eta-utility package. As a comparison, a genetic algorithm should now be trained and compared with the results of the PPO algorithm.

At the Institute for Mechatronic Systems in Mechanical Engineering, research is being conducted into energy-efficient drive concepts for electric robotic vehicles. In order to be able to better assess potentials for future drive and traffic concepts, these must be tested simulatively. In order to be able to assess the influence of a traffic environment, a traffic simulation of the city center of Darmstadt is therefore used at the IMS. This is implemented in the SUMO simulation software. In order to represent current traffic behavior here, the existing simulation must be revalidated. Current traffic data is available for this purpose.

Die Elektrifizierung der Mobilität in der Industrie rückt Lithium-Ionen-Batterien als mobile Speichertechnologie immer weiter in den Fokus der Forschung und Anwendung. Da die Kosten der Batterie einen relevanten Anteil an den Gesamtkosten der Fahrzeuge ausmachen, ist das Wissen über den aktuellen Alterungszustand der Batterie (SoH) und die zukünftig zu erwartende Restlebensdauer für Investitions- und Betriebsentscheidungen äußerst relevant. Datengetriebene Ansätze zeigen hohes Potenzial, den SoH von Li-Ion-Batterien verlässlich zu bestimmen. Die Arbeit wird in Zusammenarbeit mit einem führenden Industriepartner im Bereich elektrischer Flurförderfahrzeuge durchgeführt.

Human-robot collaboration (HRC) is a way to improve flexibility and ergonomics in the workplace. To improve the collaborative process’s efficiency, each partner should be aware of the other partner’s actions and intentions at all times.

Am Institut für Mechatronische Systeme im Maschinenbau wird an Energieeffizienten Antriebskonzepten für elektrisches Roboterfahrzeug geforscht. Um Potentiale für zukünftige Antriebs- und Verkehrskonzepte besser abschätzen zu können, müssen diese Simulativ getestet werden. Um dabei den Einfluss eines Verkehrsumfeldes beurteilen zu können, wird deshalb am IMS eine Verkehrssimulation des Stadtzentrums von Darmstadt verwendet. Dieses ist in der Simulationssoftware SUMO umgesetzt. Um hier aktuelles Verkehrsverhalten darzustellen, muss eine Neuvalidierung der bestehenden Simulation vorgenommen werden. Dafür stehen aktuelle Verkehrsdaten zur Verfügung.

Für das Training von Machine-Learning-Modellen bspw. zur Geschwindigkeitsprädiktion wird eine große Menge an Daten benötigt. In vielen Anwendungsfällen ist das Aufzeichnen von Daten jedoch mit hohem Aufwand verbunden. Im Rahmen dieser Arbeit sollen realistische künstliche Fahrten mithilfe eines Ansatzes aus dem Bereich der künstlichen Intelligenzen basierend auf einem bestehenden Datensatz von Realfahrten erzeugt werden.

At depots with several electric buses, a large aggregated capacity for storing electrical energy is available at night. Through bidirectional charging, this storage capacity can be used beyond mobility purposes. By connecting to the electricity grid, renewable energy generation fluctuations can be compensated. Further, the electricity grid can be stabilized through control energy. The aim of this work is a comprehensive analysis of the energy flexibility potential of bus depots.

At depots with several electric buses, a large aggregated capacity for storing electrical energy is available at night. Through bidirectional charging, this storage capacity can be used beyond mobility purposes. By connecting to the electricity grid, renewable energy generation fluctuations can be compensated. Further, the electricity grid can be stabilized through control energy. The aim of this work is a comprehensive analysis of the energy flexibility potential of bus depots.

At depots with several electric buses, a large aggregated capacity for storing electrical energy is available at night. Through bidirectional charging, this storage capacity can be used beyond mobility purposes. By connecting to the electricity grid, renewable energy generation fluctuations can be compensated. Further, the electricity grid can be stabilized through control energy. The aim of this work is a comprehensive analysis of the energy flexibility potential of bus depots.

Modern flywheel storages use electric machines to convert electrical into mechanical energy and vice versa. Due to power losses caused by intensive load profiles and high speeds, flywheel storages are often operated at high temperatures. In order to estimate the lifetime of the entire system and design various components, knowledge of the system thermal behavior is required. Thermal models are then aligned with experimental results and used to estimate the thermal behavior of the system.

Modern flywheel storages use electric machines to convert electrical into mechanical energy and vice versa. Due to power losses caused by intensive load profiles and high speeds, flywheel storages are often operated at high temperatures. In order to estimate the lifetime of the entire system and design various components, knowledge of the system thermal behavior is required. Thermal models are then aligned with experimental results and used to estimate the thermal behavior of the system.

Schwungmassenspeicher dienen zur Speicherung elektrischer Energie. Zur Reduktion der Reibung wird der Rotor in der Regel magnetgelagert. Im Fall einer Fehlfunktion soll eine mechanische Lagerung (Fanglagerung) ein sicheres Auslaufen des Rotors gewährleisten. Dabei wird die Fanglagerung thermisch stark beansprucht.

In der Arbeit soll ein Modell entwickelt werden, mit dem aufbauend auf den Kräften während eines Absturzes die Temperaturen im Lager berechnet werden können.