Universität Stuttgart
Permanent URI for this communityhttps://elib.uni-stuttgart.de/handle/11682/1
Browse
Item Open Access Energy-based modeling and decentralized observers for adaptive structures(Düren : Shaker Verlag, 2021) Warsewa, Alexander; Tarín, Cristina (Prof. Dr.-Ing.)Item Open Access Lifetime modeling and model-based lifetime optimization of Li-ion batteries for use in electric two-wheelers(Düren : Shaker Verlag, 2020) Rechkemmer, Sabrina; Sawodny, Oliver (Prof. Dr.-Ing. Dr. h.c.)This thesis deals with the lifetime prediction of Li-ion batteries and the applicability of such models to optimize the operation and charging strategy. A model-based optimal control is therefore designed. The focus is on typical application scenarios for electric two-wheelers (E2W) in Shanghai. A realistic driving behavior is first derived to reflect the typical use of an E2W. A data acquisition campaign is conducted to capture the local conditions in Shanghai. Based on survey and fleet data, a typical driving cycle for E2W is designed in Shanghai. Models of the electric motor and power control in combination with vehicle-to-street characteristics are used to convert the recorded typical driving cycles to equivalent power profiles. The main focus is to investigate and model a Li-ion battery from a system perspective. Different model representations are compared regarding their applicability in simulation and optimization. An aging campaign is initiated to investigate the characteristics of a lithium manganese oxide (LMO) cell. Results for calendar and cycle aging are presented and compared to further cell chemistries. A single particle model as well as a simplified electrical equivalent circuit model are introduced and adapted to the cell results. The driving cycle and system model are incorporated into a model-based optimization strategy. Since the greatest optimization potential was identified during charging and storage, a combined optimal charging and operating strategy is described. The validity of the results is proven by further cell tests and compared with simulation results, which show that an extension of the lifetime is achieved. A heuristic operating strategy is then derived, which can also be implemented online on the E2W control unit.Item Open Access Methoden und Implementierung einer vorausschauenden Fahrwerksregelung für aktive und semi-aktive Federungssysteme(Shaker Verlag, 2014) Göhrle, Christoph; Sawodny, Oliver (Prof. Dr.-Ing. Dr. h.c.)Item Open Access Model predictive energy management for induction motor drives and all-wheel-drive battery electric vehicles : a flatness based approach(Düren : Shaker Verlag, 2021) Rolle, Bernhard; Sawodny, Oliver (Prof. Dr.-Ing. Dr. h.c.)Since emission and fuel economy standards have risen significantly, car manufacturers find themselves forced to invest in new battery and related technologies. Research on methods and technologies that improve the efficiency of both the battery and the electric powertrain, pose the greatest challenges of this technological transition. Therefore, software solutions for energy management and motion control as well as economic driving strategies are becoming more and more the focus of future developments. Software solutions allowing for potential economic savings are particularly appealing, since these do not require any structural or mechanical design changes. With the main objective of increasing economic savings, this present work investigates analytical models of the electric powertrain for a battery electric vehicle with two drive modules on the front and rear axle. The modeling approach focuses on loss processes associated with the energy conversion of the voltage source inverters and induction motors. Due to a wide spectrum of involved time constants in the range of seconds to a few milliseconds, efficiency analyses of electric vehicles rarely follow model-based approaches and instead rely on characteristic loss maps, which neglect dynamic effects and physical limitations, for example, those resulting from the limited battery voltage. This widespread spectrum comes from both the longitudinal dynamics of the vehicle and the voltage and current harmonics that result from high frequency switchings of the inverters’ semi-conductors. New dynamical models are thus proposed that can be efficiently integrated into vehicle simulations and also be implemented online on embedded systems, such as the motor control unit of the investigated vehicle. In doing so, an average value model of the voltage source inverter is derived, based on a double Fourier integral analysis of the semi-conductor switching signals. Furthermore, a widely used model of the induction motor, applied for motor analysis and control design, is reformulated into an equivalent differential flat system based on the definition of a new flat output. Both component models are integrated into a vehicle simulation of a Mercedes Benz EQC prototype and are thoroughly validated through extensive simulative studies and experimental test series. With the help of the newly introduced models and with the assistance of modern vehicle sensor systems, control strategies of the electric powertrain are investigated that aim for the most energy efficient operation. In a first step, decentralized optimal control approaches are proposed that improve the efficiency of the electric drive module, not only during stationary operation, but also during transient torque conditions. This improvement is achieved by an appropriate field oriented control method. In a second step, optimization-based torque allocation strategies are investigated and evaluated experimentally. Finally, a centralized predictive control approach is presented that exploits all operational degrees of freedom, which are the variable torque allocation, the front and rear axle magnetic flux, and the adjustment of the vehicle speed according to topographical and traffic dependent conditions. Significant economic savings are demonstrated for the decentralized control methods as well as for the centralized control approach. The high level of accuracy and performance that is achieved by the proposed model-based framework and predictive operational strategies are only made possible by exploiting the positive structural properties of the newly introduced differential flat system representation of the induction motor.Item Open Access Modellierung und modellbasierte Regelung von Kontinuum-Manipulatoren(Aachen : Shaker Verlag, 2017) Falkenhahn, Valentin; Sawodny, Oliver (Prof. Dr.-Ing. Dr. h.c.)Kontinuum-Manipulatoren gehören zu einer neuartigen Roboterklasse und erschließen durch ihre gelenkfreien, kontinuierlichen Verformungen neue Anwendungsfelder. Für die praktische Anwendbarkeit von Kontinuum-Manipulatoren sind schnellere und präzisere Bewegungen erforderlich. Daher wird eine modellbasierte Folgeregelung der Aktorkoordinaten entwickelt, die ein gutes Führungsverhalten und verbessertes Störverhalten bei dynamischer Anregung ermöglicht. Die Vorgabe der Führungsgrößen erfordert Methoden der Trajektoriengenerierung. Die Entwicklung dieser Methoden und deren experimentelle Umsetzung erfolgt anhand eines Beispielmanipulators, des pneumatisch aktuierten Bionischen Handling-Assistenten (BHA). Im Rahmen der dynamischen Modellierung wird die verteilte Manipulatormasse über konzentrierte Sektionsmassen approximiert. Die kinematische Beschreibung des Manipulators erfolgt über den Ansatz konstanter Sektionskrümmungen. Darauf aufbauend wird das dynamische Modell über ein Ersatzmodell mit konzentrierten Massen hergeleitet, sodass die Bewegungsgleichungen der Aktoren in klassischer Form vorliegen. Für den Reglerentwurf werden modellbasierte Zustandslinearisierungen mit linearen PID-Reglern und Vorsteuerungen kombiniert. Der entwickelte modellbasierte Regler mit einem signifikant verbesserten Führungs- und Störverhalten erlaubt schnelle und präzise Bewegungen generierter Aktortrajektorien aus Vorgaben von Aufgabenkoordinaten. Für die Trajektoriengenerierung werden beispielhaft zeitoptimale Punkt-zu-Punkt-Trajektorien und onlinefähige TCP-Geschwindigkeitsvorgaben vorgestellt. Darüber hinaus kann das entwickelte dynamische Modell für erweiterte Reglerkonzepte hinsichtlich TCP-Regelungen oder Kraft- bzw. Impedanzregelungen verwendet werden.