Power Hardware in the Loop (PHiL) Systems
Vehicles and especially their powertrains become more complex and interconnected, whereas time development cycles are getting shorter. To increase safety, fuel efficiency or drivability, the number of highly interconnected components and complex control units is on the rise. To ensure quality and cost efficiency, PHiL test beds are utilised in early development stages. They are used to develop, calibrate and test the interaction of various subsystems, both real and virtual, with most realistic and reproducible test patterns.
In a PHiL system, a real device under test is connected to dynamic simulation models. Interaction between the real system and the virtual simulation environment (VSE) is realised by the PHiL interface. One major challenge, which arises in PHiL systems, is that operations must be as close to reality or as accurate as possible in order to be reliable when making further decisions in the development process.
Research Topics at our Institute
In a PHiL system, unwanted oscillations or even instability may occur. These effects are caused by the interaction of the VSE with the real test bed (RTB) by some imperfections. These can be imperfect measurements, noisy signals or numerically stiff models. Therefore, the interfacer represents a crucial part of stability and accuracy of a PHiL system. To reduce the adverse effects, our research group developed an interface methodology for a PHiL powertrain testbed. The methodology is used to predict and optimise the behaviour of the VSE by modifying the measured signal from the test bed before it is sent to the VSE.
Model Based Interface Methodology
In a powertrain testbed, real halfshaft torques are measured and used to drive virtual wheels and their tyres. Therefore, a good choice of an interface model that predicts oscillations would be the wheel including its tyre and road interaction. Oscillations in the wheel‘s rotational speed can be controlled by effectively reducing its second time derivative or jerk. Therefore, in the interface methodology, a corrective torque which reduces the jerk is introduced and added to the measured torque. This optimised torque is then fed into the VSE. The determination of the corrective torque is based on a predictive receding horizon optimisation. The penalty function minimises the jerk while both angular momentum as well as mechanical work introduced by the corrective torque are guaranteed to be zero. The physical significance of the PHiL test runs is ascertained, because the validity of these constraints is secured by the proposed method.
Research Projects at our Institute
Zauner, Michael, Philipp Mandl, Christoph Hametner, Oliver König, and Stefan Jakubek. "Flatness-Based Discrete-Time Control of a Battery Emulator Driving a Constant Power Load, opens an external URL in a new window" IEEE Journal of Emerging and Selected Topics in Power Electronics 9, no. 6 (2021): 6864-6874.
Aschauer, G., A. Schirrer, M. Kozek, and S. Jakubek. "PHiL pantograph testing via FE-based catenary model with absorbing boundaries, opens an external URL in a new window" Control Engineering Practice 88 (2019): 97-109.
Schirrer, A., G. Aschauer, E. Talic, M. Kozek, and S. Jakubek. "Catenary emulation for hardware-in-the-loop pantograph testing with a model predictive energy-conserving control algorithm, opens an external URL in a new window" Mechatronics 41 (2017): 17-28.
Jakubek, Stefan, Elisabeth Luchini, Alexander Oberhummer, and Felix Pfister. "A model-based interfacing concept for accurate power hardware-in-the-loop systems, opens an external URL in a new window" Mathematical and Computer Modelling of Dynamical Systems 22, no. 1 (2016): 1-20.
König, Oliver, Christoph Hametner, Günter Prochart, and Stefan Jakubek. "Battery emulation for power-HIL using local model networks and robust impedance control, opens an external URL in a new window" IEEE Transactions on Industrial Electronics 61, no. 2 (2013): 943-955.