Fault tolerant mechatronic actuator systems with realisation of an electrical duplex power steering

• Applicant: Professor Dr.-Ing. Rolf Isermann
• Subject Area: Automation, Mechatronics, Control Systems, Intelligent Technical Systems, Robotics
• Term: from 2012 to 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 213635183

The general interest on duplex actuators for safety critical systems is for instance in the automotive sector increased since initial request. The tasks of the initial request and the results of our research are basically important for fault-tolerant actuator systems. The not clarified tasks and the new developed questions are described below.The accuracy of the developed actuator models should be further increased. Therefore the motor model should be extended by a semi-physical modeling. With this approach nonlinear effects like hysteresis could be handled better. This is particularly important for fault diagnosis.The residuals, which are the result of a structure analysis of the models, must be validated at the existing test benches. Depending on the validation results, an extended fault modeling will be done.An approach for handling actuator faults was developed. The robustness of the used coupling control should be improved by a semi-physical modeling. The coupling control will be extended to a gain-scheduling controller.Especially at a total breakdown of one actuator, the choice of the right switching strategy is important. The impact of a hard or soft switching highly depends on the situation. The existing duplex-EPS test bench should be extended by actuators for generating steering force reactions. This allows an active load simulation and the test bench can operate in closed loop. Driving situational switching strategies can be researched. An integration of a vehicle simulation at the real-time target is intended.In order to get a good adaptability of the chosen controller at a fault the internal model control (IMC) was analyzed. Because of the internal model in the controller the control structure can be adapted easily by parameter adaption. A speed and a positioning control loop should be implemented with the IMC structure. Additionally several parameter estimation techniques for the adaption of the IMC should be investigated at the test bench.Faults of sensors can be easily followed by wrong actuating values. To minimize the impact of faulty current and rotor position sensors virtual measurement signals should be generated.Finally, the integration of all methods in a fault management system should be analyzed and synergies of the methods should be used in order to reduce the computing demand.

• DFG Programme: Research Grants