With increasing levels of driving automation, the responsibility for the vehicle control and passengers’ safety is shifted from the driver towards the automation system. This results in increased reliability and safety requirements for the subsystems involved in the automated vehicle motion. In case of a safety-critical failure, an automated transition to a vehicle standstill must be executed without any driver interaction and supervision for high and full driving automation. Therefore, in addition to the functionality of the safety-critical subsystems, also their reliable power supply for the duration of the transition to the standstill must be guaranteed by the vehicle powernet.
According to the functional safety norm ISO 26262, all safety-critical subsystems must be designed with an appropriate automotive safety integrity level (ASIL). Since reliable power supply is one of the prerequisites for their correct functionality, it becomes safety-critical itself. The need for new fail-operational powernet topologies and appropriate control strategies fulfilling the increased reliability and availability requirements arises. The work presented in this dissertation proposes a new generic energy management system for the safety-based range extension, supporting the optimization of the powernet and powertrain control for arrival at the safest possible location for the passengers.
The key element of the proposed energy management system is the online energy distribution optimization with an integrated degradation concept. Using the predicted values for the available energy resources and for the energy required to complete a driving mission, the control strategy automatically adapts the energy flows within the vehicle powernet. Furthermore, it estimates appropriate degradation step for the comfort loads, driving profile and driving destination with the goal to reach the safest destination with a maximum of comfort in a minimum of time under consideration of the available energy resources. With this approach, also the fault reactions of the functional safety concept, required by ISO 26262 for all safety-critical functions, are automated and optimized. In this way, the energy management system finds autonomously the best suited fault reactions for achieving the defined control goal.
The energy demand required for the completion of a driving mission depends on the velocity profile, road slope and stops on the way to the destination. Using the electronic horizon, a driving trajectory from the current vehicle position to the destination is approximated online, which is then used for the route based estimation of the required propulsion energy. In powertrain topologies with multiple traction motors, the overall driving efficiency and hereby also the driving range can be increased significantly with an appropriate strategy for the torque distribution between individual motors. Therefore, a torque distribution profile is estimated online based on the theory of optimal control for the entire driving mission, enabling an accurate and realistic prediction of the propulsion energy required for the safety based range extension. In addition to the driving efficiency increase also symmetrical discharge of independent traction batteries, required for approximately the same driving range in case of a breakdown of one battery, is incorporated in the control strategy. Furthermore, also the balancing of the energy losses in powertrain components is considered for the torque distribution, which is required to avoid their overheating possibly leading to faster aging and wear.
The application of the generic framework for the energy management system is exemplified on two different powernet and powertrain topologies with a single and multiple traction motors. The benefits are verified using simulation results for both, fault-free and failure case operation. With the proposed torque distribution strategy a decrease in the energy losses of up to 12 % for the given use case was achieved. Also the optimization of the fault reactions shows the ability of the energy management system to achieve the control goals despite multiple faults. In addition to the concepts presented in this dissertation, also runtime optimized algorithms are proposed, implemented and validated by means of simulation. By enhancing the reliability of the power supply and fail-operability of the powertrain, the work presented in this dissertation contributes to the establishment of the evolving automated driving technology and provides a generic framework for model predictive energy management for future implementation in automated vehicles.