This thesis presents new approaches for modeling and optimization for stationary base engine calibration. At first, the most suitable type of modeling for this topic is determined in an extensive comparison. Afterwards, this technique is extended by new features, e.g. outlier-robustness, in order to achieve a dependable performance of the model even under complex and difficult circumstances. Subsequently, a new model-based multi-objective online optimization is presented, which is able to automatically identify the Pareto-optimal areas of the combustion engine with a permanent online-connection between the optimization algorithms and the test bench. Various theoretical examples and practical applications demonstrate the performance of these new approaches.
«
This thesis presents new approaches for modeling and optimization for stationary base engine calibration. At first, the most suitable type of modeling for this topic is determined in an extensive comparison. Afterwards, this technique is extended by new features, e.g. outlier-robustness, in order to achieve a dependable performance of the model even under complex and difficult circumstances. Subsequently, a new model-based multi-objective online optimization is presented, which is able to automa...
»