Modellbasierte Entwicklung und Applikation von Diagnosefunktionen im Kühlkreislauf des Kraftfahrzeugs

• Model-based development and calibration of diagnosis functions in the cooling circuit of a vehicle

Hoppe, Matthias; Abel, Dirk (Thesis advisor); Brabetz, Ludwig (Thesis advisor)

Aachen (2017)
Dissertation / PhD Thesis

Dissertation, RWTH Aachen University, 2017

Abstract

The focus of this work lies on model-based methods for the development and the calibration of diagnosis functions. Using model-based calibration methods in the development of diagnosis functions reduces the required time significantly. A model-based diagnosis function allows an isolation of errors, which is not possible with diagnosis functions currently in use. The temperature diagnosis in the cooling system of a vehicle is used as an example application. In a first step, a simulation model of the cooling system is developed. For this model, a balance between the accuracy of the model and the calculation time has to be found. For the engine, a static combustion model is combined with a dynamic model of the heat transfer. For the elements in the outer cooling circuit, a static model using maps is compared with a physical, dynamic model. For the calibration of the maps used in the diagnosis, two methods are described. On the one hand, parametric methods like a least squares algorithm can be used. On the other hand, a direct optimization with the help of evolutionary algorithms is presented. Both algorithms are further extended, in order to satisfy additional requirements, like a smooth surface of the maps. For the development of a model-based diagnosis function, several sensor configurations are investigated in regard to the isolation of errors with the help of a structural analysis. A diagnosis function is implemented for a suitable sensor configuration. The functionability is tested in simulation.