The integration of safety-critical and major power-consuming electrical systems presents a challenge for the development of future automotive electrical networks. Both reliability and performance must be enhanced in order to guarantee the power supply to essential electrical consumers at a sufficient degree of power quality. Often, in order to cope with these requirements, merely an upgrade of the existing wiring harness design is used, resulting in additional complexity, weight, and cost [3].
A characterization of the wiring harness and its electrical consumers facilitates a systematic optimization approach aimed at designing new automotive power networks [1, 5]. Measurement and analysis methods to characterise the thermal behaviour of the wiring harness have been presented and discussed in a previous paper [4]
This paper presents and compares two methods aimed at modeling the electrical behavior of consumers at various voltages and temperatures. The first, conventional method relies on physics-based, analytical models. The second method is based on a generic representation of non-linear dynamic behavior using a common model frame for various electrical consumers. A dedicated parameterization strategy for the data-driven models is developed. The two modeling approaches are compared in terms of representation, flexibility and generalization capabilities. The models obtained are integrated into a modular and scalable HiL test platform used to characterize and test automotive electrical power nets in an early phase of the development process.
The methodology and tools used, as well as the results obtained, are presented and discussed in this paper.