Simulation of Automotive EMC Emission Test Procedures Based on Cable Bundle Measurements M. Gonser * , C. Keller * , J. Hansen * , V. Khillkevich † , A. Radchenko † , D. Pommerenke † , R. Weigel ‡ * Robert Bosch GmbH Stuttgart, Germany Email: markus.gonser@de.bosch.com, christoph.keller@de.bosch.com, jan.hansen@de.bosch.com † Missouri University of Science & Technology Rolla, Missouri, USA Email: khilkevichv@mst.edu, ar8r3@mst.edu, davidjp@mst.edu ‡ Friedrich-Alexander University Erlangen-Nuremberg Erlangen, Germany Email: r.weigel@ieee.org Abstract—Two time efficient simulation methods for the pre- diction of the conducted and radiated automotive EMC emission tests, respectively, are presented. To verify the correct prediction of the cable bundle model, a novel cable bundle test bench has been developed. It allows a fully automated network analysis with up to 32 ports. Several types of cable bundles are experimentally characterized under stochastic considerations. The system simu- lation for the conducted emission setup is partitioned into four parts connected on circuit level. The statistically analyzed voltage at the output of the line impedance stabilization network is in very good agreement to respective measurements. To simulate the radiated emission a novel method is presented, where the setup is partitioned by the use of Huygens principle. The simulated prediction of the antenna voltage is in good agreement with respective measurements. The presented approaches are very time efficient and therefore can be used effectively during product development. Index Terms—Electromagnetic compatibility, Simulation, Mea- surement, Measurement Techniques, Numerical analysis, Numer- ical simulation, Electromagnetic coupling I. I NTRODUCTION The rising complexity of automotive electronics along with shorter development periods challenges many design aspects of automotive products. One major issue is the electromagnetic compatibility (EMC) where simulation of associated measure- ment setups even at early design stages has become desireable. One challenge is due to the flexible nature of cable bundles. The flexible wires within the cable bundle contribute largely to deviations in measurement results. A stochastic cable bundle and statistical simulation method therefore is required. Another challenge is due to the overall complexity of the simulation task. Coarse structures combined with fine details lead to huge memory requirements and very long simulation times, if simulated with a full numerical analysis method. Therefore it is beneficial to divide the problem into subsections and to use different simulation methods. The literature shows a broad variety of methods to model cable bundles stochastically. An overview is given in [1]. Most approaches generate random geometries which aim to resemble one realization of the cable bundle. In the following the stochastic cable bundle model given in [1], [2] is presented. In order for its verification a novel cable bundle test bench was developed [3]. The verified cable bundle model is used in simulations for two EMC methods [4], [5]. II. STOCHATIC CABLE BUNDLE MODEL Fig. 1. Visualization of the stochastic cable bundle algorithm. From the broad variety of stochastic cable bundle models in the literature, those who generate random geometries are most general and capable of modelling many different kinds of cable bundle configurations. One such model from [6] was chosen and further developed. The basic principle of the algorithm is shown in Fig. 1. It consists of a hexagonal grid of positions. 978-1-4673-1088-8/12/$31.00 ©2012 IEEE