The impact of the design of automotive electrical distribution systems (EDS) is becoming more and more significant with the continuous integration of new safety-relevant functions and the substitution of mechanical systems having reached a high degree of robustness. The introduction of hybrid and electric vehicles amplify this trend and lead to the design of even more complex electrical networks with multiple voltage levels and new challenges.
To assess electrical power systems with respect to their ability to supply the involved electrical consumers in various driving and consuming situations at a high level of reliability and voltage stability simulation studies, bench testing and driving tests are conducted. However, a sustained strategy to define relevant consuming and driving situations in order to test the EDS under consistent loading conditions is missing. The total installed electrical power being much larger than the available generator power, a dedicated strategy to define which electrical loads are likely to be switched simultaneously is needed. As this is dependent upon driving situations and environment conditions and possesses a probabilistic character, an adapted Bayesian network-based strategy is proposed in this paper to generate consistent loading situations in order to assess the design of the power network.
In this paper the stimuli generation based on a Bayesian network is introduced and discussed. The test sequences generated have been implemented on a test bench for electrical power systems to prove their efficiency. The analysis conducted on the generated stimuli and the results produced on the test bench are presented and discussed.