At present, automotive functional safety and EMC engineering are largely carried out independently. Current EMC regulations aim to avoid unwanted disturbances by setting appropriate immunity threat levels and emission limits. However, with the rapidly evolving technology and complexity of automotive systems, the limits identified in standards may no longer be appropriate. Hence the identification and assessment of EMC-related risks are becoming increasingly necessary. This paper outlines the tools used to support risk analysis for functional safety and presents initial proposals for a graphical method to better align the analysis of EMC risks and functional safety.

The performance of two oscillator circuits, namely a current-starved voltage controlled oscillator and a ring oscillator, is compared with respect to multi-tone direct power injection (DPI). The objective is to investigate the impact of causal dependence between multi-tones on the immunity levels of integrated blocks with different architectures but similar functionality. The multi-tone immunity analysis performed using the probabilistic noisy-OR model reveals an increase in the probability of failure due to electromagnetic interference relatively to single-tone EM disturbance. The proportions of inhibition and positive causality, as well as the mean degree of synergy (DoS) caused by multi-tone EM disturbances, are extensively compared for both oscillator circuits.

The reliable operation of an integrated circuit can be affected by environmental changes, such as of multiple frequency electromagnetic (EM) disturbances and temperature variations. This paper compares the performance of two oscillator circuits, namely a current-starved voltage controlled oscillator and a ring oscillator integrated into a chip, in terms of their immunity to multitone direct power injection while under the influence of thermal stress. The objective is to demonstrate by the means of measurements the synergistic effect caused by multitone EM disturbances in a test chip, in contrast to the conventional single-tone EM disturbances. Moreover, the multitone immunity levels of integrated blocks with different architectures but similar functionality are analyzed at extreme temperature deviations. Bayesian networks (BN) are applied in order to visualize the probability of circuit failure due to multitone disturbances and temperature influence. Additionally, noisy-OR and improved-adaptive-recursive-noisy-OR (I-ARNOR) probabilistic models are implemented to identify the types of causal interactions (i.e. inhibition and positive causality) between multitone disturbances and to predict the probability of failure due to higher order multitone disturbances, respectively.

No abstract

Current trends in the automotive industry are reshaping the architectures and electromagnetic characteristics of road vehicles. Increasing electrification and connectivity are enabling considerable packaging flexibility and leading to radically different electromagnetic environments. At the same time, increasing automation of driving functions will require unprecedented levels of system dependability. However, existing EMC engineering processes were developed in a very different world of low system complexity and incremental technological development. In order to adapt to rising system complexity and the increasingly rapid pace of technological change, it is considered that a more agile risk-based approach is better suited to ensure the electromagnetic resilience of future vehicles and other complex systems. This paper outlines a Bayesian network approach that allows the combination of both technical and nontechnical aspects in assessing the likelihood of issues that could lead to system-level risks. This approach could be used from the earliest stages of product development, where the detailed information required to undertake detailed risk assessment is generally unavailable.