The intense electromagnetic environments (EMEs), such as the intentional electromagnetic interference and electromagnetic pulse, pose severe threats to the normal functions of electric and electronic systems. A system is usually composed of numbers of interdependently linked subsystems or equipments. The interactions of the system and the high-power EME involve large quantities of parameters and scenarios, so the complete tests or computations are usually difficult to fulfill, which leads to a hard mission to assess the system-level electromagnetic vulnerability. This paper provides the thought of divide-and-rule to cope with this problem. First, it divides the system into relatively independent and manageable subsystems, and after respective tests and computations, the subsets of data are fused to characterize the whole system. The key point for this assessment methodology is to set up one model or framework to unify all the activities, which is completed here by the causal Bayesian networks (BNs). The system-level effects and the environment threats are characterized with the probability theory. The modeling and parameter determining techniques are presented. Since fault tree analysis (FTA) is also utilized in the electromagnetic risk assessment, the assessment procedures based on relatively BN and FTA are compared. The final results indicate that BN is capable of extending the modeling and analysis power of FTA.

No abstract

The random parameters remain hard to tackle for a long time in electromagnetic (EM) computations. Aiming at the time‐dependent uncertainty in the axially symmetric structures, the formulations of the stochastic finite‐difference time‐domain (S‐FDTD) method are derived, and the effectiveness is verified by a cylindrical ring structure. The simulation results indicate that the efficiency of S‐FDTD method is much higher than that of the polynomial chaos (PC) method and the Monte Carlo (MC) method, that is, the S‐FDTD method only takes about 3% of the computation time of the PC method and 0.03% of the computation time of the MC method with slightly lower accuracy. Thus the uncertainty in the EM calculation of axisymmetric cylindrical coordinates is possible to tackle, such as the uncertainties in the calculations of lightning electromagnetic pulse, high altitude electromagnetic pulse and so on.

The vulnerability of microcontroller system against high-altitude electromagnetic pulse (HEMP) is taken as an illustration to demonstrate the assessment methodology based on Bayesian networks (BN). The complete procedure is performed by two steps: the qualitative and the quantitative. The first step focuses on the analysis of three classes of properties, the electromagnetic environment, system function/structure, and their interactions. The primary BN model is built at the end of the first step. The second step investigates the BN nodes and branches one by one, which further implemented through two stages, i.e., the data acquisition and data fusion. The susceptibilities of devises are examined with the pulsed current injection. The responses of the transmission lines to HEMP are computed using the field-line coupling model. Comparing the probability density functions of the electromagnetic stresses and strengths produces the failure probabilities of the interface components. Through two-step analysis, the critical elements and coupling paths are identified and highlighted. After neglecting those unimportant factors, many BN nodes and branches are deleted. Thus, the complexity of assessment is reduced. By assigning the probability values to the simplified BN model, the system failure probability is calculated, which characterizes the system vulnerability against HEMP environment. The illustration validates the rationality and flexibility of the BN assessment methodology.

No abstract