The conventional approaches for electromagnetic shielding structures’ design, lack the incorporation of uncertainty in the design variables/parameters. In this paper, a reliability-based design optimization approach for designing electromagnetic shielding structure is proposed. The uncertainties/variability in the design variables/parameters are dealt with using the probabilistic sufficiency factor, which is a factor of safety relative to a target probability of failure. Estimation of probabilistic sufficiency factor requires performance function evaluation at every design point, which is extremely computationally intensive. The computational burden is reduced greatly by evaluating design responses only at the selected design points from the whole design space and employing artificial neural networks to approximate probabilistic sufficiency factor as a function of design variables. Subsequently, the trained artificial neural networks are used for the probabilistic sufficiency factor evaluation in the reliability-based design optimization, where optimization part is processed with the real-coded genetic algorithm. The proposed reliability-based design optimization approach is applied to design a three-layered shielding structure for a shielding effectiveness requirement of ∼40 dB, used in many industrial/commercial applications, and for ∼80 dB used in the military applications.
Performance Analysis of Neural Networks Training using Real Coded Genetic Algorithm