An artificial intelligence-based approach for identifying the in-plane orthotropic mechanical properties of electronic circuit boards

Abstract

The finite element modeling of electronic boards is a challenging task due to the complexity of the multi-component board structure. Hence, it is acceptable to attain equivalent orthotropic in-plane mechanical properties and use them throughout the finite element analysis (FEA) simulations. This paper aims to present an artificial intelligence-based methodology, using the artificial neural networks (ANNs), to estimate the in-plane mechanical properties of the printed circuit boards (PCB). In this methodology, the ANN technique used FEA data to find the relationship between the first 10 natural frequencies and the mechanical properties, that is, modulus of elasticity, Poisson’s ratio and the shear modulus, of the test board. Subsequently, the experimentally derived natural frequency data is then imported to the ANN model to identify the equivalent orthotropic properties. The ANN-predicted properties are plugged back into FEA and provided natural frequencies and mode shapes that are in great match with experimental results.

Artificial Neural Networks Modeling of the Electricity Demands in the Jordanian Industrial Sector

“…In other words, instead of using more data points, the present paper trained more neural networks to maximize accuracy and ensure solution convergence. In fact, this approach has been previously successfully adopted in recent research studies of the authors [24][25][26][27].…”