Printed circuit boards (PCB) are critical components in many aerospace applications. In these applications harsh dynamic environments, such as shock and vibration, are encountered and it is critical that the electronic assemblies endure these conditions. An effective method for minimizing the detrimental impact of these conditions on a PCB is to increase the lowest natural frequency of the PCB through the use of structural supports. In this report we describe a global optimization method, based on surrogate models, that was found to be effective at finding optimal support locations on a PCB. Using a global optimization method within the DAKOTA optimization solver, wrapping a smooth-spline based finite element software, proves effective at handling unreliable gradients and multi-modal solution space. This report was prepared to partially fulfill the course requirements for ME575: Optimization Techniques at Brigham Young University in Provo, Utah. This report has not undergone peer-review.
Accompanying source-code repository: https://github.com/GregVernon/Dakota_CFS