Efficient Methods for Yield Optimization using CST Microwave Studio
Deviations in the manufacturing process of high-frequency electronic components, waveguides or antennas may lead to rejections due to malfunctioning. Uncertain design variables can be modeled as random variables. Then, the probability that a product fulfills its performance specifications, e.g. the scattering parameter, is called the yield. The performance specifications are restrictions involving the partial differential equations describing the electromagnetic field, i.e., Maxwell’s equations in time or frequency domain discretized by the Finite Elements Method (FEM). Recently, new methods have been proposed to more efficiently estimate the yield by using a hierarchy of low- and high-fidelity simulation models. In this project those methods shall be extended and implemented using CST Microwave Studio as FEM solver.
The problem setting must be derived, and available algorithms understood. Then, the main part is the efficient implementation and possibly extension of a new hybrid algorithm using CST Microwave Studio. It is recommended to carry out this work externally at Dassault Systemes Deutschland GmbH in Darmstadt.