Krylov-Subspace-Based Order Reduction Methods Applied to Generate Compact- Electro-Thermal Models for MEMS

The high power dissipation density in today’s miniature electronic/mechanical systems makes on-chip thermal management crucial. In order to achieve quick-to-evaluate, yet accurate electro-thermal models, needed for the thermal management of microsystems, model order reduction is necessary. In this paper, we use Krylov-subspace methods for the order reduction of a electro-thermal MEMS model, illustrated by a novel type of micropropulsion device. Comparison between different moment-matching algorithms including a new two-sided Arnoldi algorithm, is performed.