Robert Maul

Nonlinear Model Order Reduction for high Q MEMS gyroscopes

Despite continuously increasing computing speed of modern calculation clusters, the transient simulation of system models with nonlinear effects is still challenging. This is especially valid in the context of highly complex structures such as MEMS gyroscopes, because of the dramatic increase in calculation time with raising level of detail. Model Order Reduction (MOR) methods such as the modal superposition have been successful in the description of linear systems, but cannot be used to simulate nonlinear systems. Here, a nonlinear MOR method for structural mechanics simulations is presented and its application on a simple MEMS geometry demonstrated. The method is a novel approach for computationally efficient transient simulations of nonlinear mechanics occurring in MEMS high Q systems (e.g. MEMS gyroscopes). In a one mass oscillator test system a reduction of computation time by a factor of 1000 is achieved, while the physics of the system is preserved within a good accuracy.