Model Order Reduction of a Nonlinear Electromechanical Beam Actuator by Clustering Nonlinearities

This contribution presents the finite element modelling and nonlinear model order reduction of an electrostatic beam microactuator with mechanical contacts due to pull-in. These nonlinearities are rewritten as inputs to enable the application of linear model order reduction. The linear part of the model is reduced by Krylov-subspace-based methods, whereas the nonlinearities are grouped into a smaller number of inputs. Present contribution proposes an agglomerative clustering algorithm for this grouping process to improve approximation quality. A comparison of the reference model and the reduced model demonstrates the latter’s accuracy and computational efficiency. Furthermore, the reduced model enables characterization of the electrostatic hysteresis. Finally, it is deployed within a control loop, demonstrating its feasibility for designing controller architectures and for use in control loops.