Optimal shape design of electrostatic microactuators: A multiobjective formulation

Abstract

A procedure of automated optimal design of a class of micro-electromechanical devices is proposed in terms of a multiobjective optimization formulation. A comb-drive electrostatic accelerometer is considered as the case study. By optimizing the shape design of the device, it is desired to extend the actuation range of the comb drive in steady-state conditions, leading to greater gap distances between the electrodes. The design variables control the geometry of fixed and movable electrodes, in terms of width and height. The twofold goal of maximizing the drive force considered as the main criterion, while minimizing the levitation force considered as a parasitic effect, makes it possible to implement a Pareto optimality approach.

Non-linear inverse problems and optimal design of MEMS

“…The Pareto front in Figure 8(b) was previously obtained for 15 individual points and ten generations, after 10 h 30 min of computation time, on the same computing platform, but without surrogate models. The vastly required computing resources and higher probability of degeneration of the finite elements in the mesh at micro and nano scales, inhibited further improvements of the optimization method described in Chereches et al (2013). These findings corroborate the need for alternative design optimization methods, Non-linear inverse problems and optimal design addressed to applications such as MEMS: the method here presented is a contribution in this direction.…”

“…This is in fact about half the time needed for optimizing the same device without considering surrogate modeling (Chereches et al, 2013). The average values (with respect to the Pareto set) of the final input (design variables) and output data (objective functions) for the optimization problem show a close agreement between the two cases of optimization, confirming the validity of the proposed optimization method (Table I).…”

Evolutionary Computing and Optimal Design of MEMS

“…Second-order Lagrangian shape functions were considered in the finite-element model: a typical mesh (see Fig. 3) is composed of 170 000 elements with 240 000 unknowns [20], [23], approximately. The device is considered electrically isolated: the boundary condition of the air subdomain is set to zero charge density; moreover, the fixed electrodes are at the same potential as the grounded substrate, while the movable electrodes are subject to voltage u 0 = 1 V. The device components are made of polycrystalline Si exhibiting a relative permittivity ε r = 4.5.…”

Minimizing design criterion and sensitivity: Cost-effective evolutionary approach with application in mechatronics

“…Usually, the design of MEMS is approached in a systematic way in terms of non-linear multi-objective optimization of design criteria subject to a set of constraints [9,10].…”