Study On Multidisciplinary Design Optimization And Its Application In The Design Of MEMS

MEMS is a complex system which refers to several disciplines such as micro-electronics , micro-mechanics and micro-fluidics . The design of MEMS is complex due to the coupling between several disciplines and multi physics fields . Considering the problem , Multidisciplinary Design Optimization (MDO) method is introduced to the design of MEMS . A Multidisciplinary Variable Couple Design Optimization method for non-hierarchic system is studied , then the method is used in the design of micro-machined silicon accelerometer . The research is helpful to the design of complex system such as MEMS which refers to several disciplines.The main points of this paper are as follows:1. Multidisciplinary Design Optimization (MDO) method is introduced to the design of MEMS . A novel MDO framework for MEMS– MEMS Multidisciplinary Design Optimization System (MMDOS) is established,and the working principle of each part is given . Considering hierarchic system , non hierarchic system and hybrid hierarchic system, four kinds of method of MDO are advanced, including Multidisciplinary Hierarchic Compatibility , Multidisciplinary Variable Couple , Multidisciplinary Object Compatibility and Multidisciplinary Mixed Coordinative .2. A Multidisciplinary Variable Couple Design Optimization method for non-hierarchic system is studied . The working principle of system decomposition , optimal modeling of subsystem , the establishment of couple function and system-level coordination are proposed. To handle the coupling between subsystems , couple function is established by ideal value during system-level coordination , which is used to coordinate the independent optimization of subsystem and finally obtain global optimal solution . Based on this method a MVCDO algorithm is programmed .3. Two representative MEMS products—comb-finger micromachined silicon accelerometer and pendulous micromachined silicon accelerometer are studied and expressions of peel number are proposed to predict the adhesion-related failure . By analyzing the main points in the design of the two kinds of micromachined accelerometers and interactions between different disciplinarys , MDO model is established and solved by Multidisciplinary Variable Couple Design Optimization method. The results show that the performance of the two kinds of micromachined accelerometers is greatly improved , and the Multidisciplinary Variable Couple Design Optimization method as well as the MVCDO algorithm proves to be efficient.This paper is important to the theoretical study and application of related fields.