Modeling and Simulation of MEMS-based Comb Drive Actuator using Bond Graph Method

摘要

A comb drive is an important component of MEMS-based actuators used in many applications such as a dispenser and a fluidic pump, and as a sensor in an accelerometer and in MEMS gyroscope. Such systems are coupled multi-domain energy systems popularly known as mechatronic systems. The modeling and simulation of such systems poses a formidable challenge. Bond Graph Method (BGM) has emerged as a method of choice for the modeling and simulation of coupled multi-domain energy systems. The system is seen as an interplay of power variables and energy variables, thus introducing the immortal concept of cause and effect. The two primary forces, electromechanical and electrostatic, are cast into the cause and effect paradigm whilst expressing power into effort and flow variables. This research work focuses on investigating the control parameters of lateral electrostatic comb-drive actuators, where the effect of these parameters on the actuation performance is explored using a commercial bond graph package known as 20-SIM. This type of analysis is essential for the design process of system-on-a-chip MEMS applications, where a comb-drive can represent the main source of actuation within the chip system. Of particular interest to this study is the application of the electrostatic comb-drive motor as a fluidic pump in on-a-chip systems used for drug delivery or cooling of microprocessors used in space vehicles. The 20-SIM computer software is used to construct a robust comb-drive model and solve its multi-physics interaction problem using the power of bond graph method. In this model, the driving voltage is taken into account. The calculated comb displacement and the electrostatic force generated are known to be directly proportional to the square of the driving voltage. In this method, a model based on BGM is developed and then directly simulated using 20-SIM. Since a bond graph is a precise mathematical model, it can be used to unfold the state-space equations associated with the physical dynamic system. The bond graph simulations lead to the desired state space equations, which unfold the dynamic response of the physical system.