UBC Theses and Dissertations
Advanced MEMS inertial sensors : Simscape modelling and applications in Sliding Mode Control Srinivasan, Sapna
As the technology behind Microelectromechanical Systems (MEMS) continues to progress, more complex microsystems evolve, in turn demanding more sophisticated and efficient macromodels that can be interfaced with the electronic subsystems. The thesis proposes a hybrid simulation approach that combines signal flow models in MATLAB/Simulink with more accurate and reusable energy flow reduced-order macromodels in Simscape to allow the exploration of nonlinear operating modes and bi-directional electro-mechanical coupling in MEMS devices. The physical system modelling capabilities of Simscape behaviour modelling language are used to develop and test a novel MEMS library containing parameterized fundamental building blocks (area and gap-varying MEMS capacitors and displacement stoppers). Simulations of the library models compare favourably with both analytical results and structural finite element analyses performed in COMSOL Multiphysics. The thesis also aims at designing all-digital control and read-out of MEMS accelerometers and gyroscopes, simulated in the Simulink+Simscape hybrid environment. The research reports on the design and simulation of two distinct microsystems, a single-axis MEMS accelerometer and gyroscope, based on devices fabricated in a custom 50um SOI technology. Both microsystems use Sliding Mode Control (SMC) for digital feedback control. Motion cancelling SMC techniques are applied for the MEMS accelerometer and the sensing mode of the MEMS gyroscope, while a bandpass SMC scheme was designed for the driven mode of the MEMS gyroscope to track perturbations in its mechanical resonance frequency and maintain an optimum electrostatic actuation (similar to a digital PLL). In the case of the accelerometer, the digital feedback has attenuated the motion of the proof-mass by 93.2%, while the output bitstream has captured the input quantity (external acceleration) with a steady-state error of 0.06%. Similarly, the motion induced in the sensing mode by an external angular rate was attenuated by 100 times. The output bitstream was subjected to a novel digital synchronous demodulation technique to reconstruct the input angular rate with an error of just 0.16%. The bandpass SMC was able to track and electrically actuate the drive mode to the mechanical resonant frequency with a steady-state error of 0.034%, and a tracking delay of 110ms.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International