光学 精密工程, 2014, 22 (6): 1500, 网络出版: 2014-06-30
硅微谐振式加速度计驱动电路参数优化
Parameter optimization of drive circuit in silicon resonant accelerometer
硅微谐振式加速度计 闭环驱动 遗传算法 非线性系统 silicon resonant accelerometer closed loop driving genetic algorithm nonlinear system
摘要
将遗传算法与低频模型相结合, 提出了一种快捷的驱动电路设计方法, 用于提高低功耗硅微谐振式加速度计模拟驱动电路的瞬态性能, 并缩短设计周期。该方法通过对闭环驱动电路模型进行高低频解耦, 提取闭环驱动电路的低频模型; 将提取的低频模型与遗传算法相结合, 给出完整的优化方法, 得到了满足各项实际约束的最优电路参数。针对某型硅微谐振式加速度计, 建立了SIMULINK低频仿真模型, 根据实际情况制定了约束条件。应用该方法求出了系统启动速度最快的PI控制器的参数, 并对其进行了实验验证。起振实验结果表明, 采用优化参数可使超调量小于50%, 相位误差小于5°, 1%调节时间从优化前的0.42 s减少到优化后的0.19 s, 实验与仿真误差小于5%。得到的结果证明提出的方法正确有效, 具有可实施性。
Abstract
A fast design method for the drive circuit was proposed by combining Genetic Algorithm(GA) with a low frequency model to improve the transient performance of analog drive circuit for a low-power Silicon Resonant Accelerometer(SRA) and to shorten its design cycle. The method decoupled the closed drive circuit model in high and low frequencies to extract a low-frequency model from drive close-loop circuits. Combined the low-frequency model with the GA, an optimization method was proposed to optimize the circuit parameters for meeting the different actual restraints. A simulation model was established in SIMULINK based on one type of micro silicon resonant accelerometer, and the optimal parameters of PI controller with a most start-up speed were obtained under constraint conditions. Finally, a start-up experiment was performed to testify the simulation results. It shows that the start-up time is shorten from previous 0.42 s to 0.19 s and the over-shoot and phase error are less than 50% and 5°, respectively. The difference between the simulation and experiment is less than 5%, which falls within the acceptable range. It proves that the optimization method is correct and effective.
赵健, 苏岩, 赵阳, 夏国明. 硅微谐振式加速度计驱动电路参数优化[J]. 光学 精密工程, 2014, 22(6): 1500. ZHAO Jian, SU Yan, ZHAO Yang, XIA Guo-ming. Parameter optimization of drive circuit in silicon resonant accelerometer[J]. Optics and Precision Engineering, 2014, 22(6): 1500.