基于多目标优化的高量程加速度传感器小型化设计

Miniaturization Design of High-Range Accelerometer Based on Multi-Objective Optimization

为了实现MEMS高量程压阻式加速度传感器的小型化,并提高其性能,提出一种多目标优化方法。将加速度传感器的性能、尺寸作为多目标函数,并将最优空间填充设计(OSF)和克里金(Kriging)代理模型相结合,构建目标函数与几何变量之间的函数关系。然后将精度较差的样本点作为细化点、不断更新代理模型,直至模型拟合精度满足指标要求。最后利用非支配排序遗传(NSGA-Ⅱ)算法进行多约束优化得到帕累托(Pareto)最优解集。优化结果表明,传感器的尺寸减小了29.33%,一阶谐振频率提高了27.94%,在100000 g_(n)过载下的理论灵敏度提高了23.64%,且不会降低该高量程加速度传感器的抗过载能力,从而验证了该方法的有效性。

Towards the miniaturization and performance improvement of MEMS high-range piezoresistive accelerometer,a multi-objective optimization method was proposed.The performance and dimension of the accelerometer were used as multi-objective functions,and the optimal space filling design(OSF)and Kriging proxy model were combined to construct the functional relation between objective functions and geometric variables.Then the sample points with poor accuracy were taken as refinement points,and the proxy model was constantly updated until the model fitting accuracy met the index requirements.Finally the non-dominated sorting genetic(NSGA-Ⅱ)algorithm is used for multi-constraint optimization to obtain the Pareto optimal solutions.The optimization results show the dimension of the sensor is reduced by 29.33%,the first-order resonance frequency is increased by 27.94%,and the theoretical sensitivity under 100000 g_(n) overload is increased by 23.64%.And the anti-overload ability of the highrange accelerometer is not reduced,which proves the effectiveness of the method.