无失效数据的MEMS传感器可靠性分析

Reliability analysis of MEMS sensors with zero-failure data

为了解决有限试验次数情况下微机电系统(MEMS)传感器的无失效数据可靠性分析问题,针对寿命服从指数分布的MEMS传感器,通过研究最优置信限法和最小二乘估计这两种方法来获得适用于MEMS传感器无失效数据可靠性分析的置信度α、c(c>1为常数)。分析试验组数对于预测分析结果的影响,通过对两种方法进行不同分组试验,获得MEMS传感器的最佳分组数。针对某数字型双轴倾角MEMS传感器的实际算例表明置信度α=0. 4、c取值范围3~6,分组为12/13组时,得到最佳预测结果,证明了所提出方法的有效性。

In order to solve the limited number of test cases of micro-electro-mechanical system( MEMS) sensor failure data reliability analysis,in view of the MEMS sensor life obey exponential distribution,the confidence degree α and c( c > 1 and c is a constant)suitable for the reliability analysis of MEMS sensorless data are obtained by studying the optimal confidence limit method and least squares estimation. The influence of the number of test groups on the prediction analysis results was studied and analyzed. The optimal tests number of MEMS sensors was obtained by different grouping tests of the two methods. An example of a digital dual axis inclination MEMS sensor shows that the best prediction results are obtained when the values of confidence α = 0. 4 and c range 3 ~ 6 are grouped into 12/13 groups,which proves the effectiveness of the proposed method.