Development of a methodology for MEMS accelerometer health state estimation

The spread of micro electro-mechanical systems (MEMS) in industrial and civil applications that has occurred in recent years is mainly due to the small physical size of these devices and their high robustness and reliability. The diffusion has also occurred in mission critical areas of use, increasing the responsibilities of decisions based on MEMS devices' measurements. This, therefore, necessitates the design and definition of health parameters that can provide Remaining Useful Life (RUL) information for prognostic purposes to increase reliability in mission-critical applications. However, given the substantial difference in aging processes compared to other apparatuses studied in the literature, such as large industrial machinery, models for studying aging and RUL of such devices based on online measurable features have never been presented. Therefore, the objective of this study is to propose a Health Index for estimating RUL based on measurements from MEMS devices, in this case, accelerometers. The results were satisfactory, showing device failure recognition more than 40 days in advance of actual failure.