Accelerated life tests for prognostic and health management of MEMS devices

Abstract

Microelectromechanical systems (MEMS) offer numerous applications thanks to their miniaturization, low power consumption and tight integration with control and sense electronics. They are used in automotive, biomedical, aerospace and communication technologies to achieve different functions in sensing, actuating and controlling. However, these microsystems are subject to degradations and failure mechanisms which occur during their operation and impact their performances and consequently the performances of the systems in which they are used. These failures are due to different influence factors such as temperature, humidity, etc. The reliability of MEMS is then considered as a major obstacle for their development. In this context, it is necessary to continuously monitor them to assess their health status, detect abrupt faults, diagnose the causes of the faults, anticipate incipient degradations which may lead to complete failures and take appropriate decisions to avoid abnormal situations or negative outcomes. These tasks can be performed within Prognostics and Health Management (PHM) framework. This paper presents a hybrid PHM method based on physical and data-driven models and applied to a microgripper. The MEMS is first modeled in a form of differential equations. In parallel, accelerated life tests are performed to derive its degradation model from the acquired data. The nominal behavior and the degradation models are then combined and used to monitor the microgripper, assess its health state and estimate its Remaining Useful Life (RUL).