Design and Characterization of the MEMS Sensor Fusion for Real Time Damage Detection in Structures
thesisposted on 03.03.2017, 00:00 by Hossain Saboonchi
Micro-electro-mechanical systems (MEMS) have diverse manufacturing capabilities to design and manufacture various sensing elements in order to monitor various mechanical behaviors in structures. In this study, capacitive acoustic emission sensors are integrated with piezoresistive strain sensors on a small footprint device. The integrated sensing allows redundant data measurement from a given point and intelligent data collection strategy in order to increase the reliability of Structural Health Monitoring (SHM) methods. The acoustic emission (AE) sensors are designed with the principle of capacitance change under dynamic excitation, and tuned to the range of 60 kHz to 150 kHz via changing spring and mass geometry. The MetalMUMPs (Multi-User-MEMS-Processes) are implemented to manufacture the sensors on a 1 cm x 1 cm device area. The experimental characterization includes capacitance and impedance measurement, and mechanical simulation experiments including laser, ball impact and pencil lead break for the comparison with conventional piezoelectric sensors. The MEMS strain sensors are designed with the principle of piezoresistivity property of polysilicon, which have higher gauge factor as compared to conventional metal gauges. For the design of strain sensors, trenching concept is implemented to increase the strain transfer. Three strain sensors are placed in horizontal, vertical and angled directions to extract the principle strains. The influences of the sensor position on the silicon substrate and the trenching to the strain transfer from structure under loading to polysilicon layer are numerically demonstrated, and experimentally validated. The characterization experiments include monotonic, cyclic and fatigue mechanical loading, and thermal loading. Combining acoustic emission and strain sensors on the same package can tackle several limitations of SHM methods such as the need of redundant measurement to increase the reliability and defining idle/active mode of acoustic emission sensor using strain sensor to reduce the power consumption, and enabling the integration of the energy harvesting devices. The concurrent performance of MEMS strain sensors and acoustic emission sensors is tested under fatigue loading of two notched aluminum specimens. The strain sensor strengthens the interpretation of complex acoustic emission data via monitoring the driving force and allowing on-chip data filtering in order to process the data recorded from high stress levels.