Modal analysis and condition monitoring for an electric motor through MEMS accelerometers

Piezoelectric accelerometers are commonly employed for diagnosing machine faults, due to their accuracy. In the last few years, however, MEMS (Micro Electro-Mechanical Systems) accelerometers have attracted strong interest thanks to their low cost. In this work, a synchronous electric motor with an integrated MEMS sensor is studied and results are compared from both MEMS and piezoelectric sensors. A modal analysis is performed, using data from all available sensors. Comparing the frequency response functions and the natural frequencies shows the limitations of the MEMS sensor. One can then correct the MEMS measurements, by using global statistical parameters calculated on the data or by defining a “filter” function between the signals, thus improving the signal-to-noise ratio. It is found that MEMS sensors may replace piezoelectric ones for diagnostic applications. This way, an inexpensive measurement system (which needs to be calibrated only once, before installation, against higher-accuracy sensors) can be used for vibration monitoring of electric motors.