An integrated electro-mechanical model of motor-gear units—Applications to tooth fault detection by electric measurements

Fault diagnosis in geared transmissions is traditionally based on vibration monitoring but, in a number of cases, sensor implementation and signal transfer from rotary to stationary parts can cause problems. This paper presents an original integrated electro-mechanical model aimed at testing the possibility and the interest of tooth fault detection based on electric measurements on the motor stator. The motor is simulated using Kron's transformation while the mechanical transmission is accounted for by a lumped parameter model. Tooth defects are assimilated to distributions of initial separations between the mating flanks whose positions and shapes are controlled. A unique non-linear parametrically excited differential system is obtained, which provides direct access to both the electrical and mechanical variables. A number of results are presented, which illustrate the possibility of tooth fault detection by stator current measurements with regard to the position and dimensions of the defect.

► An integrated dynamic model of gears driven by asynchronous motors is proposed.
► Multiple scales are considered from the motor to tooth defects.
► Results show that tooth defects can be detected by current measurements.
► Tooth spall detection is found to be easier near the tooth central part.