Model Based Diagnostics for Gearboxes

A model-based diagnostic to detect and assess faults is applied to gearboxes. Signals measured from a gearbox were compared to corresponding signals from simulations. The gearbox simulation model includes dynamics of input, lay and output shafts, gear bodies, bearings, and kinematics and tribology of gear tooth contacts. These were assembled into a systems model via bond graphs. To make simulations emulate measurements, the model’s parameters were “tuned”. A direct physical correspondence between parameters and real system elements locates faults. Considered was gear tooth cracking. Faults disturb normal flow of power and signal through a machine. Viewing the gearbox as a communications channel, and applying information theory theorems to the perturbed signals permitted assessment of the gearbox’s functional health.