Data-driven modeling of staged planetary gearbox vibration signature

Modeling planetary gearboxes is an important field of research, since it helps in condition based monitoring of the system. Multiple models with different approaches were developed last decades. They showed interesting results, but they did not reflect all involved phenomena. Indeed, specifically in a complex rotating systems such as staged planetary gearboxes, those models only approximated the dynamic behavior of the investigated system. Hence, in this paper, an algorithm of modeling staged planetary gearbox vibration signature is proposed based on exclusively measured vibration signals. First, in section 2, the dynamic mode decomposition (DMD) will be introduced. Second, in section 3, the DMD algorithm will be implemented on simulated mathematical functions to prove its robustness on future state prediction. Then, in section 4, the experimental staged planetary gearbox test rig is presented. Finally, in section 5, the DMD algorithm is applied to multiple measured vibration signals with different operating conditions as given in section 4. Interesting results are conducted and limits of the approach are presented.