| Gears are common mechanical components used in power transmissions and they are frequently responsible for transmission failures. Unexpected failures of gears almost always have severe consequences to the bottom line and safety. The diagnosis and prognosis of gears are of growing interest today due to the need to drive-down costs and increase safety margins. Many researchers have attempted to analyze and model the propagation of a fatigue crack on a single gear tooth under a given load and an initial crack size. However, results of these studies can not be easily extended to predict residual life of gears in service because the crack size and tooth load are not directly measurable in service. This research presents a prognostic methodology utilizing a 2D FEM/fracture mechanics based crack propagation model which needs current crack size and tooth load as inputs. The utility of previously developed gear diagnostic algorithms was established to estimate crack size from measured gear vibration. A dynamic model of the gearbox was constructed to estimate tooth loading. Proper integration of the crack propagation model, diagnostic algorithms and gear dynamic model was performed to predict the residual life of a gear. Furthermore, a gear test bed was constructed to facilitate real experiments on which one can validate the crack diagnostic algorithm, the gear dynamic model and the crack propagation model individually and in combination. Extensions and refinements were carried out as appropriate. Finally, to facilitate field deployment of the proposed methodology, a “fast” crack propagation model was established. It is believed that this research will advance the art of gear fatigue crack prognosis greatly. |
