Research on gearbox wearing prognosis based on Gamma-State Space Model

Gear wearing is the result of physical and chemical effect and it is also the ultimate reason of indirect condition information changes. On one hand, the condition information which can directly reflect the gear wear degree is difficult to measure. On the other hand, there is a mount of indirect condition information that is obtained by condition monitoring system which can reflect the health of gears. On the basis of the full lifetime experiment of gearbox and a few direct condition information and plentiful indirect condition information, established the Gamma-State Space Model (SSM) which was used in analyzing the gear wear states and predicting the development of wearing out processes. Furthermore it brings forward a parameter estimation method which combines Experience Maximization (EM) algorithm and Particle Filter (PF) together. Finally, by comparing the prognostic results with the experiment, the efficiency of the model is validated.