A general regression neural network model for gearbox fault detection using motor operating parameters

Condition monitoring of a gearbox is a very important activity because of the importance of gearboxes in power transmission in many industrial processes. Thus there has always been a constant pressure to improve measuring techniques and analytical tools for early detection of faults in gearboxes. This study forces on developing gearbox monitoring methods based on operating parameters which are available in machine control processes rather than using additional measurements such as vibration and acoustics used in many studies. To utilise these parameters for gearbox monitoring, this paper examines a model based approach in which a data model has been developed using a General Regression Neural Network (GRNN) to captures the nonlinear connections between the electrical current of driving motor and control parameters such as load settings and temperatures based on a two stage helical gearbox power transmission system. Using the model a direct comparison can be made between the measured and predicted values to find abnormal gearbox conditions of different gear tooth breakages based on a threshold setup in developing the model.