Study on Bearing Load Identification of Rotor Bearing System Based on Artificial Neural Networks

The identification of bearing load is essentially a mapping process from vibration response to bearing load. The characteristics of neural networks is that using sample data without establishing a mathematical model of system may achieve nonlinear mapping of system from Rn (n is the node number of input)space to Rm (m is the node number of output )space. This paper presents a identification method of bearing load of rotor system based on neural networks technique. It has many characteristics such as simple model, controllable accuracy, fast computation, and it has memory capability and the ability to further study. It may provide a theoretical basis for online monitoring the bearing load.