Fault diagnosis of rolling bearing based on k-svd dictionary learning algorithm and BP Neural Network

Mechanical equipment has become the main force of social production and its exist makes production and engineering increasingly efficient. However, behind these advantages, there are hidden dangers. Once mechanical equipment goes wrong, the fault will affect production progress or the life safety of the people. It seems that the fault diagnosis of mechanical equipment is particularly important. In many rotating machinery, rolling bearings are widely used. If the early fault diagnosis can be offered to rolling bearing, then a lot of economic loss and personnel casualties will be avoided. Advocating the efficient security is an integral part to the modernization of engineering work. To define the fault type as soon as possible, this paper denoises the fault signal of rolling bearing by the KSVD dictionary learning algorithm, then the signal will be diagnosised by the BP neural network.