WEAR PREDICTION OF NBR IN CRUDE OIL BASED ON BP ARTIFICIAL NEURAL NETWORK

The wear behavior of rubber stator belongs to the typical non-linear problem, because it is affected by multiple factors. In the background of NBR tribology system, a wearing model has been configured in this article based on BP artificial neural, in which hexane, cyclohexane, toluene, and rubber wearing are considered as the input and output factors, respectively. The wear loss is predicted successfully in different kinds of crude oil, and the error is only 3.98% comparing with the experimental results.