Prediction Model for Erosion-Corrosion of Seawater Bend Based on Fluent and Neural Network

A model for predicting corrosion sensitive parts and erosion-corrosion rates of seawater bend was established by the simulation of flow field in a seawater bend using flow velocity as the main influencing factor of bend erosion-corrosion and FLUENT fluid simulation software, in combination with the simulation of the correlative equation between flow velocity and erosion-corrosion rate in seawater bend by neural network analysis. The validity of the model was verified by experiments in real sea. The model prediction and verification experiments got the same result that the worst corrosion damage occurred in the vicinity of the inlet with section angle of 30°. The model can effectively predict the corrosion sensitive part and the erosion-erosion rate of the bend.