Corrosion Protection Status Survey of Submarine Pipelines Based on Bayesian Regulated Neural Networks

摘要

Bayesian Regularization Neural Network is used in the estimatingrnsystem for corrosion protection status survey of submarine pipeline.rnFirstly it discusses about the theory and approach to mend backpropagationrn(BP) neural network using Bayesian Regularization trainingrnalgorithm, then using this algorithm to train and validate an actualrnNeural Network based on a particular simulation calculated model forrnthe submarine pipeline cathodic protection system, and it was provedrnthat this algorithm effectively improved the precision and generalizationrncapability of BP Neural Network. This approach is used in thernestimating software system and need further extension.