Intelligent Analysis of Macro Failure Image of Oil Country Tubular Goods
The intelligent recognition method of oil country tubular goods(OCTG)failure images based on deep learning is currently a re-search hotspot.In response to the problems of poor accuracy and poor generalization ability of traditional recognition networks/models,based on the residual block stacking design and an intelligent recognition model/algorithm,an integrating attention and residual network for OCTG failure images classification was established.This algorithm endows the network with the function of assigning weight to typical macroscopic images of OCTG,such as fracture,corrosion,wear,deformation,effectively improving the ability of the neural network to capture key fea-ture information in macroscopic images.Compared with traditional BottleNeck structured networks,the average classification accuracy of the AXBlock has been improved from 87.52%to 94.93%.This research work will provide technical support for intelligent diagnosis,failure a-nalysis,prediction,and prevention of OCTG failures.
oil country tubular goodsdeep learningneural networkfailure analysismacroscopic images
NETL
NSTL
万方数据
维普
