井下钻杆表面螺纹缺陷智能检测系统

Intelligent system of thread defect detection for downhole drill pipe surface

钻杆修复是石油勘探和生产过程中的关键问题。及时对钻杆螺纹表面多类型的缺陷进行高效检测,是油田企业安全生产的重要保障之一。为解决当前人工目检效率低的难题,该文提出一种基于深度神经网络的螺纹缺陷检测模型,模型自主学习缺陷样本分布规律,精准、快速检测缺陷位置和类别;针对检测车间环境,设计了专用的双轴机械运动系统和操控显示软件,自动实现整个螺纹缺陷检测流程。实验结果表明:该检测系统的精度已经超越了人工目检,显著提高了工厂钻杆缺陷检测的效率。

The drill pipe restoration is a key problem during exploration and exploitation of petroleum.Detecting the various thread defects on the drill pipe surface efficiently in time is a crucial guarantee for the production safety of oilfield enterprises.To address the difficulty of low efficiency of manual visual inspection,a thread defect detection model based on deep neural networks is proposed,which can autonomously learn the distribution rules of defect samples to locate the position and identify the category of defects accurately and quickly.According to the environment of the workshop,a special two-axis mechanical motion system and a control display software are designed to automatically complete the whole thread defect detection process.The experimental results show that the accuracy of the detection system is higher than manual visual inspection,the efficiency of drill pipe defect detection is significantly improved in the factory.