基于声发射和GAN-CNN的铝合金管道法兰 连接松动泄漏检测*

面向管道法兰连接松动引起的泄漏检测需求，为解决数据样本不足和减少特征指标手动选取的繁琐环节。本文，考虑到生成性对抗网络（GAN）作为数据扩充工具，已被证明能够生成与真实数据相似的样本数据。同时，卷积神经网络（CNN）作为一种深度学习方法，为自动提取数据的特征提供了一种有效的方法。开展了基于GAN和CNN的铝合金管道法兰连接松动泄漏检测研究。首先，搭建管道泄漏标定和数据采集实验台，利用声发射技术获取不同等级的原始泄漏信号。其次，采用GAN生成样本数据扩充原始数据。同时，为了评估生成模型的性能，引入统计特评估生成质量。最后，将生成的样本数据与原始数据设置为不同训练集，基于卷积神经网络构建智能分类检测模型，应用于管道泄漏检测。同时，分类检测结果与小样本智能分类方法SVM进行了比较，实验结果表明，基于GAN和CNN构建的智能分类模型可显著提高管道法兰连接松动泄漏检测精度。

Leak detection of aluminum alloy pipe due to loosening of flange connection based on acoustic emission and GAN-CNN

Facing the demand of pipe leakage detection caused by loosening of flange connection, in order to solve the insufficient data samples and reduce tedious process of manual selection of characteristic indexes. In this paper, considering generative adversarial network (GAN), as a data augmentation tool, has been proved to be able to generate sample data similar to real data, and as a deep learning method, convolutional neural network (CNN) provides an effective method for automatic extraction of data features, GAN and CNN-based leakage detection of aluminum alloy pipe due to loosening of flange connection is carried out. Firstly, the experimental platform for pipeline leakage calibration and acquisition were set up, and using the acoustic emission technology, raw leakage signals in different levels are obtained. Secondly, GAN was used to generate sample data to expand the raw data. Meanwhile, to evaluate the performance of the generated model, statistical characteristics were introduced to evaluate the generated quality. Finally, the generated sample data and raw data were set as different training sets, and an intelligent classification detection model was constructed based on CNN, which was applied to pipeline leak detection. At the same time, the classification detection results were compared with the small sample intelligent classification method SVM. The experimental results show that the intelligent classification model based on GAN and CNN can significantly improve the leakage detection accuracy of pipe due to loosening of flange connection.