海洋平台油气管道疲劳裂纹AE信号特征提取及识别研究

A study on feature extraction and recognition of fatigue crack AE signals of oil and gas pipelines in offshore platforms

为了将声发射(AE)技术实际应用到监测海洋平台油气管道疲劳裂纹中,需要解决管道振动干扰以及疲劳裂纹AE信号有效特征提取的问题,而问题的关键在于对管道结构疲劳裂纹AE信号特征提取及识别算法的研究。在已有研究的基础上,提出了一种基于经验模态分解(EMD)为特征提取的疲劳裂纹识别方法,将管道振动干扰问题和疲劳裂纹AE信号有效特征提取问题联系在一起,对特征元素进行优化并剔除无效噪声干扰信息,通过概率神经网络(PNN)对疲劳裂纹信号进行识别。试验结果表明,PNN结合基于EMD为特征提取的疲劳裂纹识别法能够取得良好的效果,为声发射技术监测海洋平台油气管道疲劳裂纹提供了试验和理论依据。

In order to apply acoustic emission(AE)technology to the monitoring of fatigue cracks on oil and gas pipelines of offshore platforms,it is necessary to solve the problems of pipeline vibration interference and effective feature extraction of fatigue crack AE signals.The key to the problem lies in the study of feature extraction and identification algorithms of AE signals for fatigue cracks in pipeline structures.Based on existing research,a fatigue crack identification method was proposed based on empirical mode decomposition(EMD)as feature extraction.The problem of pipeline vibration interference and effective feature extraction of fatigue crack AE signals were linked.The characteristic elements were optimized to eliminate invalid noise interference information,and the fatigue crack signal was identified by a probabilistic neural network(PNN).The results show that PNN combined with the fatigue crack recognition method based on EMD as feature extraction can achieve a good result,which provides an experimental and theoretical basis for acoustic emission technology to monitor fatigue cracks of oil and gas pipelines on offshore platforms,and has certain guiding significance.