示功图特征提取选择在油井故障诊断中的研究应用  

The research and application of the eigenvalue extraction and selection in oilfield fault diagnosis

作　　者：丁漪[1] 李训铭[1] 

机构地区：[1]河海大学,江苏南京210036

出　　处：《电子设计工程》2014年第17期148-150,共3页Electronic Design Engineering

摘　　要：示功图特征值提取是在功图数据中挑选具有代表性的数据,经过计算处理得到最有效的特征值,作为故障诊断系统中神经网络的输入。本文用了几何参数法、灰度矩阵法来获取示功图的特征值,减小了神经网络的输入规模,使特征值更具有针对性,提高了神经网络的训练速度,从而使诊断的准确性更高。该方法已在江苏油田的实际应用中取得了良好的效果。The eigenvalue extraction of the indicator diagram is to select the most typical data from diagram data and get the most effective eigenvalue which is used as the input of the neural network in fault diagnosis system by calculating and processing them. This paper obtained the characteristic value of the indicator diagram by the methods of geometric parameter and gray-level matrix. These methods have largely reduced the scale of input vector of the neural network, made the characteristic value more typical and improved the training speed of neural network, so as to make the diagnosis more precise. This method has obtained good effect in the practical application of Jiangsu oilfield.

关 键 词：神经网络 故障诊断 示功图 特征值提取 灰度矩阵 

分 类 号：TN-7[电子电信]