西南石油大学学报(自然科学版) ›› 2010, Vol. 32 ›› Issue (2): 133-137.DOI: 10.3863/j.issn.1674-

• 石油与天然气工程 • 上一篇    下一篇

长时压力监测数据处理新方法与应用研究

张建军a 贾永禄b 张烈辉b   

  1. 西南石油大学:a.理学院,b.石油工程学院,四川 成都 610500
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2010-04-20 发布日期:2010-04-20

A NEW METHOD FOR PROCESSING OF LONG-TERMP RESSURE DATA AND ITS APPLICATION

ZHANG Jian-juna JIA Yong-lub ZHANG Lie-huib   

  1. a.School of Sciences,Southwest Petroleum University,b.School of Petroleum Engineering,Southwest Petroleum University,Chengdu 610500,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2010-04-20 Published:2010-04-20

摘要:

以传统试井理论和小波理论为基础,经过反复试验,寻找到了适合于处理长时压力数据的小波类型。应用多重阀值方法成功地侦测并剔除掉长时压力数据中的异常点;在Chee kin(2001)的线性回归方法基础上,建立了剔除噪音的新方法。模拟和实测长时压力数据的降噪效果说明了该方法不仅能有效地剔除数据中的噪音,还能完整地保留数据中的棱角特征。应用小波模极大值理论成功地将长时压力数据离散为相对独立的一系列不稳定数据段;在保留数据基本特征的前提下,应用最大压力—最大时间准则对原始数据实施了大幅度压缩;在数据处理的基础上,综合应用移动窗分析方法和正交试验方法解释了油田长时压力数据集。解释结果与传统试井解释方法具有一致性。

关键词: 长时压力数据, 小波变换, 噪音, 数据压缩, 正交试验, 移动窗分析

Abstract:

Based on the traditional well test and wavelet theory,it is found the that types of wavelet families are particularly useful in the processing of long-term pressure data by iterative attempt.The iterative thresholding method to detect outliers from the data based on wavelet analysis is developed.This method is applied to actual field data and the method is successful in identifying outliers.Based on the linear regression method from Chee Kin (2001),a new method to remove noise from the data is developed.The denoising process from simulated and field examples is successful in suppressing the noise while the data features are still preserved.The pressure set is discretized into individual transient successfully by using the wavelet modulus maxima.The size of original data is reduced considerably by using the maximum-pressure-maximum-time criterion while the most representative points are kept.Based on the data processing,the moving data window approach,together with the orthogonal experimental method is used to interpret long-term pressure set.It is found that the interpreting result is consistent with traditional well testing interpretation.

Key words: long-term pressure data, wavelet transform, noise, data reduction, orthogonal experiment, moving window analysis

中图分类号: