基于小波变换的自适应多井对比技术

Adaptive multiwell correlation based on wavelet transformation

清华大学自动化系智能系统与技术国家重点实验室,北京100084

The State Key Laboratory of Intelligent System and Technology, Automation Department of Qinghua University, Beijing 100084

针对多井对比中的层位识别问题,提出了利用自适应小波变换进行地层自动划分的解决办法。该方法首先利用已知井的标志层信息,提取该标志层的小波尺度信息,在该小波尺度信息的基础上得到一个小波尺度序列;然后利用该小波尺度序列对未知标志层井段的测井信号进行相应的小波变换,对得到的变换序列利用奇异值点进行分割得到一组包含后选标志层的层段;最后利用动态规划算法计算已知层序列和被分割序列的最佳匹配距离。在一系列小波尺度序列对应的最佳匹配距离中,必有一最小者,而该最小者对应的小波尺度即为对该井进行分层的最佳尺度,因此可以根据各个井测井信息的不同自动调整小波变换尺度参数,大大减少用户调整参数的工作量,避免人为的误差。

WT5”BZ] This paper presents a wavelet transformation based adaptive formation zonation method for formation identification between multiwells. The method first extracts wavelet scale information from known marked beds in wells and forms a wavelet scale sequence. The wavelet scale is then used to perform wavelet transform on logging signals of unknown marked beds and the resulting transformed sequence are sorted by singular values to form groups containing candidate marked bed segments. And lastly the optimum matching distances between the sequence of known beds and the resulting sequence are calculated by dynamic programming. The wavelet scale corresponding to the minimum distance among these distances is just the optimum scale for zonation in the well, thereby the wavelet scale in different wells is adjusted automatically. Thus the amount of parameter adjustment by user is reduced greatly and unintentional errors are avoided.