本文已被:浏览次 下载次 |
码上扫一扫! |
|
子空间法单输入多输出系统混合相位地震子波提取 |
杨培杰1,2,潘勇3,穆星1, 任向东4
|
(1.胜利油田 地质科学研究院,山东 东营 257015;2.中国石油大学 地球资源与信息学院,山东 东营 257061;3.胜利油田 井下作业公司,山东 东营 257077;4.胜利油田 孤岛采油厂,山东 东营 257000)
|
|
摘要: |
针对高阶统计量混合相位地震子波提取方法要求地层反射系数是非高斯白噪声,并且子波提取速度较慢的缺陷,提出子空间法单输入多输出系统(SIMO)混合相位地震子波提取方法,即基于信号和噪声子空间正交的原理,建立二次型的目标函数,通过对目标函数进行优化就可以提取出混合相位的地震子波,它不需要对地层反射系数做任何假设,因此完全不同于高阶统计量地震子波提取。模型试验和实际应用结果表明,子空间法SIMO系统混合相位子波提取速度快且准确度高,非常适合于无井和井资料不可靠地区的地震子波提取,具有广阔的应用前景。 |
关键词: 子空间法 单输入多输出系统 二阶统计量 混合相位子波 |
DOI: |
分类号::P 631.443 |
文献标识码:A |
基金项目:国家“973”项目(2007CB209605) |
|
Mixed phase seismic wavelet extraction of SIMO system by subspace method |
YANG Pei jie1,2, PAN Yong3, MU Xing1, REN Xiang dong4
|
(1.Geological Scientific Research Institute of Shengli Oilfield, Dongying 257015, China;2.College of Geo Resources and Information in China University of Petroleum, Dongying 257061,China;3.Downhole Service Company of Shengli Oilfield, Dongying 257077, China;4.Gudao Oil Production Plant of Shengli Oilfield, Dongying 257000, China)
|
Abstract: |
The method of high order statistics seismic wavelet extraction requires reflection coefficient to be non gauss white noise and its extraction speed is slow. In order to overcome the shortcomings, a novel mixed phase seismic wavelet extraction method for single input multiple output(SIMO)system by subspace method was presented. Quadratic objective function was built based on the orthogonality between a signal and a noise subspaces, and the seismic wavelets were extracted by optimizing the objective function. The reflection coefficient need not be made any assumption. So this method is totally different from high order statistics wavelet extraction. The results of numerical simulation and application show that this method is characterized by high accuracy and good stability, and it is very suitable for the areas without well materials or reliable well materials. This method has wide application perspects. |
Key words: subspace method single input multiple output(SIMO)system second order statistics mixed phase wavelet |