基于多分量的井間地震縱橫波分離方法研究

1、基于多分量的井間地震縱橫波分離方法研究        【中文摘要】井間地震技術(shù)作為一項(xiàng)特高分辨率的勘探技術(shù),在油氣開(kāi)發(fā)與生產(chǎn)中將發(fā)揮越來(lái)越重要的作用。在井間地震資料處理技術(shù)中,主要有直達(dá)波旅行時(shí)層析成像技術(shù),和井間地震反射波成像技術(shù)。在傳統(tǒng)的這兩種處理技術(shù)中大多都只是利用了縱波信息,而對(duì)橫波的研究比較少。這一方面是由于對(duì)橫波的重要性熟悉不夠;另一方面是橫波本身的復(fù)雜性,由于橫波能量一般比縱波低,一般被縱波給覆蓋,不輕易提取。在地震勘探已由單純的構(gòu)造勘探發(fā)展到構(gòu)造巖性勘探的今天,現(xiàn)代已擁有的技術(shù)水平可以解決橫波面臨的

2、一些題目了。假如我們能較為正確地獲得橫波資料,就能獲得井間介質(zhì)橫波速度的二維分布以及高分辨率橫波疊加剖面。這對(duì)于研究地層的物性(巖性、孔隙度、流體成分等)是十分重要和有利的。而目前常用的波場(chǎng)分離法都是對(duì)地面多波記錄的,很少涉及到井間激發(fā)的情況。因此,尋求適合井間地震資料的縱橫波分離方法就成了井間地震處理技術(shù)中的關(guān)鍵。本文從井間地震資料的特征出發(fā),對(duì)于井間多分量地震資料的縱橫波分離,分離步驟基本如下:首先,對(duì)多個(gè)分量數(shù)據(jù)進(jìn)行濾波處理,消除管波等相干干擾波,假如是三分量數(shù)據(jù),還得由矢量合成法將其轉(zhuǎn)化為二分量數(shù)據(jù);接著,對(duì)水平分量和垂直分量進(jìn)行上下行波的分離;再分別對(duì)上行波和下行波,由兩分量計(jì)算出偏

3、振角,將偏振角應(yīng)用到波場(chǎng)分離矩陣,就能分離出上下行波各自的縱橫波;最后,將上行波的縱波和下行波的縱波相加得終極的縱波剖面,將上行波的橫波和下行波的橫波相加得終極的橫波剖面。對(duì)于上下行波的分離,采用頻率空間域-p變換,計(jì)算中采用廣義逆和-p域中的修邊處理能極大壓制假頻和端點(diǎn)效應(yīng),采用奇異值分解法能大大進(jìn)步計(jì)算效率。對(duì)于偏振角的求取,采用協(xié)方差矩陣法,這種方法不需要任何速度等先驗(yàn)信息,只需要輸進(jìn)分離后的二個(gè)分量的上行波(或下行波)即可。在構(gòu)建協(xié)方差矩陣時(shí),時(shí)窗的選取是一個(gè)很重要的因素,本文給出了道自相關(guān)法和過(guò)零點(diǎn)法。將偏振角應(yīng)用到由井間地震資料的偏振分析得到的波場(chǎng)分離矩陣上,也就是采用偏振旋轉(zhuǎn)變換

4、,就能分離出縱波和橫波。對(duì)理論水平層狀模型和實(shí)際測(cè)井資料模型上模擬出的記錄進(jìn)行試算表明:在終極得到的縱波剖面里,橫波能量明顯減弱,縱波得到保存。在得到的橫波剖面里,直達(dá)縱波等縱波成分得到極大的壓制,原先被縱波能量掩蓋的橫波成分顯現(xiàn)出來(lái),清楚可見(jiàn)。因此,本文給出的縱橫波波場(chǎng)分離法對(duì)于井間地震多波資料取得了良好效果,該方法充分利用了地震波的偏振特性和視速度特性,計(jì)算效率高,抗噪能力強(qiáng),而且稍加修改,還能應(yīng)用到地面和海上多分量地震資料中。');【Abstract】 Crosswell seismic technique as a high resolution exploring techn

5、ique becomes more and more usefull in the petroleum exploitation and produce.In the crosswell seismic data processing, we mainly use the seismic tomography with directed wave and obtain high resolutional stack section with reflected wave.In traditional processing techniques of the two methods,we gen

6、erally use the s-wave only,not considering the s-wave.One side,it's caused by our inadequate understand on S-wave,on the other hand,s-wave is more complicated than p-wave. The energy of the s-wave is very low and is covered by p-wave,so we can't extract s-wave from multiple waves easily.Toda

7、y,the seismic exploration has developed from single structure exploration to structure lithology exploration,and we now can resolve a few problems about S-wave.If we can obtain the s-wave data from originality wavefield exactly,then we can obtain the S-wave velocity field of the medium between the s

8、timulate well and receive well.This is very important and usefull to investigate the physical property of the stratum(lithology,porosity,fluid component,et al.).So,the separation of s-wave and s-wave from crosswell seismic data is absolute significant in the crosswell seismic processing.This article

9、 is based on the characteristic of the crosswell seismic wavefield,and give the steps of the separation of p-wave and s-wave.Firstly,make filtering techniques on originality multi-component data,and we can eliminate many coherent interference waves.If the originality seismic data is three components

10、,we should transform them to two components using vector synthesis method.Secondly,with the vertical component and horizontal component,we must separate the down-going wave from up-going wave.Thirdly,with the separated down-going wave and up-going wave,we can calculate the polarization angle by cova

11、riance matrix.Fourthly, applying the polarization angle on the wavefield separation matrix,then we can separate the s-wave from p-wave for both up-going wave and down-going wave.Lastly,adding the up-going s-wave and down-going s-wave as the final s-wave,and adding the up-going p-wave and down-going

12、p-wave as the final p-wave.When separating the up-going wave from down-going wave,we use the frequency-space domain-p transform.The using of generalized inverse and the trimming treatment can suppress false frequency and port effect effectively.Using the singular value decomposition method,we can im

13、prove calculation efficiency obviously.We found the covariance matrix to calculate the polarization angle.This technique don't need velocity information,and we should justly enter the the up-going wave and down-going wave of two components.When founding the covariance matrix,the choosing of the

14、time window is very important,and this article introduce two ways to choose the time window,the autocorrelation and passing zero point.Then,we found the wavefield separation matrix,and put the polarization angle in it,which is calculated above.We call this method for polarization rotation,and after

15、this,the separation of p-wave and s-wave is complete.The tests on the stimulant theoretically crosswell seismic data and practicality data denote:in the final p-wave profile,the energy of s-wave is subdued,and the p-wave is reserved,in the final s-wave profile,the great mass of p-wave is eliminated,and the s-wave which is formerly covered by p-wave appears out distinctly.In the article,the method to separate the p-wave from s-wave for crosswell seismic data works effectively.This method ad