CN101609161A - Based on the seismic sequence theory multi-scale data combine frequency band expanding method - Google Patents

Abstract

Based on the seismic sequence theory multi-scale data combine frequency band expanding method, comprise surface seismic poststack data is carried out time frequency analysis, surface seismic poststack data is carried out the seismologic record chopping, obtain first geologic model, first geologic model and corresponding crosshole seismic section are carried out the optimization matching treatment, thereby obtain the matching attribute of crosshole seismic section and surface seismic poststack data, under the constraint of surface seismic layer bit data, matching attribute is carried out space-variant to be handled, obtain the space-variant matching attribute, the space-variant matching attribute is acted on first geologic model obtain first data volume, first data volume is carried out the seismologic record chopping to be handled, obtaining second geologic model, is that constraint condition is carried out broad-band constrained inversion to surface seismic poststack data with second geologic model, obtains second data volume.By method of the present invention, crosshole seismic and surface seismic poststack data can be carried out joint inversion, thereby obtain the high-resolution data body.

Description

Based on the seismic sequence theory multi-scale data combine frequency band expanding method

Technical field

The present invention relates to seismic data overall treatment field, is the disposal route that a kind of use in conjunction by crosshole seismic and surface seismic poststack data improves surface seismic poststack data resolution.

Background technology

Crosshole seismic is owing to avoided the absorption of face of land weathering zone to the seismic signal radio-frequency component, therefore can obtain the seismic signal of high-resolution, the application of cross hole seismic technology in meticulous exploration in oil field and exploitation driven the Comprehensive Geophysics technical progress towards oil reservoir.Only between Liang Jing, but the advantage of three-dimensional surface seismic technology is that its area coverage is wide to crosshole seismic imaging scope, and good lateral resolution is arranged, but vertical resolution also far is not enough.Therefore, surface-seismic data and cross-well seismic data use in conjunction need be had complementary advantages.

Present existing certain methods utilizes crosshole seismic to improve the method for surface seismic poststack data resolution, all be based on Signal Matching filtering thought, all reflection coefficient is supposed, as blue color filtered and sparse spiking deconvolution, all be based upon on the reflection coefficient statistics hypothesis basis, suppose that all reflection coefficient sequence is the white noise random signal, yet the change that seismic sequence stratigraphy is thought the reflection wave frequency characteristic direction reflects the cyclical variation of deposition bedding architecture and sedimentary cycle; Think that reflection coefficient sequence is not traditional white noise random function, but corresponding with geology sequence body, as to have cyclicity distribution series, therefore existing method is the mathematical method and the processing means in signal field fully, does not embody the seismic sequence thinking.

Summary of the invention

The object of the present invention is to provide a kind ofly based on seismic sequence body thought, the joint inversion by crosshole seismic and surface seismic poststack data obtains the method for high-resolution data body.

For this reason, the invention provides a kind ofly, comprising based on the seismic sequence theory multi-scale data combine frequency band expanding method:

Surface seismic poststack data is carried out time frequency analysis;

Described surface seismic poststack data is carried out the seismologic record chopping, obtain first geologic model;

First geologic model and corresponding crosshole seismic section are carried out the optimization matching treatment, thereby obtain the matching attribute of described crosshole seismic section and described surface seismic poststack data;

Under the constraint of surface seismic layer bit data, described matching attribute is carried out space-variant handle, obtain the space-variant matching attribute;

Described space-variant matching attribute is acted on described first geologic model obtain first data volume;

Described first data volume is carried out the seismologic record chopping handle, obtain second geologic model;

With described second geologic model is that constraint condition is carried out broad-band constrained inversion to described surface seismic poststack data, obtains second data volume.

As preferably, describedly described surface seismic poststack data is carried out the seismologic record chopping be specially and select 40% threshold value that described surface seismic poststack data is carried out chopping to handle.

As preferably, describedly described first data volume is carried out the seismologic record chopping handle to be specially and select 10% threshold value that described first data volume is carried out chopping to handle.

By method of the present invention, crosshole seismic and surface seismic poststack data can be carried out joint inversion, thereby obtain the high-resolution data body.

Description of drawings

Fig. 1 is the schematic flow sheet based on the seismic sequence theory multi-scale data combine frequency band expanding method provided by the invention;

Fig. 2 is crosshole seismic section and spectrum diagram;

Fig. 3 is surface seismic poststack data and spectrum diagram;

Fig. 4 is for using frequency band expanding section and the frequency band synoptic diagram that obtains based on the seismic sequence theory multi-scale data combine frequency band expanding method provided by the invention.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the invention, the technical scheme in the embodiment of the invention is clearly and completely described, obviously, described embodiment only is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that is obtained under the creative work prerequisite.

Fig. 1 is the schematic flow sheet based on the seismic sequence theory multi-scale data combine frequency band expanding method provided by the invention, and as shown in Figure 1, this method comprises:

S01: surface seismic poststack data is carried out time frequency analysis;

S02: described surface seismic poststack data is carried out the seismologic record chopping, obtain first geologic model;

S03: first geologic model and corresponding crosshole seismic section are carried out the optimization matching treatment, thereby obtain the matching attribute of described crosshole seismic section and described surface seismic poststack data;

S04: under the constraint of surface seismic layer bit data, described matching attribute is carried out space-variant handle, obtain the space-variant matching attribute;

S05: described space-variant matching attribute is acted on described first geologic model obtain first data volume;

S06: described first data volume is handled through the seismologic record chopping, obtained second geologic model;

S07: with described second geologic model is that constraint condition is carried out broad-band constrained inversion to described surface seismic poststack data, obtains second data volume.

Among this embodiment, at first surface seismic poststack data is carried out time frequency analysis, to determine the possibility of seismic sequence body and band compensation.The time frequency analysis technology can be analyzed the frequency band distribution situation of surface seismic poststack data, and especially the analysis of high frequency section is useful signal or noise.The geology sequence body that earthquake cycle body is corresponding certain, but on seismic section, generally can not find and the corresponding independently earthquake of the inner thin layer of geology sequence body lineups, thereby the relevant information that needs the time frequency analysis technology to extract from seismic data to be comprised in the earthquake cycle body (for example, on the time frequency analysis section, can take place to identify sequence body border and pairing time period of inner thin layer thereof on the tuning frequency road of amplitude, in conjunction with well logging, these sections of analysis such as the drilling well and the geology of appearing can obtain sedimentary cycle of relevant sequence body and sedimentary facies, deposition intermittently, geological information such as reservoir and caprock distribution) in order to explanation of seismic sequence body.By time frequency analysis can also study the frequency spread of surface seismic poststack data energy and each rank, all types of sequence body shows situation and characteristics, determine frequency content to be compensated, and the possibility that realizes compensation be discussed.The key of decision band compensation possibility is the signal to noise ratio (S/N ratio) in the sequence body seismic response frequency band.The faint energy that contains of Bu Chang frequency band is shown as the useful signal energy if desired, and band compensation is feasible so.If the energy in the frequency band that needs in the original ground earthquake poststack data to compensate is a noise, so this data is used for joint inversion will be difficult to obtain effect.For example, during the frequency band spread of time frequency analysis technical Analysis seismic data, the time frequency analysis section of useful signal has clear regularity, and the directivity of frequency change is obvious; Noise does not then have this regularity, and random noise has the white noise spectrum signature, and energy evenly distributes along frequency axis, and frequency change does not have directivity.Like this, by surface seismic poststack data is carried out time frequency analysis, can determine the possibility of seismic sequence body and band compensation.

Then, described surface seismic poststack is carried out the seismologic record chopping, obtain first geologic model.Seismologic record chopping technology (SFMOD) is a kind of method that can obtain 3D geological data tectonic information, the threshold value certain to surface seismic poststack data setting keeps waveform for the part that exceeds threshold value, and shows its section with the display mode of seismologic record.Different threshold values and filtering parameter setting can be drawn different other seismic sections of sequence body level of reflection.Suitably choose threshold value, can replace reflection coefficient sequence with the chopping result of seismologic record, main layer position do not have error, and the amplitude small random is disturbed and can not appeared among the SFMOD result, can obtain the tectonic information of main layer position by SFMOD.For example, when needs obtained stratigraphic structure sequence grade section, selecting threshold value usually was 40%, and in the time of obtaining well ground and unite meticulous effective earthquake geologic model after the coupling, selecting threshold value usually is 10%.Selecting threshold value herein is 40%, to obtain first geologic model.

Then, first geologic model carries out the optimization coupling with corresponding crosshole seismic section, thereby obtains the matching attribute of described crosshole seismic section and described surface seismic poststack data.The coupling of the crosshole seismic and first geologic model realizes under dimension is received the design of optimization operator, known cross-well seismic data is y (t)={ y (0), y (1), ... y (np) }, corresponding surface seismic poststack data is x (t)={ x (o), x (1) ... x (np) }, wherein np is the sampling number in seismologic record road.

The design matching attribute h (t)=(h (and l), h (l+1) ... h (0), h (1) ... h (l) }, make x (t) * h (t) under the error sum of squares least meaning near y (t);

If z (t)=x (t) * h (t),

Then $z\left(i\right)=\underset{\mathrm{\τ}=-l}{\overset{l}{\mathrm{\Σ}}}h\left(\mathrm{\τ}\right)x(i-\mathrm{\τ}),(i=\mathrm{0,1},...\mathrm{np})$

Thereby there is following square-error formula:

$Q=\underset{k=0}{\overset{\mathrm{np}}{\mathrm{\Σ}}}{(z\left(k\right)-y\left(k\right))}^2---\left(1\right)$

Ask and make Q get the h (t) of minimum value:

$\frac{\∂Q}{\∂h\left(j\right)}=0,j=-l,-l+1,...,\mathrm{0,1},...,l---\left(2\right)$

Wherein j represents discrete signal samples point sequence number, thereby has:

$\underset{k=0}{\overset{\mathrm{np}}{\mathrm{\Σ}}}2(\underset{\mathrm{\τ}=-l}{\overset{l}{\mathrm{\Σ}}}h\left(\mathrm{\τ}\right)x(k-\mathrm{\τ})-y\left(k\right))x(k-j)=0---\left(3\right)$

That is:

$\underset{\mathrm{\τ}=-l}{\overset{l}{\mathrm{\Σ}}}h\left(\mathrm{\τ}\right)\underset{k=0}{\overset{\mathrm{np}}{\mathrm{\Σ}}}x(k-\mathrm{\τ})x(k-j)=\underset{k=0}{\overset{\mathrm{np}}{\mathrm{\Σ}}}y\left(k\right)x(k-j),j=-l,-l+1,...,\mathrm{0,1},...,l---\left(4\right)$

Promptly

$\underset{\mathrm{\τ}=-l}{\overset{l}{\mathrm{\Σ}}}h\left(\mathrm{\τ}\right)\mathrm{Rxx}(\mathrm{\τ}-j)=\mathrm{Ryx}\left(j\right),j=-l,-l+1,...,\mathrm{0,1},...,l---\left(5\right)$

Wherein (τ-j) is the auto-correlation of surface seismic data to Rxx, and Ryx is the simple crosscorrelation of surface seismic and crosshole seismic.(5) formula is found the solution, can be obtained the matching attribute h (t) of the surface seismic poststack data after crosshole seismic section and the chopping.

Then, under the constraint of surface seismic layer bit data, described matching attribute is carried out space-variant handle, obtain the space-variant matching attribute.

Utilize the interpolation of seismic interpretation layer position constraint matching attribute on three dimensions.If layer position top circle's time is st, circle, the end time is et, calculates the cross-correlation coefficient of outer each ground, road geological data of crosshole seismic section and ground, road, crosshole seismic section place geological data:

$\mathrm{cor}\left[i\right]=\frac{\underset{k=\mathrm{st}}{\overset{\mathrm{et}}{\mathrm{\Σ}}}{S}_1\left[k\right]*S_2\left[k\right]}{\sqrt{\underset{k=\mathrm{st}}{\overset{\mathrm{et}}{\mathrm{\Σ}}}{S_1}^2\left(k\right)*\underset{k=\mathrm{st}}{\overset{\mathrm{et}}{\mathrm{\Σ}}}{S_2}^2\left(k\right)}}---\left(6\right)$

Wherein, S ₁The surface seismic data in expression road, crosshole seismic section place, S ₂Outer certain ground, the road geological data of expression crosshole seismic section, cor represents their related coefficient.

If related coefficient is 1, then read corresponding line Taoist monastic name, in crosshole seismic section matching attribute storehouse, inquire about, obtain the matching attribute in its corresponding road.Thereby each road is carried out the matching attribute that same operation obtains all relevant roads on the crosshole seismic section.

For the seismic trace beyond the crosshole seismic section, its related coefficient is bound to less than 1, and its matching attribute must be subjected to the influence of distance and related coefficient, and its matching attribute can be asked for following formula:

$e\left[j\right]=\underset{i}{\mathrm{\Σ}}\mathrm{oper}\left[i\right]\left[j\right]*\frac{\mathrm{cor}\left[i\right]}{r^2}/\underset{t}{\mathrm{\Σ}}\frac{\mathrm{cor}\left[i\right]}{r^2},j=\mathrm{1,2,3},...\mathrm{np}---\left(7\right)$

E[j] the matching attribute of seismic trace beyond the expression crosshole seismic section, wherein i represents the seismic trace got, j represents the sampling spot on the sympathetic earthquakes road, oper[i] a certain matching attribute in the expression matching attribute storehouse, cor[i] the corresponding with it related coefficient of expression, r represents the distance between two seismic traces.

Because the amplitude of matching attribute is less than normal, the existing certain similarity of the matching attribute of diverse location is also variant, can not reach needed numerical value requirement in the calculating afterwards.The most reasonable in order to obtain, science numerical value, therefore need carry out the normalized of data, thereby obtain the space-variant matching attribute:

$\mathrm{oper}\left[j\right]=\mathrm{oper}\left[j\right]/\mathrm{sqrt}\left(\underset{\mathrm{np}}{\overset{i=0}{\mathrm{\Σ}}}\mathrm{oper}\right[i]*\mathrm{oper}[i\left]\right)---\left(8\right)$

Oper[j wherein] a space-variant matching attribute in the expression space-variant matching attribute storehouse.

Subsequently, described space-variant matching attribute is acted on described first geologic model and obtain first data volume.This first data volume obtains by first geologic model and described space-variant matching attribute are mated.Its principle is the convolution theory of Lu Binxun (Robinson), according to the definition of the convolution of Robinson:

S(t)＝W(t)*R(t)；

S (t)-seismologic record;

W (t)-seismic wavelet;

The reflection coefficient of R (t)-seismic section;

If s[t] expression input signal r[t] and the convolution of ω [t]

$S\left[t\right]=\underset{-\∞}{\overset{\∞}{\∫}}r(t-\mathrm{\τ})\mathrm{\ω}\left(t\right)\mathrm{d\τ}---\left(9\right)$

S[t wherein] be matching result, r[t] expression first geologic model, the space-variant matching attribute that ω [t] extracts for the front.

Then, the described first data volume chopping is handled, obtained second geologic model.With described second geologic model is that constraint condition is carried out broad-band constrained inversion to described surface seismic poststack data, obtains second data volume.Utilize the SFMOD technology that first data volume is done the chopping processing and can obtain i.e. second geologic model of meticulous effective earthquake geologic model., in the SFMOD module, select 10% herein, make the waveform in the data volume of winning become pulse signal, can obtain meticulousr tectonic information, can tell some little structures as threshold value.This meticulous effective second geologic model is a constraint condition required in the broad-band constrained inversion.

The principle that broad-band constrained inversion improves seismic data resolution is: utilize the effective high frequency composition in the high-frequency information guiding seismologic record of constraint condition, make it to be strengthened to a certain degree, and according to constraint condition, make the various radio-frequency components in the seismologic record draw close to high resolving power constraint condition as far as possible, reach the purpose that improves resolution.Broad-band constrained inversion can neither blindly improve the energy of all radio-frequency components, does not also blindly widen the frequency band of seismic data, but depends on constraint condition, improves resolution according to the quality of earthquake minute book body.The constraint condition of Shi Yonging has very high signal to noise ratio (S/N ratio) simultaneously, useful signal in refutation process in the seismologic record because of with good being strengthened of correlativity of constraint condition, noise is because of weakened with the correlativity difference of constraint condition, thereby signal to noise ratio (S/N ratio) also improves when improving resolution.If required model (inversion result) is

The known constraints condition is

Wishing to retrain inverting is satisfying under the prerequisite of seismologic record to constraint condition deflection.

To constraint condition

Biasing then has:

${\stackrel{\→}{m}}_c={[G^{T}G+{\mathrm{\α}}^{2}I]}^{-1}(G^T\stackrel{\→}{d}+{\mathrm{\β}}^2\stackrel{\→}{h})$

Inversion result is a reflection coefficient sequence

Select the high resolving power wavelet, form wavelet matrix G ^*

Then the high-resolution seismic exploration section is ${\stackrel{\→}{d}}^{*}=G^{*}{\stackrel{\→}{m}}_c$

Here be that constraint condition is for being necessary for reflection coefficient sequence.

Fig. 2 is crosshole seismic section and spectrum diagram, and Fig. 3 is surface seismic poststack data and spectrum diagram, and Fig. 4 is for using frequency band expanding section and the frequency band synoptic diagram that obtains based on the theoretical multi-scale data combine frequency band expanding method of earthquake preface body provided by the invention.

From the spectrum analysis result of Fig. 3 and Fig. 4 as can be seen, original ground earthquake poststack data dominant frequency is 40hz, and dominant frequency band is 20-100hz, dominant frequency is 60hz behind the combine frequency band expanding that obtains by method of the present invention, dominant frequency band reaches 30-140hz, and dominant frequency obviously improves, and frequency band is obviously widened.Equally, seismic section from Fig. 3 and Fig. 4, the continuity of reflection wave and stratum are in contact with one another relation obvious improvement, section is clearer, and breakpoint is more clear, and resolution improves, resolution near the crosshole seismic of Fig. 2, can tell thickness and be 7～10 meters stratum, and original ground earthquake poststack section can only be told the stratum more than 15 meters, illustrates by of the present invention the associating opening up the resolution that frequency method has improved seismic data really.

In a word, the above is embodiments of the invention only, is not to be used to limit protection scope of the present invention, but is used to illustrate the present invention.Within the spirit and principles in the present invention all, any modification of being done, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (3)

1, based on the seismic sequence theory multi-scale data combine frequency band expanding method, it is characterized in that, comprising:

Surface seismic poststack data is carried out time frequency analysis;

Described surface seismic poststack data is carried out the seismologic record chopping, obtain first geologic model;

First geologic model and corresponding crosshole seismic section are carried out the optimization matching treatment, thereby obtain the matching attribute of described crosshole seismic section and described surface seismic poststack data;

Under the constraint of surface seismic layer bit data, described matching attribute is carried out space-variant handle, obtain the space-variant matching attribute;

Described space-variant matching attribute is acted on described first geologic model obtain first data volume;

Described first data volume is carried out the seismologic record chopping handle, obtain second geologic model;

With described second geologic model is that constraint condition is carried out broad-band constrained inversion to described surface seismic poststack data, obtains second data volume.

2, the method for claim 1 is characterized in that, describedly described surface seismic poststack data is carried out the seismologic record chopping is specially and selects 40% threshold value that described surface seismic poststack data is carried out chopping to handle.

3, the method for claim 1 is characterized in that, describedly described first data volume is carried out the seismologic record chopping handles to be specially and selects 10% threshold value that described first data volume is carried out chopping to handle.

Images