CN103278849A - Method and system for performing wavelet estimation on the basis of seismic data and logging information - Google Patents

Abstract

The invention provides a method and a system for performing wavelet estimation on the basis of seismic data and logging information. The method comprises the following steps: acquiring the seismic data and the logging information of a current stratum; estimating a minimum phase wavelet according to the seismic data; performing 0-1 coding according to the minimum phase wavelet to obtain various coded data; extracting well bypass data according to the seismic data; determining a reflection coefficient according to the logging information; determining a well control evaluation model according to the well bypass data, the reflection coefficient and a convolution model; estimating a best wavelet according to the well control evaluation model and the coded data; and performing reservoir prediction according to the best wavelet to obtain a reservoir prediction result of the current stratum. The wavelet phase problem is handled by adopting genetic manipulation and combining well control as the evaluation criteria, and therefore, the calculation amount is effectively reduced, the convergence rate of root transformation wavelet phase spectrum estimation is accelerated, and the wavelet estimation accuracy is improved.

Description

Carry out wavelet estimation approach and system based on seismic data and well-log information

Technical field

The invention relates to the geophysical prospecting for oil technical field, the reservoir prediction technique about estimating based on wavelet particularly is a kind ofly to carry out wavelet estimation approach and system based on seismic data and well-log information concretely.

Background technology

In the geophysical prospecting for oil technical field, amplitude changes (Amplitude versus angle with angle, AVA) inverting is to utilize the reflection seismic wave-amplitude with the underground lithologic parameter of Changing Pattern inverting of incident angle, and then the character of estimation reservoir rock and fluid and carry out the technology of petroleum-gas prediction.The pre-stack seismic data often utilizes amplitude with variation (the Amplitude Versus Offset of offset distance, AVO) stratum three parameters (velocity of longitudinal wave, shear wave velocity and density) are carried out inverting, thereby obtain for the basic lithological information of velocity of longitudinal wave, shear wave velocity and the density etc. of identification hydrocarbon-bearing formation.

Inversion technique of the prior art (such as above-mentioned two kinds) when carrying out reservoir prediction, all needs to be applied to convolution model, and the seismic wavelet in the convolution model accurately whether, will directly affect the accuracy of follow-up reservoir prediction.In actual seismic wavelet estimation, because the influence of factors such as noise, the phase spectrum of wavelet is difficult to accurately estimate.Carry out deconvolution based on inaccurate phase place wavelet, its result's resolution will be subjected to restriction to a certain degree.

In the prior art, mainly contain following several mode about the estimation of seismic wavelet: what people such as (1), Ziolkowski proposed carries out the method for transform to geological data; What (2), Lindsey proposed carries out the method for transform to the auto-correlation of geological data; (3), the method that proposes of Ursin is that a series of expansion Yule-Walker system of equations are found the solution, and carries out the mixed-phase inverse filter that the best is estimated in the transform rooting by the maximum phase component to mixed phase wavelet only; (4), utilize enumerative technique, Li Lihong and Ma Zhaogui to propose the root of minimum phase wavelet transform is scanned to estimate best wavelet; (5), the Chang Xiangwei rotation wavelet estimation technique; (6), root conversion wavelet is estimated mode.

The technological deficiency that above-mentioned several wavelet methods of estimation mainly exist in the prior art is: the wavelet estimated accuracy of (1), Ziolkowski method is not high, and is vulnerable to The noise; (2), the Lindsey method utilizes auto-correlation information to estimate wavelet, is difficult to control the phase spectrum of wavelet; (3), the precision of Ursin method and stability are subjected to the influence of window length and operator length easily; (4), the enumerative technique calculated amount is huge, is difficult to be applied to real data; (5), the Chang Xiangwei rotation wavelet estimation technique can realize that its calculated amount is generally little by the mode of phasescan, but application conditions is relatively harsher; (6) though, root conversion wavelet estimates the phase condition of the more realistic wavelet of mode, calculated amount is huge, is difficult to be applied to during real data handles.

Therefore, for seismic exploration work area practically, how accurately to estimate seismic wavelet according to different geological datas, and then the multiple underground lithologic parameter of inverting, carrying out petroleum-gas prediction is the technical barrier that this area needs to be resolved hurrily.

Summary of the invention

In order to overcome the problems referred to above that prior art exists, the invention provides and a kind ofly carry out wavelet estimation approach and system based on seismic data and well-log information, by wavelet root conversion process is carried out specific coding and decoding, and it is tactful to introduce heredity, adopt genetic manipulation and handle the sub-wave phase problem in conjunction with well control as interpretational criteria, reduce calculated amount effectively, accelerated the speed of convergence that root varitron wave phase spectrum is estimated, improved the wavelet estimated accuracy.

One of purpose of the present invention is, provides a kind of and carries out the wavelet estimation approach based on seismic data and well-log information, comprising: seismic data, the well-log information of gathering current stratum; Estimate minimum phase wavelet according to described seismic data; Carry out the 0-1 coding according to described minimum phase wavelet, obtain multiple coded data; Extract well lie data according to described seismic data; Determine reflection coefficient according to described well-log information; Determine the well control evaluation model according to described well lie data, reflection coefficient and described convolution model; Estimate best wavelet according to described well control evaluation model and described multiple coded data; Carry out reservoir prediction according to described best wavelet, obtain the reservoir prediction result on current stratum.

One of purpose of the present invention is, provides a kind of and carried out the system that wavelet is estimated based on seismic data and well-log information, and comprising: harvester is used for gathering seismic data, the well-log information on current stratum; Minimum phase wavelet is determined device, is used for estimating minimum phase wavelet according to described seismic data; Code device is used for carrying out the 0-1 coding according to described minimum phase wavelet, obtains multiple coded data; Well lie data extract device is used for extracting well lie data according to described seismic data; Reflection coefficient is determined device, is used for determining reflection coefficient according to described well-log information; The well control evaluation model is determined device, is used for determining the well control evaluation model according to described well lie data, reflection coefficient and described convolution model; Best wavelet estimation unit is used for estimating best wavelet according to described well control evaluation model and described multiple coded data; The reservoir prediction device is used for carrying out reservoir prediction according to described best wavelet, obtains the reservoir prediction result on current stratum.

Beneficial effect of the present invention is, by wavelet root conversion process is carried out specific coding and decoding, and it is tactful to introduce heredity, adopt genetic manipulation and handle the sub-wave phase problem in conjunction with well control as interpretational criteria, reduce calculated amount effectively, accelerated the speed of convergence that root varitron wave phase spectrum is estimated, improved the wavelet estimated accuracy, when having solved of the prior art converter technique and asking for mixed phase wavelet, sub-wave phase there is the technical matters of dependence to a certain degree.

For above-mentioned and other purposes, feature and advantage of the present invention can be become apparent, preferred embodiment cited below particularly, and cooperate appended graphicly, be described in detail below.

Description of drawings

In order to be illustrated more clearly in the embodiment of the invention or technical scheme of the prior art, to do to introduce simply to the accompanying drawing of required use in embodiment or the description of the Prior Art below, apparently, accompanying drawing in describing below only is some embodiments of the present invention, for those of ordinary skills, under the prerequisite of not paying creative work, can also obtain other accompanying drawing according to these accompanying drawings.

Among Fig. 1, (a) for there not being initial wavelet composite traces figure under the situation of making an uproar, (b) for there not being raw readings figure under the situation of making an uproar, (c) for there not being the wavelet composite traces figure that the present invention estimates under the situation of making an uproar;

Fig. 2 is not for there being the comparison diagram that initial wavelet under the situation of making an uproar, true wavelet and the present invention estimate wavelet;

Among Fig. 3, be initial wavelet composite traces figure under the noisy situation (a), (b) be raw readings figure under the noisy situation, (c) estimate wavelet composite traces figure for the present invention under the noisy situation;

Fig. 4 estimates the comparison diagram of wavelet for initial wavelet, true wavelet and the present invention under the noisy situation;

Among Fig. 5, (a) being Chang Xiangwei wavelet composite traces figure, (b) is well lie real data figure, (c) the wavelet composite traces figure that estimates for the present invention;

The wavelet figure that Fig. 6 (a) estimates for the present invention;

The wavelet figure that Fig. 6 (b) estimates for the Chang Xiangwei rotation;

Fig. 7 (a) is original earthquake source map;

Fig. 7 (b) is the determinacy deconvolution of estimating wavelet based on the present invention figure as a result;

Fig. 7 (c) is Chang Xiangwei wavelet deconvolution figure as a result;

A kind of process flow diagram that carries out the wavelet estimation approach based on seismic data and well-log information that Fig. 8 provides for the embodiment of the invention;

Fig. 9 is the particular flow sheet of the step S103 among Fig. 8;

Figure 10 is the particular flow sheet of the step S107 among Fig. 8;

Figure 11 is the particular flow sheet of the step S303 among Figure 10;

A kind of structured flowchart that carries out the system that wavelet estimates based on seismic data and well-log information that Figure 12 provides for the embodiment of the invention;

A kind of concrete structure block diagram that carries out code device in the system that wavelet estimates based on seismic data and well-log information that Figure 13 provides for the embodiment of the invention;

A kind of concrete structure block diagram that carries out best wavelet estimation unit in the system that wavelet estimates based on seismic data and well-log information that Figure 14 provides for the embodiment of the invention;

A kind of concrete structure block diagram that carries out decoder module in the system that wavelet estimates based on seismic data and well-log information that Figure 15 provides for the embodiment of the invention;

A kind of concrete structure figure that carries out evaluation module in the system that wavelet estimates based on seismic data and well-log information that Figure 16 provides for the embodiment of the invention;

Figure 17 is the synoptic diagram that minimum phase wavelet is carried out the 0-1 Code And Decode.

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 obtains under the creative work prerequisite.

The present invention proposes and a kind ofly carry out wavelet estimation approach and system based on seismic data and well-log information, by wavelet root conversion process is carried out specific coding and decoding, and introducing genetic algorithm, adopt the heredity strategy and handle the sub-wave phase problem in conjunction with well control as interpretational criteria, reduced calculated amount effectively, accelerate the speed of convergence that root varitron wave phase spectrum is estimated, improved the wavelet estimated accuracy, reached the purpose that reduces the prediction of oil-gas reserve risk.

Fig. 8 a kind ofly carries out the process flow diagram of wavelet estimation approach based on seismic data and well-log information for what the embodiment of the invention provided, and as shown in Figure 8, this method specifically comprises:

S101: seismic data, the well-log information of gathering current stratum.

In oil-gas exploration, kept abundant information in seismic data, the well-log information, how utilizing these information to obtain oil gas well parameter is one of of paramount importance work of seismic prospecting for oil and gas detection and identification provide data to support.

S102: estimate minimum phase wavelet according to described seismic data.This step can realize by prior art, such as spiking deconvolution.

S103: carry out the 0-1 coding according to described minimum phase wavelet, obtain multiple coded data.

Fig. 9 is the particular flow sheet of the step S103 among Fig. 8, and as shown in Figure 9, this step specifically comprises:

S201: described minimum phase wavelet is carried out transform, obtain the polynomial expression about variable z.

A minimum phase wavelet that estimates by certainty sub wave method of estimation or statistical wavelet method of estimation is expressed as w (t)=(w ₀, w ₁..., w _n), n is the sampling number of seismic wavelet, wherein w ₀w _n≠ 0.This wavelet sequence is carried out transform obtains one only about the polynomial expression of variable z, be shown below:

Z(w(t))=w _n(z-α ₁)(z-α ₂)…(z-α _i)…(z-α _n) (1)

S202: find the solution described polynomial expression to determine the root of described polynomial expression correspondence.Make (1) formula equal 0, adopting polynomial expression to decompose and can obtaining this root of polynomial is (α ₁, α ₂..., α _i..., α _n).

S203: the root to described polynomial expression correspondence carries out the 0-1 coding, obtains multiple coded data.In order further to introduce cataloged procedure of the present invention, below summary coding principle of the present invention earlier.

Suppose that seismic data can be by Robinson(Lu Binxun) convolution model describes, then has

s(t)=w(t)*r(t)=Z ^-1[Z(w(t))]*r(t) (2)

=w _n(-α ₁,1)*(-α ₂,1)*…*(-α _i,1)*…*(-α _n,1)*(r ₀,r ₁,…,r _m)

Wherein, t represents the time, and s (t) represents seismologic record, w (t)=(w ₀, w ₁..., w _n) the expression seismic wavelet, r (t)=(r ₀, r ₁..., r _m) the expression reflection coefficient, m is the sampling number of reflection coefficient, Z ^-1The contrary transform of expression, (α ₁, α ₂..., α _i..., α _n) root of expression seismic wavelet transform, (α _i, 1) expression w (t) 2 wavelets.Say that broadly most of earthquake deconvolution or wave impedance inversion method realize by formula (2) just.

2 wavelet (α to w (t) _i, 1) and carry out Fourier transform, and make α _i=x _i+ jy _i, have

F[(-α _i,1)]=-α _i+e ^-jwt=-(x _i+jy _i)+cos(wt)-jsin(wt) (3)

=[cos(wt)-x _i]-j[sin(wt)+y _i]

Wherein, Expression imaginary unit, w represents angular frequency.Formula (3) is got spectral amplitude and phase spectrum respectively, have

$A\left[({-\mathrm{\&alpha;}}_i,1)\right]=\sqrt{{[\cos \left(\mathrm{wt}\right)-x_i]}^2+{[\sin \left(\mathrm{wt}\right)+y_i]}^2}$ (4) $={[1+x_{\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="DEST\_PATH\_HDA00003477171400071.png,DEST\_PATH\_HDA00003477171400081.png"\; state="\{\{state\}\}">i</figure-callout>}}^2+y_i^2-{2x}_i\cos \left(\mathrm{wt}\right)+{2y}_i\sin \left(\mathrm{wt}\right)]}^{1/2}$

$P\left[(-{\mathrm{\&alpha;}}_i,1)\right]=-\arctan \left[\frac{\sin \left(\mathrm{wt}\right)+y_i}{\cos \left(\mathrm{wt}\right)-x_i}\right]---\left(5\right)$

Because w (t)=w _n(α ₁, 1) and * (α ₂, 1) and * ... * (α _i, 1) and * ... * (α _n, 1), so the spectral amplitude of wavelet and phase spectrum can be expressed as respectively

$A\left[w\left(t\right)\right]=\underset{i=1}{\overset{n}{\mathrm{\&Pi;}}}A\left[({-\mathrm{\&alpha;}}_i,1)\right]=\underset{i=1}{\overset{n}{\mathrm{\&Pi;}}}[1+x_i^2+y_i^2-{2x}_i\cos \left(\mathrm{wt}\right)+{2y}_i\sin \left(\mathrm{wt}\right){]}^{1/2}---\left(6\right)$

$P\left[w\left(t\right)\right]=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}P\left[({-\mathrm{\&alpha;}}_i,1)\right]=-\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\arctan \left[\frac{\sin \left(\mathrm{wt}\right)+y_i}{\cos \left(\mathrm{wt}\right)-x_i}\right]---\left(7\right)$

2 wavelet (α as w (t) _i, 1) and quilt

After the replacement, spectral amplitude does not change, this be because

$A\left[\right|{\stackrel{\&OverBar;}{\mathrm{\&alpha;}}}_i\left|(-\frac{1}{{\stackrel{\&OverBar;}{\mathrm{\&alpha;}}}_i},1)\right]=\sqrt{{\mathrm{<figure-callout\; id="2"\; label="x\; i"\; filenames="DEST\_PATH\_HDA00003477171400071.png,DEST\_PATH\_HDA00003477171400081.png"\; state="\{\{state\}\}">x</figure-callout>}}_i^{}+y_i^2}\sqrt{{[\cos \left(\mathrm{wt}\right)-\frac{x_i}{x_i^2+y_i^2}]}^2+{[\sin \left(\mathrm{wt}\right)+\frac{y_i}{x_i^2+y_i^2}]}^2}---\left(8\right)$ (8) $=[1+x_{\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="DEST\_PATH\_HDA00003477171400071.png,DEST\_PATH\_HDA00003477171400081.png"\; state="\{\{state\}\}">i</figure-callout>}}^2+y_i^2-{2x}_i\cos \left(\mathrm{wt}\right)+{2y}_i\sin \left(\mathrm{wt}\right){]}^{1/2}=A[({-\mathrm{\&alpha;}}_i,1)]$

Wherein, x _iy _i≠ 0.2 wavelets

Phase spectrum can be expressed as

$P\left[\right|{\stackrel{\&OverBar;}{\mathrm{\&alpha;}}}_i\left|(-\frac{1}{{\stackrel{\&OverBar;}{\mathrm{\&alpha;}}}_i},1)\right]=-\mathrm{arctyan}\left[\frac{\sin \left(\mathrm{wt}\right)+\frac{y_i}{x_i^2+y_i^2}}{\cos \left(\mathrm{wt}\right)-\frac{x_i}{x_i^2+y_i^2}}\right]---\left(9\right)$

Significantly, when

Namely

The time,

(α _i, 1) phase spectrum just the same.At this moment, (α _i, 1) and

Represent same binary wavelet in fact.When

The time,

Therefore, by a conversion

Binary wavelet (α _i, 1) be

Just can scan the out of phase music score ripple of all equal length, same-amplitude spectrum.Note working as α _iBe plural number, i.e. y _i≠ 0 o'clock, its conjugation compound radical needed conversion simultaneously.

Hence one can see that, by moving part or all

Root α _i(i=1,2 ..., n) be , can scan the seismic wavelet of series of identical length, same-amplitude spectrum, out of phase spectrum.In essence, root α _iBe moved into

It is a man-to-man process.Root is outside unit circle

Move to a certain definite position in the unit circle, or in unit circle Move to the outer a certain definite position of unit circle.

Binary wavelet (the α of conversion w (t) _i, 1) be

After, the spectral amplitude of wavelet does not change, and just change has taken place phase spectrum, so the root process that move some fixed positions outside circle in the scanning of sub-wave phase some fixed positions that is the outer root of circle in the circle or the circle.Based on this feature, the present invention proposes a kind of 0-1 coding method and finish this root mode that moves.

Figure 17 is the synoptic diagram that minimum phase wavelet is carried out the 0-1 Code And Decode.As shown in Figure 17, chromosome is made of 0 or 1 of some, and chromosomal front end is represented the real number gene, only shines upon real root; Tail end is represented plural gene, only shines upon compound radical.The root that not mobile this gene position of 0 expression is shone upon, the root that mobile this gene position of 1 expression is shone upon.Chromosomal length is only relevant with the initial wavelet w (t) that estimates, after w (t) determined, chromosomal length was also fixing thereupon.In addition, plural gene string is also being controlled to determinacy another plural gene string (the present invention is referred to as controlled plural gene string).In controlled plural gene string, 0 ' and 1 ' represents the conjugation compound radical of the root that mobile and mobile corresponding plural gene position is shone upon respectively.Such coded system just can make us directly utilize selection, intersection and the variation of genetic algorithm to upgrade the phase spectrum of wavelet.

According to above-mentioned cataloged procedure, to minimum phase wavelet w (t)=(w ₀, w ₁..., w _n) carry out 0-1 coding after, can form 2 ⁿPlant coded data.

As shown in Figure 8, provided by the inventionly carry out the wavelet estimation approach based on seismic data and well-log information and also comprise:

S104: extract well lie data according to described seismic data.

S105: determine reflection coefficient according to described well-log information.This step is at first extracted interval transit time curve and densimetric curve from well-log information, calculate reflection coefficient according to interval transit time curve and densimetric curve then.

S106: determine the well control evaluation model according to described well lie data, reflection coefficient and described convolution model.Estimate the quality between the individuality, need weigh by the size of fitness value.The present invention as control criterion, adopts the related coefficient of theogram w (t) * r (t) and well lie seismologic record as the well control evaluation model, namely with well

O=max[xcorr(w(t)*r(t),s(t))] (10)

Wherein, max[] represent to get maximal value, s (t) is well lie data, the seismic wavelet that w (t) obtains for decoding, r (t) is the reflection coefficient of determining according to well-log information, and xcorr represents the related coefficient between composite traces w (t) * r (t) and the well lie record s (t), and O is objective function.

S107: estimate best wavelet according to described well control evaluation model and described multiple coded data.Figure 10 is the particular flow sheet of step S107, and as shown in Figure 10, this step specifically comprises:

S301: the coded data of from described multiple coded data, choosing predetermined number.Because the operand of genetic algorithm is a colony, the coded data individuality (chromosome) that therefore produces predetermined number is at random formed population.The scale of population is exactly number individual in the population, generally is taken as even number.Each individuality represents an initial wavelet, and population represents an initial wavelet collection.Concentrate in initial wavelet, each wavelet has identical length, identical spectral amplitude and different phase spectrums.In concrete embodiment, predetermined number can preestablish, and is set to even number in the 20-100 scope such as predetermined number.

S302: the coded data to described predetermined number is decoded, and obtains the wavelet sequence of predetermined number.This step specifically comprises:

S401: described coded data is carried out the 0-1 decoding, obtain corresponding root.Further introduce decode procedure of the present invention below.According to described binary wavelet transform, the 0-1 of coded data genetic algorithm coding is mapped to described wavelet root of polynomial.

S402: according to the described polynomial expression of asking for about variable z.Namely according to root (α ₁, α ₂..., α _i..., α _n) obtain the polynomial expression about variable z, the polynomial expression that solves such as formula (1).

S403: the polynomial expression to described variable z carries out contrary transform, obtains the wavelet sequence of described coded data correspondence.Polynomial expression about variable z is carried out contrary transform, and the wavelet sequence that obtains is w (t)=(w ₀, w ₁..., w _n), n is the sampling number of seismic wavelet, wherein w ₀w _n≠ 0.

The decoding that is basic genetic algorithmic is with { 0,1} binary string is transformed into decimal number, and then linear transformation becomes a real number again.And decoding of the present invention is with { 0,1} binary string is transformed into n root (comprising compound radical and real root), obtains a sub-wave train by asking polynomial expression and getting coefficient then.Suppose that certain individuality after decoded is

Then Xie Ma individuality has been endowed clear physical meaning.

As shown in Figure 10, the S107 step also comprises:

S303: judge in the wavelet sequence of described predetermined number whether have best wavelet sequence according to described well control evaluation model.Figure 11 is the particular flow sheet of the step S303 among Figure 10, and as shown in Figure 11, step S303 is specifically by comprising:

S401: obtain predefined well control threshold value coefficient.In concrete embodiment, can preestablish well control threshold value coefficient, as 80%.

S402: described wavelet sequence is carried out convolution with corresponding reflection coefficient, form theogram;

S403: the related coefficient of determining described theogram and well lie record according to described well control evaluation model;

S404: judge that whether described related coefficient is more than or equal to described well control threshold value coefficient.

As shown in Figure 10, the S107 step also comprises:

S304: when being judged as when being, exporting described best wavelet sequence, also is that the wavelet sequence of related coefficient maximum is best wavelet sequence.

S305: when being judged as not, upgrade the phase place of described wavelet.Return execution in step S301 according to the sub-wave phase after upgrading, choose the coded data of new a collection of predetermined number, re-execute step S302 to S303, to choose best wavelet.

Also be that this paper utilizes the scale-of-two genetic algorithm to upgrade the phase place of wavelet, and obtain best wavelet estimation by making theogram and seismic trace near well reach optimum matching.

When upgrading the phase place of wavelet, can adopt genetic manipulation to upgrade sub-wave phase, genetic manipulation generally comprises selection, intersects and variation, and its fundamental purpose is that individuality is transformed.In the present invention, their effect is the phase place of upgrading wavelet.

Wherein " selection " is to select good wavelet in order to concentrate from current wavelet, makes them have an opportunity to be breeding descendants of future generation as parent.The ideal adaptation degree is more high, and its selecteed chance is just more many.Selection is the key of genetic algorithm, and it has embodied the thought of the occurring in nature survival of the fittest.

" intersection " replaced the part-structure of two father's individualities reorganization and generated new individual operation.The purpose of this operation is in order to produce new individuality the next generation, and just as the marriage process of human society, by interlace operation, the ability of genetic algorithm search equal length, same-amplitude spectrum, out of phase music score ripple will be greatly improved." intersection " is the most important means that genetic algorithm is obtained best wavelet and reflection coefficient.

" variation " is that the gene position of individuality is carried out inversion operation.By mutation operation, can make genetic algorithm keep the diversity of population, to prevent non-ripe convergence.Generally speaking, the variation probability is all obtained less, sometimes even get 0.

Select, intersect and make a variation a lot of diverse ways are respectively arranged.At particular problem of the present invention, in embodiment, can choose roulette back-and-forth method, the even bracketing method of scale-of-two and scale-of-two alternative method respectively.

What as shown in Figure 8, the embodiment of the invention provided a kind ofly carries out the wavelet estimation approach based on seismic data and well-log information and also comprises:

S108: carry out reservoir prediction according to described best wavelet, obtain the reservoir prediction result on current stratum.

When carrying out reservoir prediction, all need to be applied to convolution model, and the seismic wavelet in the convolution model accurately whether, will directly affect the accuracy of follow-up reservoir prediction.In actual seismic wavelet estimation, because the influence of factors such as noise, the phase spectrum of wavelet is difficult to accurately estimate.Carry out deconvolution based on inaccurate phase place wavelet, its result's resolution will be subjected to restriction to a certain degree.Such as utilizing amplitude to change (Amplitude versus angle with angle, AVA) petroleum-gas prediction is carried out in inverting, (Amplitude Versus Offset AVO) carries out the identification of hydrocarbon-bearing formation to stratum three parameters (velocity of longitudinal wave, shear wave velocity and density) such as the variation that utilizes amplitude with offset distance.Therefore, the method that the present invention proposes is directly the wavelet of estimating to be carried out the transform rooting, scans the phase place of wavelet then by the scale-of-two genetic manipulation, and then the best wavelet of evolving out rapidly, improve the wavelet estimated accuracy, and then improved the precision of reservoir prediction.

Figure 12 a kind ofly carries out the structured flowchart of the system that wavelet estimates based on seismic data and well-log information for what the embodiment of the invention provided, as shown in Figure 12, shown in system specifically comprise:

Harvester 100 is used for gathering seismic data, the well-log information on current stratum.In oil-gas exploration, kept abundant information in seismic data, the well-log information, how utilizing these information to obtain oil gas well parameter is one of of paramount importance work of seismic prospecting for oil and gas detection and identification provide data to support.

Minimum phase wavelet is determined device 200, is used for estimating minimum phase wavelet according to described seismic data.This device can be realized by prior art, such as spiking deconvolution.

Code device 300 is used for carrying out the 0-1 coding according to described minimum phase wavelet, obtains multiple coded data.Figure 13 is the concrete structure block diagram of code device 300, and as shown in Figure 13, code device specifically comprises:

Transform unit 301 is used for described minimum phase wavelet is carried out transform, obtains the polynomial expression about variable z.

A minimum phase wavelet that estimates by certainty sub wave method of estimation or statistical wavelet method of estimation is expressed as w (t)=(w ₀, w ₁..., w _n), n is the sampling number of seismic wavelet, wherein w ₀w _n≠ 0.This wavelet sequence is carried out transform obtain one only about the polynomial expression of variable z, as shown in Equation (1).

Z(w(t))=w _n(z-α ₁)(z-α ₂)…(z-α _i)…(z-α _n) (1)

Polynomial solving unit 302 is used for finding the solution described polynomial expression to determine the root of described polynomial expression correspondence.Make described (1) formula equal 0, adopting polynomial expression to decompose and can obtaining this root of polynomial is (α ₁, α ₂..., α _i..., α _n).

Coding unit 303 is used for the root of described polynomial expression correspondence is carried out the 0-1 coding, obtains multiple coded data.Binary wavelet (the α of conversion w (t) _i, 1) be

After, the spectral amplitude of wavelet does not change, and just change has taken place phase spectrum, so the root process that move some fixed positions outside circle in the scanning of sub-wave phase some fixed positions that is the outer root of circle in the circle or the circle.Based on this feature, the present invention proposes a kind of 0-1 coding method and finish this root mode that moves.

Figure 17 is the synoptic diagram that minimum phase wavelet is carried out the 0-1 Code And Decode.As shown in Figure 17, chromosome is made of 0 or 1 of some, and chromosomal front end is represented the real number gene, only shines upon real root; Tail end is represented plural gene, only shines upon compound radical.The root that not mobile this gene position of 0 expression is shone upon, the root that mobile this gene position of 1 expression is shone upon.Chromosomal length is only relevant with the initial wavelet w (t) that estimates, after w (t) determined, chromosomal length was also fixing thereupon.In addition, plural gene string is also being controlled to determinacy another plural gene string (the present invention is referred to as controlled plural gene string).In controlled plural gene string, 0 ' and 1 ' represents the conjugation compound radical of the root that mobile and mobile corresponding plural gene position is shone upon respectively.Such coded system just can make us directly utilize selection, intersection and the variation of genetic algorithm to upgrade the phase spectrum of wavelet.

According to above-mentioned cataloged procedure, to minimum phase wavelet w (t)=(w ₀, w ₁..., w _n) carry out 0-1 coding after, just obtain 2 ⁿPlant coded data.

As shown in Figure 12, provided by the inventionly a kind ofly carry out the system that wavelet estimates based on seismic data and well-log information and also comprise:

Well lie data extract device 400 is used for extracting well lie data according to described seismic data;

Reflection coefficient is determined device 500, is used for determining reflection coefficient according to described well-log information.This device at first extracts interval transit time curve and densimetric curve from well-log information, calculate reflection coefficient according to interval transit time curve and densimetric curve then.

The well control evaluation model is determined device 600, is used for determining the well control evaluation model according to described well lie data, reflection coefficient and described convolution model.Estimate the quality between the individuality, need weigh by the size of fitness value.The present invention as control criterion, adopts the related coefficient of theogram w (t) * r (t) and well lie seismologic record as the well control evaluation model, namely with well

O=max[xcorr(w(t)*r(t),s(t))]

Wherein, max[] represent to get maximal value, s (t) is well lie data, the seismic wavelet that w (t) comes out for decoding, the reflection coefficient that r (t) calculates for logging trace, xcorr are represented the related coefficient between composite traces w (t) * r (t) and the well lie record s (t), and O is objective function.

Best wavelet estimation unit 700 is used for estimating best wavelet according to described well control evaluation model and described multiple coded data that Figure 14 is the concrete structure block diagram of best wavelet estimation unit, and as shown in Figure 14, best wavelet estimation unit specifically comprises:

Coded data acquisition module 701 is for the coded data of choosing predetermined number from described multiple coded data.Because the operand of genetic algorithm is a colony, the coded data individuality (chromosome) that therefore produces predetermined number is at random formed population.The scale of population is exactly number individual in the population, generally is taken as even number.Each individuality represents an initial wavelet, and population represents an initial wavelet collection.Concentrate in initial wavelet, each wavelet has identical length, identical spectral amplitude and different phase spectrums.In concrete embodiment, predetermined number can preestablish, and is set to even number in the 20-100 scope such as predetermined number.

Decoder module 702 is used for the coded data of described predetermined number is decoded, and obtains the wavelet sequence of predetermined number.The concrete structure block diagram of decoder module specifically comprises as shown in figure 15:

Decoding unit 7021 is used for described coded data is carried out the 0-1 decoding, obtains corresponding root.Further introduce decode procedure of the present invention below.According to described binary wavelet transform, coded data is mapped to described wavelet root of polynomial according to the 0-1 of genetic algorithm coding is counter.

Polynomial expression is asked for unit 7022, is used for according to the described polynomial expression of asking for about variable z.Namely according to root (α ₁, α ₂..., α _i..., α _n) obtain the polynomial expression about variable z, the polynomial expression that solves is as shown in Equation (1).

Contrary transform unit 7023 is used for the polynomial expression of described variable z is carried out contrary transform, obtains the wavelet sequence of described coded data correspondence.Polynomial expression about variable z is carried out contrary transform, and the wavelet sequence that obtains is w (t)=(w ₀, w ₁..., w _n), n is the sampling number of seismic wavelet, wherein w ₀w _n≠ 0.

The decoding that is basic genetic algorithmic is with { 0,1} binary string is transformed into decimal number, and then linear transformation becomes a real number again.And decoding of the present invention is with { 0,1} binary string is transformed into n root (comprising compound radical and real root), obtains a sub-wave train by asking polynomial expression and getting coefficient then.Suppose that certain individuality after decoded is

, then Xie Ma individuality has been endowed clear physical meaning.

As shown in Figure 14, best wavelet estimation unit also comprises:

Evaluation module 703 is used for judging according to described well control evaluation model whether the wavelet sequence of described predetermined number exists best wavelet sequence, and Figure 16 is the concrete structure figure of evaluation module, and as shown in Figure 16, evaluation module specifically comprises:

Threshold value acquiring unit 7031 is used for obtaining predefined well control threshold value coefficient;

Convolution unit 7032 is used for described wavelet sequence is carried out convolution, the formation theogram with corresponding reflection coefficient;

Related coefficient determining unit 7033 is used for determining the related coefficient that described theogram and well lie record according to described well control evaluation model;

Judging unit 7034 is used for judging that whether described related coefficient is more than or equal to described well control threshold value coefficient.

As shown in Figure 14, best wavelet estimation unit also comprises:

Best wavelet determination module 704 is used for being judged as when being when described evaluation module, exports described best wavelet sequence;

Sub-wave phase update module 705 is used for upgrading the phase place of described wavelet when described evaluation module is judged as not.Return the described coded data acquisition module of execution according to the sub-wave phase after upgrading.

Can adopt genetic manipulation to upgrade sub-wave phase, genetic manipulation generally comprises selection, intersects and variation, and its fundamental purpose is that individuality is transformed.In the present invention, their effect is the phase place of upgrading wavelet.

Wherein " selection " is to select good wavelet in order to concentrate from current wavelet, makes them have an opportunity to be breeding descendants of future generation as parent.The ideal adaptation degree is more high, and its selecteed chance is just more many.Selection is the key of genetic algorithm, and it has embodied the thought of the occurring in nature survival of the fittest.

" intersection " replaced the part-structure of two father's individualities reorganization and generated new individual operation.The purpose of this operation is in order to produce new individuality the next generation, and just as the marriage process of human society, by interlace operation, the ability of genetic algorithm search equal length, same-amplitude spectrum, out of phase music score ripple will be greatly improved." intersection " is the most important means that genetic algorithm is obtained best wavelet and reflection coefficient.

" variation " is that the gene position of individuality is carried out inversion operation.By mutation operation, can make genetic algorithm keep the diversity of population, to prevent non-ripe convergence.Generally speaking, the variation probability is all obtained less, sometimes even get 0.

Select, intersect and make a variation a lot of diverse ways are respectively arranged.At particular problem of the present invention, in embodiment, can choose roulette back-and-forth method, the even bracketing method of scale-of-two and scale-of-two alternative method respectively.Also be that this paper utilizes the scale-of-two genetic algorithm to upgrade the phase place of wavelet, and obtain best wavelet estimation by making theogram and seismic trace near well reach optimum matching.

What as shown in Figure 12, the embodiment of the invention provided a kind ofly carries out the system that wavelet estimates based on seismic data and well-log information and also comprises:

Reservoir prediction device 800 is used for carrying out reservoir prediction according to described best wavelet, obtains the reservoir prediction result on current stratum.When carrying out reservoir prediction, all need to be applied to convolution model, and the seismic wavelet in the convolution model accurately whether, will directly affect the accuracy of follow-up reservoir prediction.In actual seismic wavelet estimation, because the influence of factors such as noise, the phase spectrum of wavelet is difficult to accurately estimate.Carry out deconvolution based on inaccurate phase place wavelet, its result's resolution will be subjected to restriction to a certain degree.Such as utilizing amplitude to change (Amplitude versus angle with angle, AVA) petroleum-gas prediction is carried out in inverting, (Amplitude Versus Offset AVO) carries out the identification of hydrocarbon-bearing formation to stratum three parameters (velocity of longitudinal wave, shear wave velocity and density) with the variation of offset distance such as amplitude.Therefore, the method that the present invention proposes is directly the wavelet of estimating to be carried out the transform rooting, scans the phase place of wavelet then by the scale-of-two genetic algorithm, and then the best wavelet of evolving out rapidly, improve the wavelet estimated accuracy, and then improved the precision of reservoir prediction.

Below in conjunction with specific embodiment, introduce feasibility of carrying out wavelet estimation approach and system based on seismic data and well-log information provided by the invention in detail.

Fig. 1 (a) is not for there being initial wavelet composite traces figure under the situation of making an uproar, and Fig. 1 (b) is not for there being raw readings figure under the situation of making an uproar, and Fig. 1 (c) is not for there being the wavelet composite traces figure that the present invention estimates under the situation of making an uproar.Wherein, the related coefficient of initial wavelet composite traces and raw readings is 46%, and the present invention estimates that the related coefficient of wavelet composite traces and raw readings is 100%.Can be found that by above-mentioned accompanying drawing under nothing was made an uproar situation, the well of initial wavelet shake matching degree was very low, after estimating through the sub-wave phase spectrum of the inventive method, the well shake coupling of final mixed phase wavelet has reached 100%.

Fig. 2 is not for there being the comparison diagram that initial wavelet under the situation of making an uproar, true wavelet and the present invention estimate wavelet.As shown in Figure 2, do not contain at geological data under the situation of noise, the inventive method can accurately be asked for the mixed-phase seismic wavelet.

Fig. 3 (a) is initial wavelet composite traces figure under the noisy situation, and Fig. 3 (b) is raw readings figure under the noisy situation, and Fig. 3 (c) estimates wavelet composite traces figure for the present invention under the noisy situation; Institute's plus noise is 50% random noise (50% the meaning is that the ratio of noise energy and signal energy is 1:2).Wherein, the related coefficient of initial wavelet composite traces and raw readings is 36%, and the present invention estimates that the related coefficient of wavelet composite traces and raw readings is 82%.By above-mentioned accompanying drawing as can be known, containing under the noise situation, the well of the mixed phase wavelet of estimation shake matching degree has obtained bigger raising.

Fig. 4 estimates the comparison diagram of wavelet for initial wavelet, true wavelet and the present invention under the noisy situation, and as shown in Figure 4, when having noise, the estimation wavelet is coincide better with true wavelet.

Fig. 5 (a) is Chang Xiangwei wavelet composite traces synoptic diagram; Fig. 5 (b) is well lie real data synoptic diagram; The wavelet composite traces synoptic diagram that Fig. 5 (c) estimates for the present invention.Wherein, the related coefficient of Chang Xiangwei wavelet composite traces and well lie real data is 61.8%, and the present invention estimates that the related coefficient of wavelet composite traces and well lie real data is 67.1%.At real data, the well of the inventive method shake matching degree has improved 5.3% than classic method.

The wavelet synoptic diagram that Fig. 6 (a) estimates for the present invention; The wavelet synoptic diagram that Fig. 6 (b) estimates for the Chang Xiangwei rotation.Fig. 7 (a) is original seismic data synoptic diagram; Fig. 7 (b) is for estimating the determinacy deconvolution result schematic diagram of wavelet based on the present invention; Fig. 7 (c) is Chang Xiangwei wavelet deconvolution result schematic diagram.By above-mentioned accompanying drawing as can be known, relative classic method, the deconvolution result's of the inventive method resolution has obtained effective raising, and has correlativity preferably with well.

In sum, because mostly actual seismic wavelet is mixed-phase, Chang Xiangwei rotates the wavelet estimation technique relatively, and the present invention has more advance, and precision is higher.In addition, phase estimation is inaccurate also is a low key reason of well shake stated accuracy in the actual production.Than the well shake matching result that the Chang Xiangwei rotation obtains, this invention promotes precision and reaches more than 5%.

When existing converter technique asked for mixed phase wavelet, sub-wave phase there is to a certain degree dependence.If the condition of lax estimator wave phase, its calculated amount is huge, is difficult at present put in the actual production.And the present invention is by carrying out specific coding and decoding to wavelet root conversion process, and introducing genetic algorithm, adopt genetic manipulation and handle the sub-wave phase problem in conjunction with well control as interpretational criteria, reduced calculated amount effectively, can realize the quick convergence that mixed phase wavelet is estimated.

Useful achievement of the present invention is:

1) the present invention utilizes root in the feature of the inside and outside movement of unit circle, carries out specific coding to moving root mode, and then uses selection, intersection and the variation of genetic algorithm to upgrade sub-wave phase heuristicly, scans best wavelet after repeatedly evolving.

2) the present invention is by carrying out specific coding and decoding to wavelet root conversion process, and introducing genetic algorithm, adopt genetic manipulation and handle the sub-wave phase problem in conjunction with well control as interpretational criteria, reduced calculated amount effectively, accelerate the quick convergence that root varitron wave phase spectrum is estimated, improved the wavelet estimated accuracy.

One of ordinary skill in the art will appreciate that all or part of flow process that realizes in above-described embodiment method, can instruct relevant hardware to finish by computer program, described program can be stored in the general computer read/write memory medium, this program can comprise the flow process as the embodiment of above-mentioned each side method when carrying out.Wherein, described storage medium can be magnetic disc, CD, read-only storage memory body (Read-Only Memory, ROM) or at random store memory body (Random Access Memory, RAM) etc.

Those skilled in the art can also recognize that the various functions that the embodiment of the invention is listed are to realize depending on the designing requirement of specific application and total system by hardware or software.Those skilled in the art can be for every kind of specific application, and can make ins all sorts of ways realizes described function, but this realization should not be understood that to exceed the scope of embodiment of the invention protection.

Used specific embodiment among the present invention principle of the present invention and embodiment are set forth, the explanation of above embodiment just is used for helping to understand method of the present invention and core concept thereof; Simultaneously, for one of ordinary skill in the art, according to thought of the present invention, the part that all can change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.

Claims (12)

1. one kind is carried out the wavelet estimation approach based on seismic data and well-log information, it is characterized in that described method comprises:

Gather seismic data, the well-log information on current stratum;

Estimate minimum phase wavelet according to described seismic data;

Carry out the 0-1 coding according to described minimum phase wavelet, obtain multiple coded data;

Extract well lie data according to described seismic data;

Determine reflection coefficient according to described well-log information;

Determine the well control evaluation model according to described well lie data, reflection coefficient and described convolution model;

Estimate best wavelet according to described well control evaluation model and described multiple coded data;

Carry out reservoir prediction according to described best wavelet, obtain the reservoir prediction result on current stratum.

2. method according to claim 1 is characterized in that, carries out the 0-1 coding according to described minimum phase wavelet, obtains multiple coded data and specifically comprises:

Described minimum phase wavelet is carried out transform, obtain the polynomial expression about variable z;

Find the solution described polynomial expression to determine the root of described polynomial expression correspondence;

Root to described polynomial expression correspondence carries out the 0-1 coding, obtains multiple coded data.

3. method according to claim 2 is characterized in that, the well control evaluation model of determining according to described well lie data, reflection coefficient and described convolution model is:

O=max[xcorr(w(t)*r(t),s(t))]

Wherein, max[] expression gets maximal value, s (t) is well lie data, the seismic wavelet that w (t) obtains for decoding, r (t) is the reflection coefficient of determining according to described well-log information, xcorr is the related coefficient between theogram w (t) * r (t) and the well lie record s (t), and O is objective function.

4. method according to claim 3 is characterized in that, estimates that according to described well control evaluation model and described multiple coded data best wavelet specifically comprises:

From described multiple coded data, choose the coded data of predetermined number;

Coded data to described predetermined number is decoded, and obtains the wavelet sequence of predetermined number;

Judge according to described well control evaluation model in the wavelet sequence of described predetermined number and whether have best wavelet sequence;

When being judged as when being, export described best wavelet sequence;

When being judged as not, upgrade the phase place of described wavelet;

Return predetermined number is chosen in execution from described multiple coded data coded data according to the sub-wave phase after upgrading.

5. method according to claim 4 is characterized in that, the coded data of described predetermined number is decoded, and the wavelet sequence that obtains predetermined number specifically comprises:

Described coded data is carried out the 0-1 decoding, obtain corresponding root;

According to the described polynomial expression of asking for about variable z;

Polynomial expression to described variable z carries out contrary transform, obtains the wavelet sequence of described coded data correspondence.

6. whether method according to claim 4 is characterized in that, judge according to described well control evaluation model in the wavelet sequence of described predetermined number to exist the wavelet sequence that satisfies condition specifically to comprise:

Obtain predefined well control threshold value coefficient;

Described wavelet sequence is carried out convolution with corresponding reflection coefficient, form theogram;

Determine the related coefficient of described theogram and well lie record according to described well control evaluation model;

Judge that whether described related coefficient is more than or equal to described well control threshold value coefficient.

7. one kind is carried out the system that wavelet is estimated based on seismic data and well-log information, it is characterized in that described system comprises:

Harvester is used for gathering seismic data, the well-log information on current stratum;

Minimum phase wavelet is determined device, is used for estimating minimum phase wavelet according to described seismic data;

Code device is used for carrying out the 0-1 coding according to described minimum phase wavelet, obtains multiple coded data;

Well lie data extract device is used for extracting well lie data according to described seismic data;

Reflection coefficient is determined device, is used for determining reflection coefficient according to described well-log information;

The well control evaluation model is determined device, is used for determining the well control evaluation model according to described well lie data, reflection coefficient and described convolution model;

Best wavelet estimation unit is used for estimating best wavelet according to described well control evaluation model and described multiple coded data;

The reservoir prediction device is used for carrying out reservoir prediction according to described best wavelet, obtains the reservoir prediction result on current stratum.

8. system according to claim 7 is characterized in that, described code device specifically comprises:

The transform unit is used for described minimum phase wavelet is carried out transform, obtains the polynomial expression about variable z;

The polynomial solving unit is used for finding the solution described polynomial expression to determine the root of described polynomial expression correspondence;

Coding unit is used for the root of described polynomial expression correspondence is carried out the 0-1 coding, obtains multiple coded data.

9. system according to claim 8 is characterized in that, described well control evaluation model determines that the well control evaluation model that device is determined is:

O=max[xcorr(w(t)*r(t),s(t))]

Wherein, max[] represent to get maximal value, s (t) is well lie data, the seismic wavelet that w (t) obtains for decoding, r (t) is the reflection coefficient of determining according to well-log information, and xcorr is the related coefficient between composite traces w (t) * r (t) and the well lie record s (t), and O is objective function.

10. system according to claim 9 is characterized in that, described best wavelet estimation unit specifically comprises:

The coded data acquisition module is for the coded data of choosing predetermined number from described multiple coded data;

Decoder module is used for the coded data of described predetermined number is decoded, and obtains the wavelet sequence of predetermined number;

Evaluation module is used for judging according to described well control evaluation model whether the wavelet sequence of described predetermined number exists best wavelet sequence;

Best wavelet output module is used for being judged as when being when described evaluation module, exports described best wavelet sequence;

The sub-wave phase update module is used for upgrading the phase place of described wavelet when described evaluation module is judged as not, returns according to the sub-wave phase after upgrading and carries out described coded data acquisition module.

11. system according to claim 10 is characterized in that, described decoder module specifically comprises:

Decoding unit is used for described coded data is carried out the 0-1 decoding, obtains corresponding root;

Polynomial expression is asked for the unit, is used for according to the described polynomial expression of asking for about variable z;

Contrary transform unit is used for the polynomial expression of described variable z is carried out contrary transform, obtains the wavelet sequence of described coded data correspondence.

12. system according to claim 10 is characterized in that, described evaluation module specifically comprises:

The threshold value acquiring unit is used for obtaining predefined well control threshold value coefficient;

The convolution unit is used for described wavelet sequence is carried out convolution, the formation theogram with corresponding reflection coefficient;

The related coefficient determining unit is used for determining the related coefficient that described theogram and well lie record according to described well control evaluation model;

Judging unit is used for judging that whether described related coefficient is more than or equal to described well control threshold value coefficient.

Images