CN104237941A - Coal bed gas prediction method based on frequency attenuation - Google Patents

Abstract

The invention provides a coal bed gas prediction method based on frequency attenuation, and belongs to the field of seismic exploration and development of oil gas and coal bed gas. The coal bed gas prediction method comprises the steps that (1) seismic data x(t) are input; (2) pulse deconvolution frequency increase processing is conducted on the seismic data x(t), and a seismic record S(t) after frequency increase is obtained; (3) wavelet transformation spectral decomposition is conducted on the seismic record S(t) obtained after the frequency increase, and frequency spectra at the top and the bottom of a coal bed are obtained; (4) the difference value between the frequency spectra at the top and the bottom of the coal bed is obtained, contrastive analysis is conducted on the coal bed and the difference between the frequency spectra, and a frequency attenuation result of the coal reservoir is obtained; (5) the frequency attenuation result of the coal reservoir is output. According to the coal bed gas prediction method based on frequency attenuation, the pulse deconvolution frequency increase technology is applied, so that dominant frequency of the seismic data is effectively increased, an effective frequency band is broadened, and a data guarantee is provided for analysis of seismic wave frequency absorption attenuation features.

Description

A kind of coal-seam gas Forecasting Methodology based on frequency decay

Technical field

The invention belongs to oil gas and coalbed gas seismic exploration and development field, be specifically related to a kind of coal-seam gas Forecasting Methodology based on frequency decay, the integrated forecasting for coal-seam gas Favorable Reservoir gas-bearing property is studied.

Background technology

The research of scattering theory shows, when seismic event is through the energy attenuation of meeting spot seismic wave during hydrocarbon-bearing formation, this relates to stratum to oil gas in the rubbing action of seismic event and stratum to the energy absorption of seismic event.Experiment proves, the energy attenuation of seismic event increases along with the increase of seismic wave propagation distance, it is very fast that oil-bearing reservoir then shows as high-frequency energy decay to the energy attenuation that seismic event causes, and low frequency energy decay is comparatively slow, thus causes the seismic event frequency band through oil-bearing reservoir to narrow.

Coalbed methane reservoir belongs to a kind of unconventionaloil pool reservoir, and its feature is different from general oil and gas reservoir.Certain development is obtained in recent years to the research of coalbed methane reservoir, but the geophysics correlation technique related to is not also ripe especially, in the urgent need to appearance and the development of coalbed methane reservoir forecasting techniques stronger targetedly, therefore study the coal-seam gas forecasting techniques based on seismic wave energy decay here, wish the precision promoting coalbed methane reservoir gas distribution prediction.

The feature mainly decayed by high-frequency energy for the research of this decay is at present showed, and these irregular decay are very useful to hydro carbons instruction.Because instantaneous spectrum analysis can extract the frequency spectrum of each sampling point of seismic trace, seismic frequency is decayed the spectral change that can be described to based on frequency, utilizes spectral change can detect the hydrocarbon-bearing pool relevant with high frequency attenuation.

Spectral Imaging Technology adopts the algorithm based on discrete Fourier transformation in the past usually, but the method also exists obvious limitation, this is because the key character of the seismic amplitude spectrum of estimation is the function of selected window length.If selected time window is shorter, spectral amplitude can with transforming function transformation function convolution, thus lose the local characteristic of frequency; Another shortcoming is, shorter time window can make the secondary lobe of wavelet be rendered as the illusion of individual reflection.Increase the resolution that window length can improve frequency, if but selected time window is long, time window in multiple reflections spectral amplitude can be made to take flute mark as feature, the time window problem of Fourier transform related algorithm is difficult to the spectral amplitude feature distinguishing single reflection, so can make the estimation of spectral amplitude produce deviation.In practice, the selection of window when being difficult to grasp, and cannot quantitative test window length produce deviation.In this case, the time-frequency analysis technology based on wavelet transformation becomes the important tools of analysis of non-stationary signal, instead of Fourier transform analytical approach in a lot of practice.Pass through theory calculate, frequency splitting technology such as contrast wavelet transformation and Fast Fourier Transform (FFT), discrete Fourier transformation and maximum entropy method etc., the result Wigner distribution demonstrated based on wavelet transformation can obtain accurate time frequency analysis result, window problem when simultaneously can avoid; Meanwhile, the approach time that instantaneous spectrum analysis during elimination after window can be best and frequency location, the energy that spectral imaging obtains and phase spectrum can the time variation in thickness of describing reservoir the horizontal geology of instruction is discontinuous better.Therefore, it is feasible for carrying out frequency decay research based on wavelet transformation spectrum analysis, the method is used for innovatively the analysis of coalbed methane reservoir gas distribution prediction and belongs to precedent, have primacy, can provide support for the research of coal-seam gas forecasting techniques.

Coal-seam gas is the unconventionaloil pool that a kind of adsorbability is stronger, distinguishes comparatively large with conventional gas and oil reservoir characteristic, and main manifestations is its free property, adsorbability etc.There is blindness and non-specific aim in the forecasting research that direct utilization carries out coal-seam gas for the gas distribution prediction technology of conventional gas and oil reservoir.Stronger and the effective coal-seam gas forecasting techniques of development research specific aim be unconventionaloil pool exploration and development in the urgent need to.

Summary of the invention

The object of the invention is to solve the difficult problem existed in above-mentioned prior art, a kind of coal-seam gas Forecasting Methodology based on frequency decay is provided, improve prediction effect.

The present invention is achieved by the following technical solutions:

Based on a coal-seam gas Forecasting Methodology for frequency decay, comprising:

(1) geological data x (t) is inputted;

(2) spiking deconvolution is carried out to geological data x (t) and carry seismologic record S (t) frequently processing and obtain carrying frequently;

(3) carry out wavelet transformation spectral decomposition obtain top, coal seam, the frequency spectrum at the end to carrying seismologic record S (t) frequently;

(4) obtain top, coal seam, the spectrum difference at the end, carry out the comparative analysis of coal seam and spectrum difference, obtain the frequency decay result of coal seam reservoirs;

(5) the frequency decay result of coal seam reservoirs is exported.

Described step (2) is achieved in that

Input geological data x (t), obtains frequency field seismologic record X (ω) according to formula (2), then obtains new seismologic record S (t) according to formula (18), is and carries the seismologic record frequently;

x(t)＝b(τ)*ξ(t)， (2)

Fourier transform is asked to both sides, then obtains frequency field seismologic record X (ω):

X(ω)＝B(ω)*ξ(ω)，(3)

X (ω), B (ω) ξ (ω) are respectively the frequency spectrum of seismic spectrum, wavelet spectrum and reflection coefficient;

S(t)＝a(t)*x(t)， (18)

Wherein, a (t) is the inverse filtering factor, is tried to achieve by (17) formula:

$\left[\begin{array}{cccc}{r}_{\mathrm{xx}}\left(0\right)& r_{\mathrm{xx}}\left(1\right)& ...& r_{\mathrm{xx}}\left(m\right)\\ r_{\mathrm{xx}}\left(1\right)& r_{\mathrm{xx}}\left(0\right)& ...& r_{\mathrm{xx}}(m-1)\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ r_{\mathrm{xx}}\left(m\right)& r_{\mathrm{xx}}(m-1)& ...& r_{\mathrm{xx}}\left(0\right)\end{array}\right]\left[\begin{array}{c}{r}_{\mathrm{xx}}\left(0\right)\\ a\left(1\right)\\ .\\ .\\ .\\ a\left(m\right)\end{array}\right]\left[\begin{array}{c}a\left(0\right)\\ a\left(1\right)\\ .\\ .\\ .\\ a\left(m\right)\end{array}\right]=\left[\begin{array}{c}1\\ 0\\ .\\ .\\ .\\ 0\end{array}\right]---\left(17\right)$

(18) r in formula _xx(τ) be the auto-correlation of seismologic record x (t).

Described step (3) is achieved in that

Utilize formula (19) respectively to the frequency spectrum imaging analysis that top, coal seam, seismologic record S (t) at the end carry out based on wavelet transformation, obtain frequency spectrum P corresponding to top, coal seam, the end _on, P _under;

${\mathrm{WT}}_x(a,b)=\frac{1}{\sqrt{a}}{\∫}_{-\∞}^{+\∞}x\left(t\right){\mathrm{\ψ}}^{*}\left(\frac{t-b}{a}\right)\mathrm{dt}={\∫}_{-\∞}^{+\∞}x\left(t\right){\mathrm{\ψ}}_{a,b}^{*}\left(t\right)\mathrm{dt}---\left(19\right)$

Calculate P _ontime, replace the x (t) in formula (19) to calculate with seismologic record S (t) on top, coal seam, calculate P _undertime, replace the x (t) in formula (19) to calculate with seismologic record S (t) at the bottom of coal seam.

Described step (4) is achieved in that

By frequency spectrum P corresponding for top, coal seam _onthe frequency spectrum P corresponding with at the bottom of coal seam _undercarry out subtracting each other obtaining spectrum difference P _{on -under}, then analyze P corresponding to coal seam _{previous-next}the size of value and the corresponding relation in coal seam, P _{previous-next}value show that greatly the frequency decay in coal seam is large, otherwise then show that the frequency decay in coal seam is little; The P that frequency decay is large and corresponding _{previous-next}the region that value is larger is the favourable band of coal-seam gas gassiness of prioritizing selection.

Compared with prior art, the invention has the beneficial effects as follows:

(1) the present invention effectively overcomes the time window problem of routine techniques, obtains correct time and frequency location, effectively can portray the thickness of reservoir and the horizontal uncontinuity on instruction stratum.Carry frequently process and wavelet transform base are carried out based on frequency decay coal-seam gas forecasting techniques, there is higher reliability, effectively confirm the feature that the seismic event occurrence frequency through coal seam is decayed, the seismic band of its underlying formation is narrower compared with overlying strata, coal seam, radio-frequency component content is lower.It is more obvious that application the present invention carries out distributive province, coal seam coalbed methane reservoir gas distribution prediction effect.

(2) effectively the geological data of time domain is converted to the seismic energy data of frequency field based on the FREQUENCY SPECTRUM DECOMPOSITION TECHNIQUE of wavelet transformation, and the FREQUENCY SPECTRUM DECOMPOSITION TECHNIQUE of this technology and routine as window problem when there is compared with discrete Fourier transformation, Fast Fourier Transform (FFT), maximum entropy method etc. elimination, exact picture time and frequency location, portray reservoir thickness and indicate the advantage of the horizontal uncontinuity feature in stratum, lay a good foundation for better carrying out frequency decay analysis.

Accompanying drawing explanation

Fig. 1 (a) is the frequency spectrum of coal seam overlying strata.

Fig. 1 (b) is the frequency spectrum of coal seam sub-surface.

Fig. 2 is that seismic event passes spectral decay feature behind coal seam.

Fig. 3 is through the planar distribution of the frequency decay feature in coal seam.

Fig. 4 is the step block diagram of the inventive method.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in further detail:

Conventional seismic processing data dominant frequency is lower, frequency band is narrower, is used for carrying out coal-seam gas and predicts that resolution is lower.What by spiking deconvolution, the present invention carried that frequently technology carries out seismic data proposes process frequently, and dominant frequency is increased to about 50Hz, and frequency span have also been obtained increase.Carry the geological data to be frequently conducive to carrying out the coal-seam gas forecasting techniques research based on frequency decay, be summed up as the energy absorption that coalbed methane reservoir is stronger to seismic event, particularly to the absorption of the high-frequency energy absorption much larger than low frequency energy, greatly, prediction effect is obvious for seismic wave energy change difference before and after frequency decay.

Theoretical research shows when seismic event is through the attenuation by absorption of high-frequency energy can occur during oil-bearing reservoir, thus makes seismic wave characteristic show as high-frequency energy to reduce and the relative increase of low frequency energy, through relative the narrowing of seismic event frequency band of oil and gas reservoir.Utilize this theoretical foundation, the technical method based on frequency decay is used for the analysis and research of coal-seam gas gas distribution prediction by the present invention innovatively.Window problem when the conventional frequency spectrum analysis method for frequency decay analysis is as intrinsic in having based on the spectrum analysis technique of discrete Fourier transformation and Fast Fourier Transform (FFT), maximum entropy method, and window problem when based on the Wigner distribution technology of wavelet transformation there is elimination, can exact picture time and frequency location, effectively portray the time thickness of reservoir and indicate the advantage of horizontal geology uncontinuity feature.Therefore the present invention determines frequency decay technology the researching and analysing coal-seam gas gas distribution prediction based on wavelet transformation frequency spectrum analysis method, has novelty.

(1) spiking deconvolution carries the seismic data resolution of technology raising frequently

Propose process frequently by spiking deconvolution to coal-seam gas seismic data, carry the seismic data frequency band frequently and broaden, dominant frequency can reach 45-50Hz.Spiking deconvolution algorithm is as follows:

If seismologic record: $x\left(t\right)=S\left(t\right)+n\left(t\right)={\mathrm{\Σ}}_{\mathrm{\τ}=0}^{\∞}b\left(\mathrm{\τ}\right)\mathrm{\ξ}(t-\mathrm{\τ})+n\left(t\right),---\left(1\right)$

Wherein S (t) is useful signal, and n (t) is interference wave.

First suppose to there is not interference wave n (t), that is:

x(t)＝S(t)＝b(τ)*ξ(t)，(2)

B (τ) in above formula represents seismic wavelet, and ξ (t) represents reflection coefficient.

Fourier transform is asked to both sides, then obtains the seismologic record expression formula of frequency field:

X(ω)＝B(ω)*ξ(ω)，(3)

In formula, X (ω), B (ω) and ξ (ω) are respectively the frequency spectrum of seismic spectrum, wavelet spectrum and reflection coefficient, obviously,

$\mathrm{\ξ}\left(\mathrm{\ω}\right)=\frac{1}{B\left(\mathrm{\ω}\right)}\·\left(\mathrm{\ω}\right),---\left(4\right)$

If order: $A\left(\mathrm{\ω}\right)=\frac{1}{B\left(\mathrm{\ω}\right)},---\left(5\right)$

Then have: ξ (ω)=A (ω) X (ω), (6)

Again inversefouriertransform is done to time domain to (6) formula, obtains:

ξ(ω)＝a(t)*x(t)＝a(t)*b(t)*ξ(t)，(7)

In formula, a (t) is the function of time of A (ω).

Obtain according to (7) formula:

a(t)*b(t)＝δ(t)，，(8)

Because b (t) is seismic wavelet, and there is frequency spectrum relation reciprocal each other between a (t) and b (t) (namely ), so a (t) is called anti-wavelet, be called deconvolution operator again.

First suppose in research process that desired output is burst pulse d (t), when wavelet is known, if inverse filtering factor a (t) initial time to be asked is-m ₀, continuity length is (m+1).That is:

a(t)＝[a(-m ₀)，a(-m ₀+1)，a(-m ₀+2)，...，a(-m ₀+m)]，(9)

When known input seismic wavelet b (t)=[b (0), b (1) ..., b (n)] time, actual output is:

$c\left(t\right)=a\left(t\right)*b\left(t\right)={\mathrm{\Σ}}_{\mathrm{\τ}=-m_0}^{-m_0+m}a\left(\mathrm{\τ}\right)b(t-\mathrm{\τ}),---\left(10\right)$

Actual output with the error sum of squares of desired output is:

$Q={\mathrm{\Σ}}_{t=-m_0}^{-m_0+m+n}{[{\mathrm{\Σ}}_{\mathrm{\τ}=-m_0}^{-m_0+m}a\left(\mathrm{\τ}\right)b(t-\mathrm{\τ})-d\left(t\right)]}^2,---\left(11\right)$

Make Q be minimum, be exactly mathematically the extreme-value problem asking Q, namely ask satisfied:

$\frac{\∂Q}{\∂a\left(l\right)}=0(l=-m_0,-m_0+1,...,-m_0+m),---\left(12\right)$

Filtering factor a (t).Because, ${\mathrm{\Σ}}_{t=-m_0}^{-m_0+m+n}b(t-\mathrm{\τ})b(t-l)=r_{\mathrm{bb}}(l-\mathrm{\τ})$ For the autocorrelation function of seismic wavelet, and ${\mathrm{\Σ}}_{t=-m_0}^{-m_0+m+n}d(t-\mathrm{\τ})b(t-l)r_{\mathrm{bd}}\left(l\right)$ For the cross correlation function of seismic wavelet and desired output, therefore (12) formula can be write as:

$\left[\begin{array}{cccc}{r}_{\mathrm{bb}}\left(0\right)& r_{\mathrm{bb}}\left(1\right)& ...& r_{\mathrm{bb}}\left(m\right)\\ r_{\mathrm{bb}}\left(1\right)& r_{\mathrm{bb}}\left(0\right)& ...& r_{\mathrm{bb}}(m-1)\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ r_{\mathrm{bb}}\left(m\right)& r_{\mathrm{bb}}(m-1)& ...& r_{\mathrm{bb}}\left(0\right)\end{array}\right]\left[\begin{array}{c}a(-m_0)\\ a(-m_0+1)\\ .\\ .\\ .\\ a(-m_0+m)\end{array}\right]=\left[\begin{array}{c}{r}_{\mathrm{bd}}(-m_0)\\ r_{\mathrm{bd}}(-m_0+1)\\ .\\ .\\ .\\ r_{\mathrm{bd}}(-m_0+m)\end{array}\right],---\left(13\right)$

This equation coefficient matrix is mop Ritz matrix.

If desired output is δ pulse, then cross-correlation is:

$r_{\mathrm{bd}}\left(l\right)={\mathrm{\Σ}}_{t=-m0}^{-m_0+m+n}\mathrm{\δ}\left(t\right)b(t-l)=b(-l),\left(14\right)$

Fundamental equation becomes:

$\left[\begin{array}{cccc}{r}_{\mathrm{bb}}\left(0\right)& r_{\mathrm{bb}}\left(1\right)& ...& r_{\mathrm{bb}}\left(m\right)\\ r_{\mathrm{bb}}\left(1\right)& r_{\mathrm{bb}}\left(0\right)& ...& r_{\mathrm{bb}}(m-1)\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ r_{\mathrm{bb}}\left(m\right)& r_{\mathrm{bb}}(m-1)& ...& r_{\mathrm{bb}}\left(0\right)\end{array}\right]\left[\begin{array}{c}a(-m_0)\\ a(-m_0+1)\\ .\\ .\\ .\\ a(-m_0+m)\end{array}\right]=\left[\begin{array}{c}{b}_{}(-m_0)\\ {}_{}b(m_0-1)\\ .\\ .\\ .\\ b(m_0+m)\end{array}\right],---\left(15\right)$

Generally, seismic wavelet is unknown, and filtering factor of negating when unknown wavelet, must add certain assumed condition to seismic wavelet and reflection coefficient sequence, comprise:

A. suppose that reflection coefficient sequence R (t) is random white noise sequence, namely its auto-correlation is:

B. suppose that seismic wavelet is minimum phase.

According to hypothesis A, seismic wavelet auto-correlation r _bb(τ) the auto-correlation r of seismologic record x (t) can be used _xx(τ) replace.According to hypothesis B, the zero point of transform B (z) of known seismic wavelet, all outside unit circle, is also the transform of inverse filtering factor a (t) zero point of denominator polynomials entirely outside unit circle, therefore a (t) is stable, physically realizable.Therefore, m ₀=0, free term becomes [b (0), b (-1) ..., b (-m)] ^tagain because b (t) must be physically realizable, therefore b (-1)=0, b (-2)=0 ..., b (-m)=0.Make a (t)=a (t)/b (0), then fundamental equation becomes:

$\left[\begin{array}{cccc}{r}_{\mathrm{xx}}\left(0\right)& r_{\mathrm{xx}}\left(1\right)& ...& r_{\mathrm{xx}}\left(m\right)\\ r_{\mathrm{xx}}\left(1\right)& r_{\mathrm{xx}}\left(0\right)& ...& r_{\mathrm{xx}}(m-1)\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ r_{\mathrm{xx}}\left(m\right)& r_{\mathrm{xx}}(m-1)& ...& r_{\mathrm{xx}}\left(0\right)\end{array}\right]\left[\begin{array}{c}a\left(0\right)\\ a\left(1\right)\\ .\\ .\\ .\\ a\left(m\right)\end{array}\right]=\left[\begin{array}{c}1\\ 0\\ .\\ .\\ .\\ 0\end{array}\right]---\left(17\right)$

The fundamental equation of Here it is spiking deconvolution, in its matrix of coefficients, each element can directly be tried to achieve by seismologic record, r in matrix _xx(τ) value represents the autocorrelation function (above-described assumed condition) of seismologic record respectively, and wherein seismologic record is known, according to formula ${\mathrm{\Σ}}_{t=-m_0}^{-m_0+m+n}b(t-\mathrm{\τ})b(t-l)=r_{\mathrm{bb}}(l-\mathrm{\τ})$ Computing can to obtain in matrix each value, namely obtain matrix of coefficients.Inverse filtering factor a ' (t) obtained only differs constant times with a (t).(17) target of formula obtains a ' (t), convolution according to relational expression seismologic record and filtering factor obtains a (t) * b (t)=δ (t) (see (8) formula), here a ' (t) is δ (t)/b (t), a (t) differs constant times δ (t) with the inverse filtering factor, and application division rule can obtain a (t).Do not affect compact wavelet, propose high-resolution inverse filtering effect.After having asked for inverse filtering factor a (t), itself and seismologic record x (t) has been made to carry out convolution operation, that is:

S(t)＝a(t)*x(t)， (18)

Then S (t) is the new seismologic record exported after spiking deconvolution proposes process frequently.

(2) spectral imaging is carried out based on wavelet transformation spectral decomposition algorithm to carrying the seismic data frequently

Spectral imaging is in theory mainly according to the Tuning Principle of thin bed reflections.Be less than for quarter-wave thin layer for thickness, along with the increase of thickness of thin layer in time domain, seismic reflection amplitude increases gradually; When thickness of thin layer is increased to quarter-wave tuning thickness, reflection amplitude reaches maximal value, along with the increase reflection amplitude of thickness of thin layer reduces gradually.The maximum reflection amplitude of time domain correspond to the peak swing energy value of frequency field, according to Wavelet Transformation Algorithm, the seismologic record of time domain is converted to the seismologic record of frequency field here, is convenient to utilize the attenuation by absorption theory of oil-bearing reservoir to seismic wave energy to carry out the research of coalbed methane reservoir gas distribution prediction further.The principle of wavelet transformation is as follows:

If consecutive shock signal is x (t) ∈ L ²(R), its correspondence continuous wavelet transform be defined as:

${\mathrm{WT}}_x(a,b)=\frac{1}{\sqrt{a}}{\∫}_{-\∞}^{+\∞}x\left(t\right){\mathrm{\ψ}}^{*}\left(\frac{t-b}{a}\right)\mathrm{dt}={\∫}_{-\∞}^{+\∞}x\left(t\right){\mathrm{\ψ}}_{a,b}^{*}\left(t\right)\mathrm{dt},---\left(19\right)$

In formula, ψ (t) is called as morther wavelet (wavelet); Coefficient a, b are respectively scale parameter and displacement parameter, and a > 0.

Can be known by formula (19), ψ _{a, b}t () is the result of morther wavelet after telescopic moving, when a, b constantly change, just can produce family of functions { ψ, (t) }, be referred to as wavelet basis.Wherein, each ψ _{a, b}t () is called as a small echo.Classical class small echo can be divided into again Haar small echo, Morlet small echo, Mexican hat small echo and Gaussian small echo etc.ψ (t) meets formula (20):

$C_{\mathrm{\&psi;}}={\&Integral;}_{-\&infin;}^{+\&infin;}\frac{{\left|\mathrm{\&psi;}\left(\mathrm{\&omega;}\right)\right|}^2}{\mathrm{\&omega;}}\mathrm{d\&omega;}<\&infin;,$ C _ψ＞0， (20)

Its inverse transformation is defined as:

$x\left(t\right)=\frac{1}{C_{\mathrm{\&psi;}}}{\&Integral;}_{-\&infin;}^{+\&infin;}\frac{\mathrm{da}}{a^2}{\&Integral;}_{-\&infin;}^{+\&infin;}{\mathrm{WT}}_x(a,b){\mathrm{\&psi;}}_{a,b}\left(t\right)\mathrm{db},---\left(21\right)$

Namely the seismologic record of time domain can be transformed into the data of frequency field by through type (19), the analysis and research of seismic wave energy can be carried out, analyze the attenuation of seismic wave energy at the bottom of coal seam reservoirs top further.Frequency field then can complete by through type (21) to the conversion of time domain.

(3) coal seam reservoirs frequency decay is analyzed

Utilize and carry out the research of coal-seam gas forecasting techniques based on frequency decay analytical approach, carry at seismic data and frequently and on the basis of spectral imaging realize the analysis of coal seam reservoirs frequency decay:

Propose process frequently: input geological data x (t), frequency field seismologic record X (ω) is obtained according to formula (2), then obtain new seismologic record S (t) according to formula (3)-(18), be and carry the seismologic record frequently;

Spectral imaging process: according to the spectral imaging carrying out seismic data based on wavelet transformation spectral decomposition algorithm, its concrete grammar will carry seismologic record S (t) frequently and be converted into the data WT of frequency field according to formula (19) _x(a, b), namely obtains the frequency spectrum of seismic event, as shown in Figure 2.

As shown in Figure 4, frequency decay analysis needs to carry out three steps, and concrete steps are as follows:

1. top, coal seam, end spectrum analysis.By the frequency spectrum imaging analysis that spectral analysis algorithm and application of formula (19) are carried out based on wavelet transformation to top, coal seam, end seismic event, obtain seismic wave frequency spectrum P corresponding to top, coal seam, the end _on, P _under.Spectral analysis algorithm is as follows:.

${\mathrm{WT}}_x(a,b)=\frac{1}{\sqrt{a}}{\&Integral;}_{-\&infin;}^{+\&infin;}x\left(t\right){\mathrm{\&psi;}}^{*}\left(\frac{t-b}{a}\right)\mathrm{dt}={\&Integral;}_{-\&infin;}^{+\&infin;}x\left(t\right){\mathrm{\&psi;}}_{a,b}^{*}\left(t\right)\mathrm{dt}$

2. upper and lower stratum, coal seam frequency decay Characteristic Contrast.According to the result of spectrum analysis on the upper and lower stratum, coal seam of 1. middle calculating, in conjunction with the horizontal Distribution Characteristics in coal seam and the thickness, structural attitude etc. in coal seam, analyze the corresponding top, coal seam of coal seam zones of different, the change of end spectrum signature.

3. the calculating along layer of coal seam reservoirs frequency decay.Calculate upper and lower stratum, coal seam frequency decay change size, the spectral energy values corresponding with at the bottom of coal seam by the spectral energy values that top, coal seam is corresponding carries out subtracting each other process, obtains spectrum energy difference P _{previous-next}.Analyze the P that coal seam is corresponding _{previous-next}the corresponding relation in the size of value and coal seam (comprises Seams with Different Thickness, the different construction locations etc. in coal seam and P _{previous-next}the magnitude relationship of value), P _{previous-next}show that greatly the frequency decay in coal seam is large, otherwise then show that the frequency decay in coal seam is little.The P that frequency decay is large and corresponding _{previous-next}the region that value is larger is can the favourable band of coal-seam gas gassiness of prioritizing selection.

Because all can gain be given for the value of seismologic record in seismic processing process, the amount of gain determines according to the size of the earthquake-capturing record value of reality, the gain of giving in the little process of record value gathered is large, and then gain is little on the contrary, and bad description quantitatively goes out the size of final earthquake record value.Judge that spectrum difference size is according to the spectrum difference aggregate performance qualitative description in whole research range here, if there occurs frequency decay, the difference mean value in corresponding spectrum difference and whole research range has obvious difference.

The spectrum analysis on the upper and lower stratum of coal seam reservoirs effectively confirms the absorption of coal seam to seismic event high frequency, and compared with overlying strata, the radio-frequency component of coal seam underlying formation reduces, frequency band narrows.

The calculating of the upper and lower frequency decay of coal seam reservoirs, the when plane of frequency decay portrayed indicate the Absorption Characteristics of coal seam reservoirs to seismic wave energy: coal seam is thicker, and seismic wave energy decay is larger; Coal seam reservoirs air content is larger, and seismic wave energy decay is comparatively large, also reacts the feature of coal seam inhomogeneity simultaneously.

The spectrum analysis that Fig. 1 (a) is coal seam overlying strata, Fig. 1 (b) is coal seam sub-surface, by the band ratio underlying formation of overlying strata, comparative analysis visible coal seam compared with wide, radio-frequency component is many.The radio-frequency component of underlying formation disappearance is depicted in Fig. 1 (b).

Accompanying drawing 2 is for seismic event is through spectral decay feature behind coal seam, wherein upper figure is the differential chart of the upper and lower seismic wave frequency spectrum energy of spectrogram and single track coal seam of the upper and lower seismic event in coal seam in one-channel record, figure below is integral spectrum decay field section (wherein coal seam can be considered as the stacked general performance in the little coal seam of multilayer, and frequency decay is then the stacked general performance of multilayer little coal seam frequency decay) in coal seam.Artwork master Oxford gray represents the frequency spectrum below coal seam, Dark grey) long-pending large, represent bandwidth; Be the light grey frequency spectrum represented above coal seam in a solid timing window in artwork master, in artwork master, the black frequency spectrum represented on coal seam deducts the frequency spectrum under coal seam; Can find out, growth position, coal seam spectral decay is serious.

Accompanying drawing 3 is the planar distribution of the frequency decay feature through coal seam.In artwork master, light gray areas represents the large region of frequency decay; Dark gray areas represents the little region of frequency decay, and this is conducive to the spatial prediction of coalbed methane reservoir.

An embodiment of the inventive method is as follows:

The first step, analyzes the seismic data feature of distributive province, coal seam, and utilize spiking deconvolution to put forward technology frequently and improve seismic resolution, widen seismic band, putting forward the seismic data dominant frequency frequently can reach 50Hz.

Second step, the spectral decomposition of carrying out based on wavelet transformation calculates.The present invention have selected the Spectral Imaging Technology based on wavelet transformation, achieves the conversion of time domain geological data to frequency field geological data.Window problem when this technology has an elimination compared with conventional spectral technology, exact picture time and frequency location, portraying reservoir thickness and indicate the advantage of the horizontal uncontinuity feature in stratum, laying a good foundation for carrying out frequency decay analysis better.

3rd step, coal seam reservoirs frequency decay is analyzed.By analyzing frequency spectrum and the frequency decay signature analysis on upper and lower stratum, coal seam, obtain the frequency decay Variation Features of coal output layer: 1. seismic event is through the decay that there occurs energy during coal seam, high-frequency energy absorbs serious, and seismic event frequency band narrows; 2. the position frequency decay that coal seam is thicker is relatively large; 3. the horizontal change of frequency decay shows the Lateral heterogeneity in coal seam; 4. the position that frequency decay is larger is coal seam reservoirs gassiness Favorable Zones.

Technique scheme is one embodiment of the present invention, for those skilled in the art, on the basis that the invention discloses application process and principle, be easy to make various types of improvement or distortion, and the method be not limited only to described by the above-mentioned embodiment of the present invention, therefore previously described mode is just preferred, and does not have restrictive meaning.

Claims (4)

1. based on a coal-seam gas Forecasting Methodology for frequency decay, it is characterized in that: described method comprises:

(1) geological data x (t) is inputted;

(2) spiking deconvolution is carried out to geological data x (t) and carry seismologic record S (t) frequently processing and obtain carrying frequently;

(3) carry out wavelet transformation spectral decomposition obtain top, coal seam, the frequency spectrum at the end to carrying seismologic record S (t) frequently;

(4) obtain top, coal seam, the spectrum difference at the end, carry out the comparative analysis of coal seam and spectrum difference, obtain the frequency decay result of coal seam reservoirs;

(5) the frequency decay result of coal seam reservoirs is exported.

2. the coal-seam gas Forecasting Methodology based on frequency decay according to claim 1, is characterized in that: described step (2) is achieved in that

Input geological data x (t), obtains frequency field seismologic record X (ω) according to formula (2), then obtains new seismologic record S (t) according to formula (18), is and carries the seismologic record frequently;

x(t)＝b(τ)*ξ(t)， (2)

Fourier transform is asked to both sides, then obtains frequency field seismologic record X (ω):

X(ω)＝B(ω)*ξ(ω)，(3)

X (ω), B (ω) ξ (ω) are respectively the frequency spectrum of seismic spectrum, wavelet spectrum and reflection coefficient;

S(t)＝a(t)*x(t)， (118

Wherein, a (t) is the inverse filtering factor, is tried to achieve by (17) formula:

$\left[\begin{array}{cccc}{r}_{\mathrm{xx}}\left(0\right)& r_{\mathrm{xx}}\left(1\right)& ...& r_{\mathrm{xx}}\left(m\right)\\ r_{\mathrm{xx}}\left(1\right)& r_{\mathrm{xx}}\left(0\right)& ...& r_{\mathrm{xx}}(m-1)\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ r_{\mathrm{xx}}\left(m\right)& r_{\mathrm{xx}}(m-1)& ...& r_{\mathrm{xx}}\left(0\right)\end{array}\right]\left[\begin{array}{c}{r}_{\mathrm{xx}}\left(0\right)\\ a\left(1\right)\\ .\\ .\\ .\\ a\left(m\right)\end{array}\right]\left[\begin{array}{c}a\left(0\right)\\ a\left(1\right)\\ .\\ .\\ .\\ a\left(m\right)\end{array}\right]=\left[\begin{array}{c}1\\ 0\\ .\\ .\\ .\\ 0\end{array}\right]---\left(17\right)$

(17) r in formula _xx(τ) be the auto-correlation of seismologic record x (t).

3. the coal-seam gas Forecasting Methodology based on frequency decay according to claim 2, is characterized in that: described step (3) is achieved in that

Utilize formula (19) respectively to the frequency spectrum imaging analysis that top, coal seam, seismologic record S (t) at the end carry out based on wavelet transformation, obtain frequency spectrum P corresponding to top, coal seam, the end _on, P _under;

${\mathrm{WT}}_x(a,b)=\frac{1}{\sqrt{a}}{\&Integral;}_{-\&infin;}^{+\&infin;}x\left(t\right){\mathrm{\&psi;}}^{*}\left(\frac{t-b}{a}\right)\mathrm{dt}={\&Integral;}_{-\&infin;}^{+\&infin;}x\left(t\right){\mathrm{\&psi;}}_{a,b}^{*}\left(t\right)\mathrm{dt}---\left(19\right)$

Calculate P _ontime, replace the x (t) in formula (19) to calculate with seismologic record S (t) on top, coal seam, calculate P _undertime, replace the x (t) in formula (19) to calculate with seismologic record S (t) at the bottom of coal seam.

4. the coal-seam gas Forecasting Methodology based on frequency decay according to claim 3, is characterized in that: described step (4) is achieved in that

By frequency spectrum P corresponding for top, coal seam _onthe frequency spectrum P corresponding with at the bottom of coal seam _undercarry out subtracting each other obtaining spectrum difference P _{on -under}, then analyze P corresponding to coal seam _{previous-next}the size of value and the corresponding relation in coal seam, P _{previous-next}value show that greatly the frequency decay in coal seam is large, otherwise then show that the frequency decay in coal seam is little; The P that frequency decay is large and corresponding _{previous-next}the region that value is larger is the favourable band of coal-seam gas gassiness of prioritizing selection.