A weight coefficient constrained spectral decomposition prediction quality factor method and system

Abstract

The application discloses a weight coefficient constrained spectral decomposition prediction quality factor method and system, and the weight coefficient constrained spectral decomposition prediction quality factor method comprises the following steps: constructing a weight coefficient of a frequency component according to an average amplitude spectrum of seismic data; determining a target function of a quality factor based on the weight coefficient; and determining the quality factor based on the target function of the quality factor. The application adopts the energy of the average amplitude spectrum to determine the weight coefficient of different frequency components, has stronger robustness to noise, and test results also show that the greater the frequency width is, the more helpful to improve the accuracy of quality factor calculation. The application of actual data proves that there is a better relationship between the low quality factor and the high gas production rate.

Description

Technical Field

[0001] This invention belongs to the field of geophysical exploration technology, and specifically relates to a method and system for predicting quality factors by spectral decomposition with weighted coefficient constraints. Background Technology

[0002] Fluid flow and friction between grains convert the energy of seismic waves into heat, leading to viscoelastic attenuation of seismic data. The quality factor Q is used to measure the viscoelastic attenuation of seismic waves propagating underground. Viscoelastic attenuation of seismic waves is commonly used for reservoir prediction. It is particularly sensitive to gas-bearing media; abnormally high attenuation (low quality factor) can serve as an important indicator of gas-bearing reservoirs. Variations in viscoelastic attenuation with azimuth are also used to analyze fracture development direction in fractured media. Therefore, viscoelastic attenuation of seismic waves is an important seismic property characterizing subsurface features.

[0003] Researchers have proposed methods for extracting the quality factor Q in both the time and frequency domains. Time-domain algorithms utilize pulse amplitude information. However, seismic pulses are frequently affected by scattering, noise, and other factors. Therefore, obtaining the quality factor Q in the time domain is unstable. Frequency-domain algorithms include the spectral ratio method (SR), centroid frequency shift method (CFS), and peak frequency shift method (PFS). Researchers have made significant efforts to improve the stability of quality factor calculations in the frequency domain (SR, CFS, and PFS). However, for SR-based algorithms, small changes in the high and low frequency components of the amplitude spectrum can drastically alter the quality factor Q. Therefore, selecting an appropriate frequency band and improving the spectral ratio method are crucial. Unfortunately, selecting an appropriate frequency band is also very difficult in practical applications. The center frequency method, peak frequency method, improved center frequency method, and improved peak frequency method all require specific shapes of the amplitude spectrum at the source and receiver locations. However, in practical applications, selecting a suitable amplitude spectrum shape is extremely challenging. Summary of the Invention

[0004] To address the above problems, this invention discloses a method for predicting quality factors using spectral decomposition with weighted coefficient constraints, comprising:

[0005] Weighting coefficients for frequency components are constructed based on the average amplitude spectrum of earthquake data;

[0006] The objective function for determining the quality factor is based on the weighting coefficients.

[0007] The quality factor is determined based on the objective function of the quality factor.

[0008] Furthermore, the weighting coefficients for constructing the frequency components based on the average amplitude spectrum of seismic data are determined through the following steps:

[0009] The average amplitude spectrum is obtained by transforming the seismic data from the time domain to the frequency domain using Fourier transform.

[0010] The weighting coefficients of the frequency components are determined based on the energy percentage of the average amplitude spectrum.

[0011] Furthermore, the weighting coefficients are determined by the following formula:

[0012] w(f) = w avg (f-(f p -f p_avg ))

[0013] Where w(f) is the weighting coefficient; w avg f is the average coefficient; f is the frequency; f p f is the peak frequency at position t0; p_avg This represents the peak frequency of the average amplitude spectrum of the earthquake data.

[0014] Furthermore, the average coefficient is determined by the following formula:

[0015]

[0016] Among them, S avg (f) represents the average amplitude spectrum of the seismic data; F1 represents the minimum bandwidth; and F2 represents the maximum bandwidth.

[0017] Furthermore, the objective function for the quality factor is as follows:

[0018]

[0019] Where G is the objective function; α is the absorption coefficient; C is the frequency-independent attenuation term; and V is a function of frequency.

[0020] Furthermore, the function relating to frequency is determined by the following formula:

[0021]

[0022] Where S1(f) is the amplitude spectrum at the source location; S2(f) is the amplitude spectrum at the receiver location; t 1,2 For the travel time from the source to the receiving location; α 1,2 denoted as , where is the absorption coefficient from the source to the receiver; e is the natural constant.

[0023] A weighted coefficient-constrained spectral decomposition prediction quality factor system includes:

[0024] Weighting coefficient unit, used to construct the weighting coefficients of frequency components based on the average amplitude spectrum of seismic data;

[0025] A determining unit is used to determine the objective function of the quality factor based on the weighting coefficients;

[0026] The quality factor unit is used to determine the quality factor based on the objective function of the quality factor.

[0027] Furthermore, the weighting coefficient unit is specifically used for:

[0028] The average amplitude spectrum is obtained by transforming the seismic data from the time domain to the frequency domain using Fourier transform.

[0029] The weighting coefficients of the frequency components are determined based on the energy percentage of the average amplitude spectrum.

[0030] Furthermore, the weighting coefficients are determined by the following formula:

[0031] w(f) = w avg (f-(f p -f p_avg ))

[0032] Where w(f) is the weighting coefficient; w avg f is the average coefficient; f is the frequency; f p f is the peak frequency at position t0; p_avg This represents the peak frequency of the average amplitude spectrum of the earthquake data.

[0033] Furthermore, the objective function for the quality factor is as follows:

[0034]

[0035] Where G is the objective function; α is the absorption coefficient; C is the frequency-independent attenuation term; and V is a function of frequency.

[0036] Compared with the prior art, the beneficial effects of the present invention are: the weighting coefficients of different frequency components are determined by the energy of the average amplitude spectrum; it has stronger robustness to noise and is less sensitive to the selection of seismic bandwidth; test results also show that a larger bandwidth helps to improve the accuracy of quality factor calculation; the application of actual data confirms that there is a good relationship between low quality factor and high gas production rate.

[0037] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the methods / processes pointed out in the description, claims, and drawings. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 A seismic profile acquired on land according to an embodiment of the present invention is shown;

[0040] Figure 2 for Figure 1 The first seismic trace and its corresponding time spectrum;

[0041] Figure 3 for Figure 1 The amplitude spectrum of the first seismic trace at t0 = 0.74s;

[0042] Figure 4 The average amplitude spectrum according to an embodiment of the present invention is shown;

[0043] Figure 5 The amplitude spectra at different times are shown according to embodiments of the present invention;

[0044] Figure 6 The quality factor Q model and the corresponding post-stack seismic profile according to an embodiment of the present invention are shown.

[0045] Figure 7 The temporal spectra of different seismic traces in the post-stack seismic profile of the quality factor Q model according to an embodiment of the present invention are shown.

[0046] Figure 8 The quality factor for a bandwidth of 10Hz-90Hz according to an embodiment of the present invention is shown;

[0047] Figure 9 The quality factor for a bandwidth of 20Hz-60Hz according to an embodiment of the present invention is shown;

[0048] Figure 10 The quality factor for a bandwidth of 30Hz-50Hz according to an embodiment of the present invention is shown;

[0049] Figure 11 A seismic profile of two production wells and one dry well is shown according to an embodiment of the present invention.

[0050] Figure 12 The temporal spectrum of seismic trace data for different well locations according to embodiments of the present invention is shown;

[0051] Figure 13 Based on Figure 11The quality factor of the seismic profile calculation. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] The amplitude spectrum relationship between the source and receiver locations in existing technologies can be expressed as:

[0054]

[0055] Where S1(f) and S2(f) represent the amplitude spectra at the source and receiver locations, respectively; the coefficient C is a frequency-independent attenuation term; f is the frequency; e is the natural constant; t 1,2 Q is the distance traveled from the epicenter to the receiving location. 1,2 This is the constant quality factor (i.e., a constant quality factor) from the source to the receiver location.

[0056] The spectral ratio method performs a logarithmic operation on formula (1) to obtain the following formula:

[0057]

[0058] Where b = log(C), and another Substituting into formula (2), the objective function G for calculating the quality factor is expressed as:

[0059]

[0060] Here, (F1, F2) represents the bandwidth required to calculate the quality factor. F1 is the minimum bandwidth, and F2 is the maximum bandwidth. The least squares method is used to solve the quadratic function in formula (3). As shown in formula (3), the spectral ratio method does not require any assumptions about the shape of the seismic data spectrum. However, Blias points out that the high-frequency components with smaller amplitudes have smaller errors. This will introduce a large error; therefore, errors in the lower frequency components of the amplitude spectrum will make the calculation of the quality factor unstable. Therefore, Blias proposes the following objective function G:

[0061]

[0062] The constant C in formula (4) can be obtained by solving the following equation:

[0063]

[0064]

[0065] Formula (4) is a quadratic optimization problem that can be solved by the least squares method.

[0066] The method in formula (4) is called the modified spectral ratio method. However, Blias points out that given the range of absorption coefficient α, the value of the minimization function G(α,C) is determined, and this method is still relatively stable. As can be seen from formula (4), a bandwidth for calculating the quality factor needs to be defined. Different bandwidths will produce different quality factors. Therefore, choosing an appropriate bandwidth is crucial for calculating the accurate quality factor.

[0067] The low-frequency and high-frequency components of seismic data amplitude spectra typically have low values. It is well known that noise in seismic data can easily mask the geological information carried by low-amplitude frequency components. Therefore, frequency components with higher amplitude (energy) should have a larger weight in the quality factor calculation. However, as shown in formula (4), each frequency component has the same weight in the objective function, which is unreasonable in practical applications.

[0068] Therefore, this invention proposes a method for predicting quality factors using spectral decomposition with weighted coefficient constraints, comprising the following steps:

[0069] Weighting coefficients for frequency components are constructed based on the average amplitude spectrum of earthquake data;

[0070] The objective function for determining the quality factor is based on the weighting coefficients.

[0071] The quality factor is determined based on the objective function of the quality factor.

[0072] The weighting coefficients for the frequency components, constructed from the average amplitude spectrum of seismic data, are determined through the following steps:

[0073] The average amplitude spectrum is obtained by transforming the seismic data from the time domain to the frequency domain using Fourier transform.

[0074] The weighting coefficients of the frequency components are determined based on the energy percentage of the average amplitude spectrum. The energy percentage refers to the ratio of the amplitude intensity of different frequency components to the intensity of the average amplitude spectrum.

[0075] like Figure 1 As shown, a seismic profile acquired on land is presented. Figure 2 As shown, Figure 2 (a) and Figure 2 (b) respectively Figure 1 The first seismic trace and corresponding time spectrum of the seismic profile. Figure 3 for Figure 2 (a) shows the amplitude spectrum of the first seismic trace at t0 = 0.74 s. It is worth noting that... Figure 3 The amplitude spectrum in the image shows a "depression" around 35 Hz. The amplitude (energy) of the frequency components near this "depression" is relatively smaller than the amplitude of the peak frequency, even though the depressor and peak frequencies are very close. The interference of reflected waves is... Figure 3 One of the most likely reasons for the depression phenomenon shown is that if only the absolute value of the amplitude spectrum is considered, the frequency components near the depression frequency have a smaller weight, while those near the peak frequency have a larger weight. Therefore, there might be an unreasonable weighting coefficient function that changes abruptly with frequency.

[0076] Figure 4 The solid line and the dashed line curves are respectively Figure 2 (a) shows the average amplitude spectrum of the seismic trace and Figure 1 The average amplitude spectrum of the seismic profile is shown. It is worth noting that the average amplitude spectrum of a seismic trace still has a small dip. The average amplitude spectrum of the seismic profile changes continuously and smoothly with frequency. Statistical wavelet estimation methods assume that the wavelet amplitude spectrum is similar to the average amplitude spectrum of the seismic data. In this invention, it is assumed that the shape of the wavelet amplitude spectrum does not change significantly. Then, the weighting coefficients for each frequency component are constructed using the average amplitude spectrum. Considering that the peak frequency may change with the time axis, the weight of the frequency component at t0 can be expressed as:

[0077] w(f) = w avg (f-(f p -f p_avg (7-1)

[0078]

[0079] Where w(f) is the weighting coefficient; w avg f is the average coefficient; f is the frequency; f p f is the peak frequency at position t0; p_avg S represents the peak frequency of the average amplitude spectrum of the seismic data; avg (f) represents the average amplitude spectrum of the seismic data.

[0080] Therefore, the objective function for the optimal quality factor scan can be expressed as:

[0081]

[0082]

[0083] Where α is the absorption coefficient; V is a function of frequency; C is an attenuation term independent of frequency; α 1,2 The absorption coefficient is the distance from the seismic source to the receiving location.

[0084] In practical applications, it is difficult to obtain the amplitude spectrum of the source wavelet. Therefore, the average amplitude spectrum of shallow continuous seismic phase axes is usually chosen as the reference amplitude spectrum. However, it is recommended that the reference amplitude spectrum be changed when calculating the quality factor of seismic phase axes reflected from deeper layers. Figure 1 The seismic profile shown indicates the presence of a continuous seismic phase axis at 1.6s. Figure 5 The solid and dashed colored curves in the image are respectively Figure 1 The amplitude spectra at 1.6s and 1.85s in the seismic profile show a significant difference between the two. Therefore, careful selection of the bandwidth may be necessary to calculate the quality factor. Otherwise, using the amplitude spectrum at 1.6s as a reference for the amplitude spectrum at 1.85s may result in an unreasonable quality factor. Classical convolution theory indicates that seismic reflections occur at both the top and bottom of rock strata. Therefore, this invention proposes first establishing a subsurface layered model, and then using the average amplitude spectrum at the top of each layer as a reference amplitude spectrum. The reflection data at the top of the subsurface layered model are transformed from the time domain to the frequency domain using a Fourier transform to obtain the average amplitude spectrum and bandwidth.

[0085] Figure 6 (a) and Figure 6 (b) shows the quality factor Q model and its corresponding post-stack seismic profiles. For example... Figure 6 As shown in (a), a gas-bearing reservoir exists at the location indicated by the white arrow. Figure 6 (b) shows a seismic profile with 101 data channels, each with 301 time sampling points at a sampling rate of 2 ms. The synthesized seismic profile uses the Ricker wavelet with a dominant frequency of 40 Hz. Gaussian noise was added to the post-stack seismic data, resulting in a signal-to-noise ratio of 5. Figure 7 (a) is Figure 6 (b) is the amplitude spectrum of the first data passage of the seismic profile, and 7(b) is... Figure 6 (b) Amplitude spectrum of the 6th data from the seismic profile. Compared to shallow reflection data, the amplitude spectrum of deep reflection data corresponds to low-frequency components.

[0086] The quality factor scan range is 100-200, with an interval of 1. In the quality factor Q model, there is no reflection at the top and bottom layers, so the quality factor Q value of the first layer is the same as that of the second layer, and the Q value of the last layer is the same as that of the penultimate layer. Figure 8 (a) and (b) are the quality factors Q calculated using the modified spectral ratio method and weighted coefficient constraint spectral decomposition prediction method, respectively. The data bandwidth used to calculate the quality factors is 10Hz-90Hz. Figure 9 (a) and 9(b) respectively use the modified spectral ratio method and the spectral decomposition method with weighted coefficient constraints to predict the quality factor for data with a bandwidth of 20Hz-60Hz to calculate the quality factor Q. Figure 10(a) and (b) use a modified spectral ratio method and a weighted coefficient-constrained spectral decomposition method, respectively, to calculate the quality factor Q for data with a bandwidth of 30Hz-50Hz. Due to waveform interference, the quality factors calculated by either method are inaccurate. However, both methods yield relatively accurate air layer quality factors. Figure 8 , 9 10 and 10 respectively Figure 6 (a) Compare and synthesize Figure 8 , 9 Figures 10 and 10 show that the results calculated by the weighted coefficient-constrained spectral decomposition method for predicting the quality factor are closer to the results calculated by the modified spectral ratio method than those calculated by the weighted coefficient-constrained method. Figure 6 (a) shows that the spectral decomposition method with weighted coefficient constraints is more accurate than the modified spectral ratio method, indicating that the spectral decomposition method with weighted coefficient constraints has better robustness to noise. Figure 8 , 9 The values ​​of 10 and 10 also indicate that for the modified spectral ratio method and the spectral decomposition prediction quality factor method with weight coefficient constraints, the wider the seismic data bandwidth, the higher the accuracy of the calculated quality factor (i.e., compared with...). Figure 6 (a) The closer they are.

[0087] Taking terrestrial seismic data as an example, the reservoir lithology is limestone. Figure 11 This is a seismic profile containing 1300 data points. The profile passes through two production wells (A and C) and one dry well (B). The seismic phase axis indicated by the black arrow represents the producing layer. Drilling results show that the production wells A and C have a production rate of 45,000 m³ / s. 3 / d、87000m 3 / d. Figure 11 The three vertical curves are sonic logging curves (unit: m / s). The sonic logging curve values ​​increase from left to right. The logging interpretation results show that there is a good correlation between high gas production and low sonic velocity. Figure 12 (a), (b), and (c) are the time-frequency spectra of the seismic traces located at production well A, dry well B, and production well C, respectively. The white arrows indicate the two-way travel time at the producing layer location, suggesting that the seismic data at production wells A and C have lower frequency components than those at dry well B. Figure 13 The quality factors calculated using the method of this invention are shown in the figure. The quality factors near production wells A and C are relatively low, demonstrating a good correlation between low quality factors and production wells. Low quality factors represent strong absorption, indicating hydrocarbon growth. This invention improves upon the traditional spectral ratio method by weighting the amplitude spectrum according to the energy percentage weight. The energy percentage weight ensures a larger weight for frequency components with strong amplitude values, making the quality factor calculation more stable and further reducing the influence of bandwidth on the quality factor. The calculated quality factor can serve as an indicator factor for reservoir prediction.

[0088] Based on the above-mentioned method for predicting quality factors using spectral decomposition with weighted coefficient constraints, this invention proposes a system for predicting quality factors using spectral decomposition with weighted coefficient constraints, comprising:

[0089] Weighting coefficient unit, used to construct the weighting coefficients of frequency components based on the average amplitude spectrum of seismic data;

[0090] A determining unit is used to determine the objective function of the quality factor based on the weighting coefficients;

[0091] The quality factor unit is used to determine the quality factor based on the objective function of the quality factor.

[0092] Weight coefficient unit, specifically used for:

[0093] The average amplitude spectrum is obtained by transforming the seismic data from the time domain to the frequency domain using Fourier transform.

[0094] The weighting coefficients of the frequency components are determined based on the energy percentage of the average amplitude spectrum.

[0095] The weighting coefficients are determined using the following formula:

[0096] w(f) = w avg (f-(f p -f p_avg ))

[0097] Where w(f) is the weighting coefficient; w avg f is the average coefficient; f is the frequency; f p f is the peak frequency at position t0; p_avg This represents the peak frequency of the average amplitude spectrum of the earthquake data.

[0098] The objective function for the quality factor is as follows:

[0099]

[0100] Where G is the objective function; α is the absorption coefficient; C is the frequency-independent attenuation term; and V is a function of frequency.

[0101] The present invention proposes a method and system for predicting quality factors by weighted coefficient constraint spectral decomposition, which uses the energy of the average amplitude spectrum to determine the weighting coefficients of different frequency components; it has stronger robustness to noise and is less sensitive to the selection of seismic bandwidth; test results also show that broadband data helps to improve the accuracy of quality factor calculation; the application of actual data confirms that there is a good relationship between low quality factor and high gas production rate.

[0102] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting quality factors using spectral decomposition with weighted coefficient constraints, characterized in that, include: Weighting coefficients for frequency components are constructed based on the average amplitude spectrum of earthquake data; The objective function for determining the quality factor is based on the weighting coefficients. The quality factors are determined based on the objective function of the aforementioned quality factors. The weighting coefficients for constructing the frequency components based on the average amplitude spectrum of seismic data are determined through the following steps: The average amplitude spectrum is obtained by transforming the seismic data from the time domain to the frequency domain using Fourier transform. The weighting coefficients of the frequency components are determined based on the energy percentage of the average amplitude spectrum; The weighting coefficients are determined by the following formula: in, For weighting coefficients; The average coefficient; For frequency; The peak frequency at position t0; This represents the peak frequency of the average amplitude spectrum of the earthquake data.

2. The method for predicting quality factors by spectral decomposition with weighted coefficient constraints according to claim 1, characterized in that, The objective function for the quality factor is as follows: in, The objective function is... The absorption coefficient; This is an attenuation term independent of frequency. It is a function of frequency; This is the minimum bandwidth value; This represents the maximum bandwidth.

3. The method for predicting quality factors by spectral decomposition with weighted coefficient constraints according to claim 2, characterized in that, The function relating frequency is determined by the following formula: in, The amplitude spectrum at the epicenter location; The amplitude spectrum at the receiving location; For the time it takes for the earthquake to travel from the epicenter to the receiving location; The absorption coefficient is the distance from the seismic source to the receiving location. It is a natural constant.

4. A spectral decomposition prediction quality factor system with weighted coefficient constraints, characterized in that, include: Weighting coefficient unit, used to construct the weighting coefficients of frequency components based on the average amplitude spectrum of seismic data; A determining unit is used to determine the objective function of the quality factor based on the weighting coefficients; A quality factor unit is used to determine the quality factor based on the objective function of the quality factor. The weight coefficient unit is specifically used for: The average amplitude spectrum is obtained by transforming the seismic data from the time domain to the frequency domain using Fourier transform. The weighting coefficients of the frequency components are determined based on the energy percentage of the average amplitude spectrum; The weighting coefficients are determined by the following formula: in, For weighting coefficients; The average coefficient; For frequency; The peak frequency at position t0; This represents the peak frequency of the average amplitude spectrum of the earthquake data.

5. The spectral decomposition prediction quality factor system with weighted coefficient constraints according to claim 4, characterized in that, The objective function for the quality factor is as follows: in, The objective function is... The absorption coefficient; This is an attenuation term independent of frequency. It is a function of frequency; This is the minimum bandwidth value; This represents the maximum bandwidth.

Images