Frequency spectrum consistency Q value inversion method under phase constraint

Abstract

The invention discloses a frequency spectrum consistency Q value inversion method under phase constraint, and belongs to the field of exploration geophysical research. According to the method, a multi-attribute joint inversion strategy is introduced, seismic signals received at different moments or positions are reversely propagated back to a seismic source position, an objective function is established by using amplitude and phase consistency of seismic waves, and then a Q value is inversed. According to the new method, hypothesis of seismic source wavelet characteristics is not needed, overall changes of seismic signals are fully considered, and the Q value can be accurately estimated while data continuity and fidelity are kept. Simulation data show that the method has high estimation precision and stability under the conditions of noise influence and wavelet interference, and the Q value estimated by the method can effectively improve seismic data resolution in practical application.

Description

Technical Field

[0001] The invention belongs to the field of exploration geophysical research, and in particular relates to a spectrum consistency Q value inversion method under phase constraint. Background Art

[0002] The propagation of seismic waves in underground media is affected by the viscoelastic properties of the medium, causing its energy to gradually decay with increasing propagation distance, manifested as a decrease in amplitude and a change in phase. In order to quantify this phenomenon, the quality factor (Q value) is usually used to describe it. As an inherent characteristic parameter of rock, the Q value has important application value in seismic exploration. By accurately estimating the Q value, the resolution of seismic data can be improved, thereby more effectively solving deep geological problems. The Q value is not only closely related to the physical properties of the rock, such as heterogeneity and permeability, but also has a significant relationship with the oil and gas content of the reservoir and the degree of development of pores and fractures. Therefore, accurate estimation of the Q value is of great significance for the reasonable evaluation of geological targets.

[0003] At present, the estimation methods of Q value are mainly divided into three categories: time domain, frequency domain and time-frequency domain. The time domain method estimates the Q value by analyzing the changes in amplitude and width of seismic waves during propagation, but the amplitude information is easily affected by factors such as geometric expansion and propagation loss, resulting in low accuracy. Frequency domain methods, such as the centroid frequency shift method and the spectral ratio method, remove the influence of factors such as geometric expansion by analyzing the changes in spectral characteristics, and have become the most widely used Q value estimation methods. However, the centroid method assumes that the source waveform is Gaussian, which is easily interfered by noise and has poor noise resistance; while the spectral ratio method is limited by frequency band selection, and it is very difficult to select a suitable frequency band in practical applications. The time-frequency domain method inverts the Q value through time-frequency analysis technology, including fixed window function transformation and variable window function transformation methods, which can improve the accuracy and stability of the estimation to a certain extent, but its computational complexity is high and has high requirements on data quality.

[0004] In recent years, optimization theory has been introduced into Q-value inversion in order to improve the accuracy and stability of estimation. However, current research focuses on single signal features or local spatial variation of seismic wave features, and the estimation results under low signal-to-noise ratio conditions are still unstable. Therefore, developing a Q-value estimation method that can integrate amplitude and phase information and is suitable for complex geological conditions has important research significance and broad application prospects. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes a spectral consistency Q value inversion method under phase constraint, which can significantly improve the estimation accuracy and stability of the Q value. Starting from the theoretical basis of seismic wave propagation attenuation, this method introduces a multi-attribute joint inversion strategy, establishes the objective function using the amplitude and phase consistency of seismic waves, and solves it using the gradient descent method, and finally obtains a more accurate Q estimation value. Compared with the spectral ratio method and the spectral consistency method without phase constraint, this method has higher estimation accuracy and stability under the conditions of noise influence and wavelet interference.

[0006] In order to achieve the above object, the present invention adopts the following technical solution to achieve it: a spectrum consistency Q value inversion method under phase constraint, comprising the following steps:

[0007] Step 1: Perform Fourier transform on the seismic signals received at different times or locations, and compensate each data to the source location;

[0008] Step 2: Extract the amplitude and phase information of the compensated data, and use the compensated mean to establish the spectrum consistency Q value inversion objective function I(Q) under phase constraints;

[0009] Step 3: Use the gradient descent method to solve the objective function I(Q).

[0010] Preferably, the specific process of compensating the seismic signals picked up at different times or positions in step 1 is as follows:

[0011] a. According to the wave theory of viscoelastic media, the received Nth seismic wave spectrum R N The frequency domain expression of (f) is:

[0012]

[0013] in, represents the source spectrum, x j represents the propagation distance from the j-1th receiving wave to the jth receiving wave, f represents the frequency, v j (f) represents the phase velocity, α j (f) is the attenuation coefficient;

[0014] b. In the Kolsky-Futterman phase velocity dispersion model, the attenuation coefficient and phase velocity have the following corresponding relationship:

[0015]

[0016]

[0017] Where f0 represents the reference frequency, v j (f0) is the phase velocity when the reference frequency is f0, Q jis the absorption attenuation factor at point j;

[0018] c. Using the Taylor expansion of the exponential function, when all frequencies satisfy When , formula (1) can be expressed as:

[0019]

[0020] in, Indicates travel time;

[0021] d. According to formula (4), the seismic signals received at different locations are sent to the source location The compensation formula for compensation is:

[0022]

[0023] Preferably, the objective function I(Q) process established in step 2 is:

[0024] a. Define the reference amplitude spectrum separately and the reference phase spectrum for:

[0025]

[0026] Among them, [f1, f2] and [f3, f4] represent the amplitude frequency band and phase frequency band related to the Q value, respectively, and s N and Respectively represent the amplitude and phase of the Nth received signal after compensation;

[0027] b. Define the spectral consistency Q value inversion objective function I(Q) under phase constraint as:

[0028]

[0029] Where Q = {Q1, Q2, ... Q N}, β represents the weighting coefficient.

[0030] Preferably, the gradient calculation formula and update rule for solving the objective function using the gradient descent method in step 3 are defined as follows:

[0031]

[0032] Among them, t represents the number of iterations and η represents the step size.

[0033] The present invention has the following beneficial effects:

[0034] The present invention proposes a spectrum consistency Q value inversion method under phase constraints, introduces a multi-attribute joint inversion strategy, and reversely propagates the seismic signals received at different times or locations back to the source location, uses the amplitude and phase consistency of the seismic waves to establish the objective function, and uses the gradient descent method to solve it. This method does not need to assume the characteristics of the source sub-waves, fully considers the overall changes of the seismic signal, and can accurately estimate the Q value while maintaining data continuity and fidelity.

[0035] This paper proposes a spectral consistency Q value inversion method under phase constraint. Synthetic data show that compared with the spectral consistency inversion method and spectral ratio method without phase constraint, this method has higher noise robustness and anti-wavelet interference ability. By estimating and compensating the Q value of two-dimensional seismic profiles, the results show that the Q value estimated by this method can effectively improve the resolution of seismic data in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is the velocity and Q value model, and the noise-free VSP data generated under the model;

[0037] Figure 2 (a) The Q-value estimation result of the spectrum-consistent Q-value inversion method under phase constraint applied to the noise-free VSP data;

[0038] Figure 2 (b) The Q-value estimation result of the spectrum-consistent Q-value inversion method without phase constraint applied to the noise-free VSP data;

[0039] Figure 2 (c) is the Q value estimation result of the spectral ratio method applied to the noise-free VSP data;

[0040] Figure 3 (a)-(c) are respectively Figure 1 VSP data with 5%, 10%, and 15% Gaussian random noise added;

[0041] Figure 3 (d)-(f) are the results of 200 independent tests under three noise intensity conditions using the spectral consistency Q value inversion method under phase constraints, the spectral consistency Q value inversion method without phase constraints, and the spectral ratio method;

[0042] Figure 4 (a) is the wedge-shaped model

[0043] Figure 4 (b) is the wedge-shaped model after attenuation (Q=50);

[0044] Figure 5(a) is the estimation result of the attenuated wedge model using the spectral consistency Q-value inversion method under phase constraint;

[0045] Figure 5 (b) is the estimation result of the attenuated wedge model using the spectrum consistent Q value inversion method without phase constraint;

[0046] Figure 5 (c) is the estimation result of the attenuation wedge model using the spectral ratio method;

[0047] Figure 6 (a) and (b) are the seismic sections before and after attenuation compensation, respectively. DETAILED DESCRIPTION

[0048] A spectrum consistency Q value inversion method under phase constraint will be described in detail below in conjunction with the accompanying drawings of the present invention.

[0049] The present invention proposes a method for inverting the spectral consistency Q value under phase constraint, comprising the following steps:

[0050] Step 1: According to the wave theory of viscoelastic media, the received Nth seismic wave spectrum R N The frequency domain expression of (f) is:

[0051]

[0052]

[0053] in, represents the source spectrum, x j represents the propagation distance from the j-1th receiving wave to the jth receiving wave, f represents the frequency, v j (f) represents the phase velocity, α j (f) is the attenuation coefficient, f0 represents the reference frequency, v j (f0) is the phase velocity when the reference frequency is f0, Q j is the absorption attenuation factor at point j;

[0054] According to Taylor expansion, when all frequencies satisfy When , formula (1) can be expressed as:

[0055]

[0056] in, Indicates travel time;

[0057] According to formula (4), the seismic signals received at different locations can be transmitted to the source location. The compensation formula for compensation is:

[0058]

[0059] Step 2: Extract the amplitude and phase information of the compensation data and define the reference amplitude and reference phase Then the compensated mean is used to establish the spectrum consistency Q value inversion objective function I(Q) under phase constraint:

[0060]

[0061] Among them, [f1, f2] and [f3, f4] represent the amplitude frequency band and phase frequency band related to the Q value, respectively, and s N and represent the amplitude and phase of the Nth received signal after compensation, Q = {Q1, Q2, ... Q N}, β represents the weighting coefficient.

[0062] Step 3: Select the appropriate frequency band, weighting coefficient β, number of iterations t and step size η, and use the gradient descent method to solve the objective function I(Q):

[0063]

[0064] Example:

[0065] (1) Synthetic seismic data example

[0066] We use synthetic seismic data with noise-free, noise-containing and wedge-shaped models to verify the effectiveness of the present invention. Figure 1 are the velocity model and Q value model parameters, and the four-layer noise-free VSP data synthesized under this model, the source wavelet is a 40HZ Ricker wavelet, the depth is set to 0-1600m, and a detector is placed every 20m. The results of estimating the Q value of each layer using the method proposed by the present invention, the estimation method without phase constraint, and the spectral ratio method are as follows: Figure 2 As shown in (a)-(c), it can be seen that the Q values ​​estimated by the three methods are basically consistent with the theoretical Q values.

[0067] Because in actual processing, seismic data will inevitably be affected by noise and wavelet interference, we Figure 1 5%, 10% and 15% Gaussian random noise were added to the VSP data, respectively, to generate the following Figure 3 (a)-(c) are earthquake records shown. Under these three noise intensities, 200 independent tests were performed using the three estimation methods. The results show that when the random noise intensity is 5% ( Figure 3 (d)), the Q values ​​estimated by the three methods are close to the theoretical Q values, and the standard deviations are relatively small, except for the spectral ratio method which is slightly higher. When the random noise intensity increases to 10% ( Figure 3(e)), the deviation of the Q value estimation of the inversion method without considering the phase and the spectral ratio method from the theoretical value increases, and the standard deviation also increases significantly. In the case of 15% random noise ( Figure 3 (f)), the proposed method still shows good stability and high estimation accuracy, while the estimation performance of the other two methods is greatly affected by noise, not only the stability is reduced, but also the estimation accuracy is significantly reduced. Figure 4 ), due to the wavelet interference phenomenon, the inversion method that does not consider the phase ( Figure 5 (b)) and spectral ratio method ( Figure 5 (c)) has a large error in the estimation result, and the stability of the estimation result is also significantly affected. However, using the method in this paper, we can obtain a Q value inversion result with smaller error and higher stability ( Figure 5 (a)).

[0068] (2) Calculation example with actual seismic data

[0069] We apply the estimation method proposed in this invention to actual seismic data to perform attenuation estimation and attenuation compensation. The two-dimensional actual seismic profile used in the study is as follows: Figure 6 (a) shows that the time sampling interval is 2 milliseconds and contains 301 data. Each seismic data is intercepted by a window to extract the received seismic wave signals at different time positions, and the Q value is inverted using the method of the present invention. Subsequently, the estimated Q value is used to perform attenuation compensation on the original seismic profile, and the following is obtained: Figure 6 (b) shows the results. As can be seen from the figure, after Q value compensation, the overall resolution of the profile is significantly improved, especially the high-frequency details are enhanced, the reflection interface is clearer, the intensity of the deep reflection signal is restored, and the lateral continuity is improved. These results fully demonstrate the effectiveness of this method in seismic profile attenuation estimation and compensation.

[0070] The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, a person skilled in the art can still modify or make equivalent substitutions to the specific implementation schemes of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention are within the scope of protection of the claims of the present invention.

Claims

1. A spectrum consistency Q value inversion method under phase constraint, characterized by: The following steps are involved: Step 1: Perform Fourier transform on the seismic signals received at different times or locations, and compensate each data to the source location; Step 2: Extract the amplitude and phase information of the compensated data, and use the compensated mean to establish the spectrum consistency Q value inversion objective function I(Q) under phase constraint; Step 3: Use gradient descent method to solve the objective function I(Q).

2. The method according to claim 1, characterized in that The specific method for compensating the seismic signals picked up at different times or positions in step 1 is as follows: a. According to the wave theory of viscoelastic media, the received Nth seismic wave spectrum R N The frequency domain expression of (f) is: in, represents the source spectrum, x j represents the propagation distance from the j-1th receiving wave to the jth receiving wave, f represents the frequency, v j (f) represents the phase velocity, α j (f) is the attenuation coefficient; b. In the Kolsky-Futterman phase velocity dispersion model, the attenuation coefficient and phase velocity have the following corresponding relationship: Where f0 represents the reference frequency, v j (f0) is the phase velocity when the reference frequency is f0, Q j is the absorption attenuation factor at point j; c. Using the Taylor expansion of the exponential function, when all frequencies satisfy When , formula (1) can be expressed as: in, Indicates travel time; d. According to formula (4), the seismic signals received at different locations are sent to the source location. The compensation formula is:

3. The method according to claim 1, characterized in that The objective function I(Q) process established in step 2 is: a. Define the reference amplitude spectrum separately and the reference phase spectrum Among them, [f1,f2] and [f3,f4] represent the amplitude frequency band and phase frequency band related to the Q value respectively, s N and Respectively represent the amplitude and phase of the Nth received signal after compensation; b. Define the spectrum consistency Q value inversion objective function I(Q) under phase constraint as: Where Q={Q1,Q2,…Q N }, β represents the weighting coefficient.

4. The method according to claim 1, characterized in that: In step 3, the gradient calculation formula and update rule for solving the objective function using the gradient descent method are defined as follows: Where t represents the number of iterations and η represents the step size.

Images