Salt mine cavern connectivity fine discrimination method based on three-dimensional earthquake time-frequency analysis technology

Abstract

The invention provides a salt mine cavern connectivity fine discrimination method based on a three-dimensional earthquake time-frequency analysis technology. The method comprises the following steps: carrying out target processing on a three-dimensional earthquake data volume; converting the seismic data after target processing from a time domain to a frequency domain by using a generalized S transformation time-frequency analysis technology; carrying out parameter test on the regulatory factor lambda, and selecting an optimal lambda value according to a test result; carrying out parameter test on an attenuation factor P, and selecting an optimal P value according to a test result; a single-frequency data volume is obtained through conversion according to an optimized regulation factor lambda and an attenuation factor P, and the boundary range of the cavern is delineated by analyzing seismic response characteristics under different frequencies, so that fine discrimination of the connectivity of adjacent salt mine caverns is realized. According to the method, the time domain seismic data is converted into the frequency domain, and the single-frequency data body is used for explaining, so that the close cavern which cannot be distinguished by the time domain can be identified, and the connectivity of the cavern can be distinguished.

Description

Technical Field

[0001] The invention relates to the technical field of geophysical exploration, and in particular to a method for finely distinguishing connectivity of salt mine dissolution cavities based on three-dimensional seismic time-frequency analysis technology. Background Art

[0002] As the main mining method for well salt, water-soluble mining will form a large number of cavities in the mining area during the mining process. These cavities continue to expand as mining continues, and the distance between adjacent cavities gradually decreases. When the distance between cavities is too close, the cavities may be interconnected, resulting in excessive development of the cavities and a significant decrease in the supporting capacity of the cavities roof. This geological change can easily cause natural disasters such as ground brine, subsidence and collapse, posing a serious threat to the surrounding environment and the safety of human life and property.

[0003] In order to effectively guide the design of production wells, rationally utilize salt cavern space, and ensure safe production in salt mines, it is particularly important to find out the development of the dissolution cavity and the degree of damage to the salt layer roof. However, traditional exploration methods often find it difficult to accurately obtain relevant information when faced with complex and changing geological conditions.

[0004] In recent years, 3D seismic exploration technology has become an indispensable means in the field of salt mine exploration because it can determine the spatial morphology of salt mine cavities, the distance between the top of the cavities and the top of the salt mine layer, and the integrity of the overlying strata. However, when the distance between cavities is close, due to the mutual interference of the time domain seismic response, the imaging results will become messy, making it difficult to accurately distinguish the dissolution boundaries, which in turn increases the difficulty of determining the connectivity of cavities at relatively close distances. This problem is particularly prominent in current salt mine exploration and needs to be solved urgently.

[0005] In addition, affected by various factors such as sediments, cavities often show a low frequency characteristic in seismic response. This characteristic makes it more difficult to accurately identify cavities in the time domain. In order to overcome this limitation, the time-frequency analysis technology of three-dimensional seismic exploration came into being. This technology transforms three-dimensional seismic data from the time domain to the frequency domain through a certain mathematical transformation, and uses the response characteristics of different frequencies to identify geological structures and characterize the morphology of geological anomalies. This method can eliminate the mutual interference of different frequency components in the time domain and significantly improve the ability of seismic data to identify special geological bodies.

[0006] Among the many time-frequency analysis methods, the generalized S transform has attracted much attention due to its unique advantages. This method combines the advantages of short-time Fourier transform and wavelet transform, and can not only provide joint information of time and frequency, but also achieve frequency adaptability of resolution. This makes the generalized S transform particularly suitable for non-stationary signals with large instantaneous frequency changes in seismic response. In addition, the generalized S transform shows obvious advantages in time-frequency resolution, global optimality, phase information accuracy and anti-noise performance. Its parameters are easy to adjust and the calculation efficiency is high, which makes this method widely used in fault structure exploration and reservoir prediction in the oil, gas and coal fields, and shows great potential in three-dimensional seismic exploration of salt mine dissolution cavities.

[0007] In summary, in order to solve the problem of determining the connectivity of adjacent cavities in current salt mine exploration, the present invention proposes a three-dimensional seismic time-frequency analysis technology based on generalized S transform, aiming to improve the accuracy of determining the connectivity of cavities through this technology and provide strong guarantee for the safe production of salt mines. Summary of the invention

[0008] In view of the shortcomings of the prior art, the present invention provides a method for finely distinguishing the connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology. The method converts time-domain seismic data into frequency domain and interprets them using single-frequency data bodies, thereby being able to identify close cavities that cannot be distinguished in the time domain and distinguish their connectivity.

[0009] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0010] A method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology comprises the following steps:

[0011] Target processing of 3D seismic data to eliminate interference and improve the resolution of geological anomalies;

[0012] The generalized S-transform time-frequency analysis technique is used to convert the processed seismic data from the time domain to the frequency domain to eliminate the mutual interference of seismic responses of different frequencies and to highlight the seismic information that may be submerged in the time domain in the frequency domain.

[0013] Conduct parameter tests on the adjustment factor λ and select the optimal λ value based on the test results to ensure a balance of resolution in the time domain and frequency domain;

[0014] Conduct parameter tests on the attenuation factor P and select the optimal P value based on the test results;

[0015] The single-frequency data volume is converted according to the preferred adjustment factor λ and attenuation factor P. The boundary range of the cavity is delineated by analyzing the seismic response characteristics at different frequencies, thereby achieving a fine discrimination of the connectivity of adjacent salt mine cavities.

[0016] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, the seismic signal of actual seismic exploration is a finite and discrete signal, expressed as x(n), where n=0, 1, 2..., N-1, N is a positive integer, and the discrete Fourier transform DFT of the seismic signal is calculated using the following formula:

[0017]

[0018] Among them, 0≤k≤N-1,

[0019] The inverse transform of the discrete Fourier transform DFT is expressed by the following formula:

[0020]

[0021] Among them, 0≤n≤N-1.

[0022] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, a generalized S transform time-frequency analysis technology adopts a Gaussian window function, and the Gaussian window function is expressed as the following formula:

[0023]

[0024] Where: δ is the standard deviation of the Gaussian function, which is used to determine the width of the window function to obtain time-frequency analysis results with different resolutions.

[0025] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, a Gaussian window function is used to obtain local information of the original signal and reflect the time-varying characteristics of the original signal, wherein the shape and size of the time window are fixed, and only the time domain is shifted to obtain a short-time Fourier transform STFT, which is expressed as the following formula:

[0026]

[0027] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, a function δ related to the time-varying frequency is determined by the following formula:

[0028]

[0029] Where f is the frequency function of the original signal, and the expression of the standard S transform is:

[0030]

[0031] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, the standard S inverse transform expression is:

[0032]

[0033] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, parameters are introduced into the Gaussian window function, including:

[0034] The attenuation factor P is introduced to adjust the resolution of time-frequency analysis, and the adjustment factor λ is introduced to flexibly control the width of the time-frequency analysis window. The width σ of the Gaussian window function is parameterized to make it a function of the frequency f and these factors, which is expressed as the following formula:

[0035]

[0036] Among them, the size of the Gaussian time window in the generalized S transform is controlled by two parameters, namely the attenuation factor P and the adjustment factor λ. When P = 1, the generalized S transform can be approximated as the standard S transform; when P is not equal to 1, the size of the time window changes nonlinearly with the size of P, and λ is the linear coefficient that controls the nonlinear change.

[0037] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, the expression of generalized S transform time-frequency analysis technology is:

[0038]

[0039] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, the target processing of the three-dimensional seismic data volume includes:

[0040] Pre-stack multi-domain denoising: For surface wave interference in three-dimensional seismic data, the technology of suppressing surface waves in the time-frequency domain is used to suppress the interference of low-frequency surface waves; for oblique interference, it is suppressed in the τ-p domain to improve the signal-to-noise ratio of medium-deep data; for random energy interference and outliers, the regional abnormal energy suppression processing technology is applied, and the strategy of multi-channel identification and single-channel processing is adopted to analyze the seismic data in time windows and conduct comprehensive statistics on the recorded amplitude by frequency division, and the weighted median is used as the identification parameter to identify and process noise;

[0041] Energy compensation and adjustment: The energy of the mid-deep layer is compensated in the time domain through the spherical geometry compensation method; at the same time, the surface consistency amplitude compensation method is used to eliminate the difference in longitudinal and lateral energy of the reflected wave amplitude caused by different surface excitation and receiving conditions;

[0042] Multiple wave attenuation processing: For the multiple wave problems that may exist in the 3D seismic data volume, the early multiple wave attenuation processing is carried out; for the multiple waves remaining in the deep layer, the fk domain filtering method is used to further remove them;

[0043] Identification of special geological anomalies: Identify special geological anomalies in 3D seismic data.

[0044] According to a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology provided by the present invention, when performing parameter testing of the adjustment factor λ, a set of candidate value ranges of λ is set, and for each candidate value of λ, a machine learning algorithm or a deep learning framework is applied to perform data processing or model training respectively; the trained model or processed data under each candidate value of λ is evaluated using a validation data set, and based on the evaluation results, the performance under different candidate values ​​of λ is compared, and the λ value that makes the performance index reach the optimal value is selected as the optimal λ value;

[0045] When conducting parameter testing of the attenuation factor P, a set of candidate value ranges for P is set, and for each candidate value of P, a machine learning algorithm or a deep learning framework is applied to perform data processing or model training. The trained model or processed data under each candidate value of P is evaluated using a validation data set, and based on the evaluation results, the performance under different candidate values ​​of P is compared, and the P value that achieves the best performance indicator is selected as the optimal P value.

[0046] It can be seen that compared with the prior art, the present invention has the following beneficial effects:

[0047] The present invention introduces the generalized S transform (GST) time-frequency analysis technology into the three-dimensional seismic exploration of salt mine cavities, and successfully converts the time domain seismic data into the frequency domain. This conversion process effectively eliminates the mutual interference of seismic responses of different frequencies, so that the subtle seismic information that was originally difficult to distinguish in the time domain can be clearly displayed in the frequency domain. By comparing the actual application effects, the present invention significantly improves the recognition accuracy of adjacent and close cavities, especially those cavities that cannot be accurately distinguished in the time domain, thereby providing strong technical support for the detailed interpretation of salt mine cavities.

[0048] In order to solve the problem of connectivity detection of brine dissolution cavities in salt mines, the present invention proposes a time-frequency analysis method based on generalized S transform. By extracting multiple single-frequency data bodies and selectively selecting low-frequency bodies to identify the connectivity of dissolution cavities, the present invention can accurately reveal the connectivity status between adjacent dissolution cavities. The introduction of this method not only solves the limitations of traditional technology in the identification of dissolution cavity connectivity, but also provides an important basis for safety management and efficient resource utilization in the process of salt mining.

[0049] The successful application of this invention provides a new idea and method for the seismic interpretation of salt mine cavities. Compared with the traditional time domain seismic interpretation, the frequency domain interpretation has more advantages and can reveal more information hidden in the seismic data. This new interpretation idea is not only applicable to the exploration of salt mine cavities, but can also be extended to the identification of other geological bodies with similar geological characteristics and seismic responses, bringing innovative development to the field of geological exploration.

[0050] The present invention has achieved remarkable results in three-dimensional seismic exploration of solution cavities, accurately explained the distribution and connectivity of solution cavities, and provided a scientific basis for the mining plan of salt mines. This practical application result not only verifies the effectiveness and reliability of the present invention, but also further enhances its practical application value in the field of geological exploration. With the continuous promotion and application of technology, the present invention is expected to provide strong support for the exploration of more salt mines and other geological bodies.

[0051] In summary, the three-dimensional seismic exploration method for salt mine cavity based on the generalized S transform (GST) time-frequency analysis technology of the present invention has significant beneficial effects such as improving resolution and recognition accuracy, effectively judging the connectivity of cavity, providing new seismic interpretation ideas and enhancing practical application value.

[0052] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of an embodiment of a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology of the present invention.

[0054] Figure 2 It is a projection diagram on the cavity plane in an embodiment of a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology of the present invention.

[0055] Figure 3 It is a schematic diagram of the time section of the well A2 in an embodiment of a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology of the present invention.

[0056] Figure 4 It is a schematic diagram of time-frequency decomposition response for different adjustment factor λ values ​​in an embodiment of a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology of the present invention.

[0057] Figure 5 It is a schematic diagram of time-frequency decomposition responses for different attenuation factor values ​​in an embodiment of a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology of the present invention.

[0058] Figure 6This is an intentional diagram of single-frequency data volume sections of different frequencies in an embodiment of a method for finely distinguishing connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0061] See also Figure 1 This embodiment provides a method for finely distinguishing the connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology, and the method includes the following steps:

[0062] Step S1, target processing is performed on the three-dimensional seismic data volume to eliminate interference and improve the resolution of geological anomalies;

[0063] Step S2, using the generalized S-transform time-frequency analysis technique to convert the processed target seismic data from the time domain to the frequency domain, so as to eliminate the mutual interference of seismic responses of different frequencies and make the seismic information that may be submerged in the time domain stand out in the frequency domain;

[0064] Step S3, perform parameter test of adjustment factor λ, and select the optimal λ value according to the test result to ensure the balance of resolution in time domain and frequency domain; wherein, the larger the λ value and P value, the higher the time domain resolution, and the lower the frequency domain resolution, that is, the smaller the difference between different frequencies, the smoother the spectrum. The smaller the λ value and P value, the lower the time domain resolution, and the higher the frequency domain resolution.

[0065] Step S4, perform parameter test of attenuation factor P, and select the optimal P value according to the test result; wherein, the larger the λ value and P value, the higher the time domain resolution, and the lower the frequency domain resolution, that is, the smaller the difference between different frequencies, the smoother the spectrum. The smaller the λ value and P value, the lower the time domain resolution, and the higher the frequency domain resolution.

[0066] Step S5, converting the single-frequency data volume according to the preferred adjustment factor λ and the attenuation factor P, and defining the boundary range of the cavity by analyzing the seismic response characteristics at different frequencies, thereby achieving a fine discrimination of the connectivity of adjacent salt mine cavities.

[0067] In this embodiment, S transform is a linear time-frequency analysis method based on Fourier transform. The seismic signal of actual seismic exploration is a finite and discrete signal, represented by x(n), where n=0, 1, 2..., N-1, N is a positive integer, and the discrete Fourier transform DFT of the seismic signal is calculated using the following formula:

[0068]

[0069] Among them, 0≤k≤N-1,

[0070] The inverse transform of the discrete Fourier transform DFT is expressed by the following formula:

[0071]

[0072] Among them, 0≤n≤N-1.

[0073] In order to overcome the global average problem of Fourier transform spectrum analysis, the time window function is introduced, that is, the generalized S transform time-frequency analysis technology uses the Gaussian window function, which is expressed as the following formula:

[0074]

[0075] Among them, δ is the standard deviation value of the Gaussian function, which is used to determine the width of the window function, so as to obtain time-frequency analysis results of different resolutions. At this time, the local information of the original signal can be obtained, but the shape and size of the time window are fixed, and only the time domain is shifted, which does not reflect the time-varying characteristics of the frequency. The Fourier transform at this time is the short-time Fourier transform (STFT) using the Gaussian time window, which is expressed as the following formula:

[0076]

[0077] Substituting formula (3) into (4), the resolution of the time-frequency analysis result can be changed by adjusting the size of δ. In order to establish a relationship between the result and the frequency of the original signal, it is defined as a function related to the time-varying frequency, which is determined by the following formula:

[0078]

[0079] Where f is the frequency function of the original signal. Substituting formula (5) into (4), we get the standard S transform, which is expressed as follows:

[0080]

[0081] In this embodiment, the standard S inverse transform expression is:

[0082]

[0083] At this time, although the time-frequency analysis results have multiple resolutions related to frequency, the width of the time window is only related to the frequency and is reciprocal to each other, which makes it less flexible. In order to solve this problem, two parameters are introduced, including:

[0084] The attenuation factor P is introduced to adjust the resolution of time-frequency analysis, and the adjustment factor λ is introduced to flexibly control the width of the time-frequency analysis window. The width σ of the Gaussian window function is parameterized to make it a function of the frequency f and these factors, that is, the Gaussian time window expression is improved to a generalized S transform. At this time, the window function is expressed as the following formula:

[0085]

[0086] It can be seen from the above formula that the size of the Gaussian time window in the generalized S transform is controlled by two parameters, namely the attenuation factor P and the adjustment factor λ. When P = 1, the generalized S transform can be approximated as the standard S transform; when P is not equal to 1, the size of the time window changes nonlinearly with the size of P, and λ is the linear coefficient that controls the nonlinear change.

[0087] Substituting (8) into (7), we can get the expression of generalized S transform time-frequency analysis technology as:

[0088]

[0089] In the above step S1, the three-dimensional seismic data volume is subjected to target processing, including:

[0090] Pre-stack multi-domain denoising processing: For surface wave interference in three-dimensional seismic data, the technology of suppressing surface waves in the time-frequency domain is used to effectively suppress the interference of low-frequency surface waves; for oblique interference, it is suppressed in the τ-p domain to improve the signal-to-noise ratio of medium-deep data; for random energy interference and outliers, regional abnormal energy suppression processing technology is applied, and the strategy of multi-channel identification and single-channel processing is adopted to perform time-by-time window analysis on seismic data and conduct comprehensive statistics on the recorded amplitude by frequency division, and the weighted median is used as the identification parameter to accurately identify and process noise.

[0091] Energy compensation and adjustment: In order to eliminate the difference in wavefront energy attenuation as the propagation distance of seismic waves increases, the spherical geometry compensation method is used to compensate for the energy of mid-deep layers in the time domain. At the same time, the surface consistency amplitude compensation method is used to eliminate the differences in longitudinal and lateral energy of the reflected wave amplitude caused by different surface excitation and receiving conditions, and to achieve surface consistency amplitude balance, thereby improving the overall quality of seismic data.

[0092] Multiple wave attenuation processing: In order to solve the multiple wave problems that may exist in the three-dimensional seismic data body, especially the multiple waves generated during the data acquisition process in the frozen soil area, multiple methods such as superposition method, velocity differentiation method, and deconvolution method are adopted to carry out the preliminary multiple wave attenuation processing; for the multiple waves remaining in the deep layer, the fk domain filtering method is used to further remove them, so as to significantly improve the signal-to-noise ratio and resolution of the seismic data.

[0093] Identification of special geological anomalies: Combine signal processing, image processing, and deep learning technologies to identify special geological anomalies in 3D seismic data. For example, through coherence volume analysis, seismic attribute fusion, connectivity constraint segmentation and other methods, the ability to depict the edges of anomalies is enhanced, and the accuracy and resolution of anomaly identification are improved.

[0094] Through further processing of the above steps, the present invention can more effectively eliminate interference factors in the three-dimensional seismic data body, improve the resolution of geological anomalies, and provide more accurate and reliable geological information for subsequent oil and gas exploration and development.

[0095] In the above step S3, when performing parameter testing of the adjustment factor λ, a set of candidate value ranges of λ is set, and for each candidate value of λ, a machine learning algorithm or a deep learning framework is applied to perform data processing or model training; the trained model or processed data under each candidate value of λ is evaluated using a validation data set, and based on the evaluation results, the performance under different candidate values ​​of λ is compared, and the λ value that makes the performance index reach the best is selected as the optimal λ value;

[0096] In the above step S4, when performing parameter testing of the attenuation factor P, a set of candidate value ranges for P is set, and for each P candidate value, a machine learning algorithm or a deep learning framework is applied to perform data processing or model training; the trained model or processed data under each P candidate value is evaluated using a validation data set, and based on the evaluation results, the performance under different P candidate values ​​is compared, and the P value that makes the performance index reach the best is selected as the optimal P value.

[0097] In addition, this embodiment can also combine three-dimensional seismic data with multi-source geological information such as geology, drilling, and logging to conduct comprehensive interpretation and comparative analysis to improve the reliability of the determination of the connectivity of the dissolution cavity.

[0098] Specifically, the original 3D seismic data is preprocessed, including denoising, static correction, deconvolution, etc., to improve the signal-to-noise ratio and resolution of seismic data. Through 3D seismic interpretation technology, such as layer tracing and fault interpretation, seismic profiles and attribute bodies, such as amplitude, frequency, phase, etc., are obtained to provide a basis for subsequent comprehensive interpretation.

[0099] Then, collect regional geological data, including geological maps, geological profiles, and structural development history, to understand the regional geological background and structural characteristics. Based on the geological data, establish a preliminary geological model to provide geological constraints for subsequent seismic data interpretation. Collect drilling and logging data, including core descriptions, logging curves (such as sound waves, density, resistivity, etc.), etc., to understand the underground lithology and physical properties. Through drilling and logging data, establish wellside seismic synthetic records, calibrate seismic reflection layers, and provide a bridge for the fusion of seismic data and geological data.

[0100] Next, the seismic profile is compared with the geological profile to verify the accuracy of the seismic interpretation, and the seismic interpretation is revised based on the geological data. The geological model is used to constrain the seismic inversion, improve the inversion accuracy of the seismic data, and obtain more accurate information such as stratum thickness and lithology distribution.

[0101] Through the synthetic seismic recording near the well, the seismic reflection layer is calibrated with the drilling and logging data to achieve the deep integration of seismic data, drilling and logging data. The seismic attributes are calibrated and verified using the drilling and logging data, such as using the acoustic logging curve to calibrate the seismic amplitude attribute to obtain more accurate formation lithology and physical property information.

[0102] Finally, the multi-source geological information, such as seismic data, geological data, drilling and logging data, is comprehensively analyzed to construct an underground three-dimensional geological model. In the three-dimensional geological model, the boundaries and morphology of the solution cavity are identified, and the connectivity of the solution cavity is analyzed. The reliability of the determination of the connectivity of the solution cavity is verified by comparing the geological information from different sources, such as comparing the seismic reflection characteristics, drilling core descriptions and logging curve responses.

[0103] Furthermore, this embodiment can also utilize the geological model to constrain seismic inversion, thereby improving the inversion accuracy of seismic data and providing a more accurate geological basis for the identification of the connectivity of the solution cavity.

[0104] Furthermore, this embodiment may also use geostatistical methods, such as Kriging interpolation, sequential Gaussian simulation, etc., to perform spatial interpolation and uncertainty analysis on seismic attributes to evaluate the probability distribution and uncertainty of cavity connectivity.

[0105] In practical applications, the cavity formed by the brine well A2 is R1, the cavity formed by the brine wells A4 and A5 is R2, and the cavity formed by the brine well A3 is R3. Figure 2 As shown, Figure 2 This is a projection of the planar dissolution boundaries of the three cavities in 2022. It can be seen that the three cavities are not connected at this time, but the plane distance is very close, and there is a risk of dissolution. In order to explore the development of the cavities in the mining area, a new round of 3D seismic exploration will be carried out in 2024. The mining company has confirmed through other technical means that R1 and R2 are not dissolved, but the distance between the cavities cannot be determined. Figure 3is the time section through well A2, and the section position is shown in Figure 2 As can be seen from the figure, on the time section, the solution cavity is affected by the sediments, showing a strong low-frequency anomaly. The reflection wave anomaly areas of the two solution cavities R1 and R2 are mixed together and cannot be distinguished. Therefore, the solubility of the solution cavity that is close to the time domain seismic data cannot be determined.

[0106] Since the dissolution cavity appears as a low-frequency anomaly in the seismic response, the present invention uses the generalized S transform with high longitudinal (time) and lateral (frequency) resolution to perform time-frequency decomposition, and decomposes the three-dimensional seismic data volume in the time domain into a four-dimensional frequency volume. The four dimensions are line, trace, time and frequency value, and the solubility of the dissolution cavity is interpreted using the frequency domain analysis method.

[0107] Parameter optimization 1: Adjustment factor λ. Before time-frequency decomposition, the parameters must be tested first. There are two main parameters that need to be tested for the generalized S transform: λ and P. The empirical range of the two parameters is 0.1-3. The larger the λ value and the P value, the higher the time domain resolution, and the lower the frequency domain resolution, that is, the smaller the difference between different frequencies, and the smoother the spectrum. The smaller the λ value and the P value, the lower the time domain resolution and the higher the frequency domain resolution. The selection of the two values ​​must be tested based on actual data to determine the final value. Select the target channel to test the parameters. Figure 4 The time-frequency decomposition responses when the P value is fixed at 1.1 and λ is 0.5, 1.0, 1.5, and 2.0 respectively. It can be seen from the figure that when λ is 0.5, the lateral frequency change is the largest, and when λ is 2.0, the lateral frequency change is the smallest. Considering the time domain resolution comprehensively, λ is finally determined to be 1.0.

[0108] Parameter optimization 2: attenuation factor P. Figure 5 The time-frequency decomposition results for a fixed λ of 1.0 and P of 0.5, 0.8, 1.2, and 1.6 respectively. As can be seen from the figure, when P is 0.5, the lateral frequency change is the largest, and there is almost no lateral frequency change. Considering the time domain resolution comprehensively, P is finally determined to be 0.8.

[0109] After determining λ and P, the generalized S transform is used to perform time-frequency calculation on the data volume to obtain a four-dimensional frequency volume, and the frequency volume is subjected to frequency division processing to obtain a single-frequency data volume. Figure 6 for Figure 3 The time section corresponds to the single-frequency data volume sections of 5Hz, 10Hz, 15Hz and 20Hz. It can be seen from the figure that the vertical resolution of several data volumes varies greatly, and the 5Hz single-frequency data volume has the highest vertical resolution. The horizontal changes of different single-frequency data volumes are not obvious. The seismic frequency responses of the two cavities R1 and R2 do not interfere with each other. There is an obvious distance between the cavities, and they are not connected, which is consistent with the actual exposure. Through calculation, it can be known that the distance is about 35m.

[0110] From the known to the unknown, the single-frequency data volume can be used to interpret the dissolution boundary of the cavity in the entire mining area. For example, the seismic anomalies of the two cavities R2 and R3 cannot be judged to be connected in the time section, while the seismic anomalies between the two cavities in the frequency volume section have a clear distance and are not connected.

[0111] In order to solve the problem of detecting the connectivity of adjacent cavities, this paper proposes a set of time-frequency analysis methods based on the low-frequency seismic response characteristics of brine cavities in salt mines to identify the connectivity of adjacent cavities. The generalized S transform is used to convert the time domain seismic data into the frequency domain. In order to eliminate the mutual interference of multi-frequency seismic wave responses, multiple single-frequency data bodies are extracted, and low-frequency bodies are selected in a targeted manner to identify the connectivity of the cavities.

[0112] Compared with the time-domain seismic data, the frequency body obtained based on the generalized S transform has higher resolution. By comparing with the actual exposure situation, the interpretation results obtained by the time-frequency analysis technology are more accurate. The successful application of this technology in the three-dimensional seismic exploration of the dissolution cavity in a salt mine mining area in the Dingyuan Basin provides a new idea for the fine seismic interpretation of the salt mine dissolution cavity, and can be extended to the identification of other geological bodies with similar geological characteristics and seismic responses.

[0113] In summary, the present invention introduces the generalized S transform (GST) time-frequency analysis technology into the three-dimensional seismic exploration of salt mine cavities, and successfully converts the time domain seismic data into the frequency domain. This conversion process effectively eliminates the mutual interference of seismic responses of different frequencies, so that the subtle seismic information that was originally difficult to distinguish in the time domain can be clearly displayed in the frequency domain. By comparing the actual application effects, the present invention significantly improves the recognition accuracy of adjacent and close cavities, especially those cavities that cannot be accurately distinguished in the time domain, thereby providing strong technical support for the detailed interpretation of salt mine cavities.

[0114] Furthermore, by extracting multiple single-frequency data bodies and selectively selecting low-frequency bodies to determine the connectivity of the solution cavities, the present invention can accurately reveal the connectivity status between adjacent solution cavities, which not only solves the limitations of traditional technologies in determining the connectivity of solution cavities, but also provides an important basis for safety management and efficient utilization of resources in the salt mining process.

[0115] Furthermore, compared with traditional time-domain seismic interpretation, the frequency-domain interpretation of the present invention has more advantages and can reveal more information hidden in the seismic data. It is not only suitable for the exploration of salt mine cavities, but can also be extended to the identification of other geological bodies with similar geological characteristics and seismic responses, bringing innovative development to the field of geological exploration.

[0116] Furthermore, the present invention has achieved remarkable results in three-dimensional seismic exploration of solution cavities, accurately explained the distribution and connectivity of solution cavities, and provided a scientific basis for the mining plan of salt mines. This practical application result not only verifies the effectiveness and reliability of the present invention, but also further enhances its practical application value in the field of geological exploration. With the continuous promotion and application of technology, the present invention is expected to provide strong support for the exploration of more salt mines and other geological bodies.

[0117] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0118] The above-mentioned embodiments are only preferred embodiments of the present invention and cannot be used to limit the scope of protection of the present invention. Any non-substantial changes and substitutions made by technicians in this field on the basis of the present invention shall fall within the scope of protection required by the present invention.

Claims

1. A method for finely distinguishing the connectivity of salt mine cavities based on three-dimensional seismic time-frequency analysis technology, characterized in that: The following steps are involved: Target processing of 3D seismic data to eliminate interference and improve the resolution of geological anomalies; The generalized S-transform time-frequency analysis technique is used to convert the processed seismic data from the time domain to the frequency domain to eliminate the mutual interference of seismic responses of different frequencies and to highlight the seismic information that may be submerged in the time domain in the frequency domain. Conduct parameter tests on the adjustment factor λ and select the optimal λ value based on the test results to ensure a balance of resolution in the time domain and frequency domain; Conduct parameter tests on the attenuation factor P and select the optimal P value based on the test results; The single-frequency data volume is converted according to the preferred adjustment factor λ and attenuation factor P. The boundary range of the cavity is delineated by analyzing the seismic response characteristics at different frequencies, thereby achieving a fine discrimination of the connectivity of adjacent salt mine cavities.

2. The method according to claim 1, characterized in that: The seismic signal of actual seismic exploration is a finite and discrete signal, expressed as x(n), where n = 0, 1, 2..., N-1, and N is a positive integer. The discrete Fourier transform DFT of the seismic signal is calculated using the following formula: Among them, 0≤k≤N-1, The inverse transform of the discrete Fourier transform DFT is expressed by the following formula: Among them, 0≤n≤N-1.

3. The method according to claim 2, characterized in that: The generalized S transform time-frequency analysis technique uses a Gaussian window function, which is expressed as the following formula: Where: δ is the standard deviation of the Gaussian function, which is used to determine the width of the window function to obtain time-frequency analysis results with different resolutions.

4. The method according to claim 3, characterized in that: The Gaussian window function is used to obtain the local information of the original signal and reflect the time-varying characteristics of the original signal. The shape and size of the time window are fixed, and only the time domain is shifted to obtain the short-time Fourier transform STFT, which is expressed as the following formula:

5. The method according to claim 4, characterized in that: The function δ, which is related to the time-varying frequency, is given by the following formula: Where f is the frequency function of the original signal, and the expression of the standard S transform is:

6. The method according to claim 5, characterized in that: The standard inverse S transform expression is:

7. The method according to claim 6, characterized in that: Introduce parameters into the Gaussian window function, including: The attenuation factor P is introduced to adjust the resolution of time-frequency analysis, and the adjustment factor λ is introduced to flexibly control the width of the time-frequency analysis window. The width σ of the Gaussian window function is parameterized to make it a function of the frequency f and these factors, which is expressed as the following formula: Among them, the size of the Gaussian time window in the generalized S transform is controlled by two parameters, namely the attenuation factor P and the adjustment factor λ. When P = 1, the generalized S transform can be approximated as the standard S transform; when P is not equal to 1, the size of the time window changes nonlinearly with the size of P, and λ is the linear coefficient that controls the nonlinear change.

8. The method according to claim 7, characterized in that: The expression of generalized S transform time-frequency analysis technology is:

9. The method according to any one of claims 1 to 8, characterized in that: The target processing of the three-dimensional seismic data volume includes: Pre-stack multi-domain denoising: For surface wave interference in three-dimensional seismic data, the technology of suppressing surface waves in the time-frequency domain is used to suppress the interference of low-frequency surface waves; for oblique interference, it is suppressed in the τ-p domain to improve the signal-to-noise ratio of medium-deep data; for random energy interference and outliers, the regional abnormal energy suppression processing technology is applied, and the strategy of multi-channel identification and single-channel processing is adopted to analyze the seismic data in time windows and conduct comprehensive statistics on the recorded amplitude by frequency division, and the weighted median is used as the identification parameter to identify and process noise; Energy compensation and adjustment: The energy of the mid-deep layer is compensated in the time domain through the spherical geometry compensation method; at the same time, the surface consistency amplitude compensation method is used to eliminate the difference in longitudinal and lateral energy of the reflected wave amplitude caused by different surface excitation and receiving conditions; Multiple wave attenuation processing: For the multiple wave problems that may exist in the 3D seismic data volume, the early multiple wave attenuation processing is carried out; for the multiple waves remaining in the deep layer, the fk domain filtering method is used to further remove them; Identification of special geological anomalies: Identify special geological anomalies in 3D seismic data.

10. The method according to any one of claims 1 to 8, characterized in that: When testing the parameters of the adjustment factor λ, a set of candidate value ranges for λ is set, and for each candidate value of λ, a machine learning algorithm or a deep learning framework is applied to perform data processing or model training; the trained model or processed data under each candidate value of λ is evaluated using a validation data set, and the performance under different candidate values ​​of λ is compared based on the evaluation results, and the λ value that makes the performance index reach the best is selected as the optimal λ value; When conducting parameter testing of the attenuation factor P, a set of candidate value ranges for P is set, and for each candidate value of P, a machine learning algorithm or a deep learning framework is applied to perform data processing or model training. The trained model or processed data under each candidate value of P is evaluated using a validation data set, and based on the evaluation results, the performance under different candidate values ​​of P is compared, and the P value that achieves the best performance indicator is selected as the optimal P value.

Images