Seismic prediction method for thin reservoirs

Abstract

The present application provides a thin reservoir seismic prediction method, comprising the following steps: determining a target well; performing time-frequency spectrum decomposition on a well-side seismic trace of the target well, calculating frequency division data of multiple frequency bands according to a result of the time-frequency spectrum decomposition, and determining an optimal frequency band based on the frequency division data; selecting an optimal logging curve according to a corresponding relationship between the logging curve and the thin reservoir; calculating a seismic attribute and a correlation coefficient between the seismic attribute and the optimal logging curve based on the optimal frequency band; calculating a thin reservoir attribute indicating factor according to the correlation coefficient; and performing thin reservoir prediction according to the thin reservoir attribute indicating factor. The method can realize high-precision seismic prediction of the thin reservoir.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of oil and gas exploration and development, and particularly relates to a thin reservoir seismic prediction method. BACKGROUND

[0002] The reservoir thickness that can be identified by seismic data amplitude is generally 1 / 4 wavelength, and the reservoir with thickness lower than 1 / 4 wavelength is defined as a thin reservoir. The thin reservoir has the characteristics of thin thickness, strong heterogeneity, significant lateral variation, small single-layer thickness, and small wave impedance property difference between sandstone and mudstone, which brings great challenges to the exploration and development of oil and gas fields.

[0003] At present, the main methods for predicting thin reservoirs include seismic attribute analysis method and seismic inversion method. The seismic attribute analysis method extracts useful seismic attributes such as amplitude, frequency, phase and other information from preprocessed seismic data, and uses dimension reduction mapping, attribute selection and clustering methods to optimize seismic attributes for better prediction of thin reservoirs. Seismic attributes are a comprehensive reflection of structure, stratum, lithology and oil and gas, and conventional seismic attribute analysis methods usually extract seismic attributes in the time domain without considering frequency bands, which has limited prediction effect on thin reservoirs and needs to be further improved in prediction accuracy.

[0004] On the other hand, the seismic inversion method infers the information of the internal structure and properties of the earth by inverting the seismic observation data. The widely used deterministic inversion technology such as sparse pulse does not consider the dominant frequency band, but performs inversion in the time domain, which is equivalent to considering the entire seismic frequency band. Therefore, due to the limitation of seismic resolution and band-limited wavelet, the resolution is low and the multi-solution is strong, and the prediction effect is usually good for thick reservoirs, but not good for thin layer prediction applications. SUMMARY

[0005] In view of the deficiencies of the conventional thin reservoir seismic prediction method, the present application proposes a thin reservoir prediction method considering the dominant seismic frequency band, which can realize high-precision seismic prediction of thin reservoirs.

[0006] According to an aspect of the present application, a thin reservoir seismic prediction method is provided, which comprises:

[0007] Step 1: determining a target well;

[0008] Step 2: performing time-frequency spectrum decomposition on the well-side seismic trace of the target well, calculating frequency division data of a plurality of frequency bands according to the result of the time-frequency spectrum decomposition, and determining the best frequency band based on the frequency division data of the plurality of frequency bands;

[0009] Step 3: selecting the best well logging curve according to the corresponding relationship between the well logging curve and the thin reservoir;

[0010] Step 4: calculating seismic attributes based on the optimal frequency band and the correlation coefficient of the seismic attributes and the optimal well log curve;

[0011] Step 5: calculating a thin reservoir attribute indicating factor according to the correlation coefficient;

[0012] Step 6: predicting a thin reservoir according to the thin reservoir attribute indicating factor.

[0013] Preferably, the step 1 comprises:

[0014] Step 11: calculating a main frequency fp of a seismic data of a target layer of a work area and extracting a seismic wavelet;

[0015] Step 12: for each well, synthesizing a seismic record of the well according to the seismic wavelet and the well log data, and calculating a correlation coefficient of a well-side seismic trace of the well and the synthesized seismic record;

[0016] Step 13: selecting a well with the largest correlation coefficient as the target well.

[0017] Preferably, in the step 2, time-frequency spectrum decomposition is performed according to the following formula (1):

[0018]

[0019] wherein, represents a window function considering frequency, t represents time, τ represents a center time of a time window, f represents frequency, x(t) represents a well-side seismic trace signal, f0(τ) represents a main frequency at time τ, k represents a factor balancing time resolution and frequency resolution, Δf(τ)=f0(τ)-f, and X represents a time-frequency spectrum.

[0020] Preferably, in the step 2, the multiple frequency bands are divided with the main frequency fp as the center, and the frequency ranges of the multiple frequency bands are [f low , fp-M*a],…,[fp-M,fp+M],[fp+M~fp+3*M],...,[fp+M*b,f high ], wherein 2*M is a frequency band width, a and b are both odd numbers, fp-M*a>f low , fp+M*b<f high , f low , f high are the lowest frequency and the highest frequency for analysis.

[0021] Preferably, in the step 2, the determining of the optimal frequency band based on the frequency-divided data of the multiple frequency bands comprises:

[0022] The frequency division data of the plurality of frequency bands are compared with the well logging interpretation result, and a frequency band with the highest correlation between the frequency division data and the well logging interpretation result is selected as the optimal frequency band.

[0023] Preferably, the well logging curves include multiple of velocity logging curves, porosity logging curves, gamma logging curves, and spontaneous potential logging curves.

[0024] Preferably, in the step 4, the seismic attributes include amplitude, frequency, phase, and coherence attributes.

[0025] Preferably, the step 4 further includes:

[0026] The correlation coefficients of the seismic attributes and the optimal well logging curve are sorted in descending order, and the three largest correlation coefficients are denoted as c1, c2, and c3, respectively.

[0027] The seismic attributes corresponding to the three largest correlation coefficients are denoted as A1, A2, and A3, and the seismic attributes A1, A2, and A3 are normalized, i.e. i =A i / max(|A i |), i=1, 2, 3, max(|A i |) represents the maximum value of the absolute values of the elements in the vector A i .

[0028] Preferably, in the step 5, the thin reservoir attribute indicating factor is calculated according to the following formula (2):

[0029] A=k1*A1+k2*A2+k3*A3 (2)

[0030] wherein A represents the thin reservoir attribute indicating factor,

[0031] Another aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the thin reservoir seismic prediction method.

[0032] Still another aspect of the present application provides an electronic device, which includes:

[0033] a memory storing executable instructions;

[0034] a processor running the executable instructions in the memory to implement the thin reservoir seismic prediction method.

[0035] The thin reservoir seismic prediction method has the beneficial effects that: the time-frequency analysis method is used for high-precision time-frequency spectrum decomposition of seismic data, the relationship between different frequency band data and the thin reservoir is analyzed through well-seismic combination, the most sensitive advantage frequency band to the thin reservoir is found, then the thin reservoir sensitive attribute is optimized on the advantage frequency band data and the thin reservoir indicator is constructed, and finally the high-precision seismic prediction of the thin reservoir is realized. The method can be used for exploration and development of different types of thin reservoir oil and gas reservoirs such as clastic rock and shale, and guides the optimization of favorable areas, well deployment and optimization of exploration and development schemes.

[0036] The method of the present application has other characteristics and advantages that will be apparent from or set forth in the accompanying drawings and the detailed description that follows, which together serve to explain certain principles of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0037] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout the figures, and wherein:

[0038] Figure 1 A flow chart of a thin reservoir seismic prediction method according to one embodiment of the present application is shown.

[0039] Figure 2 A post-stack seismic data crosswell profile according to one exemplary embodiment of the present application is shown.

[0040] Figure 3 A thin reservoir seismic prediction result according to one exemplary embodiment of the present application is shown. DETAILED DESCRIPTION

[0041] Preferred embodiments of the present application will be described herein below with reference to the accompanying drawings. While the preferred embodiments of the present application are shown in the drawings, it is understood that the present application can be embodied in various forms without being limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0042] The present application proposes a thin reservoir seismic prediction method, comprising the following steps:

[0043] Step 1: determining a target well;

[0044] Step 2: performing time-frequency spectrum decomposition on the seismic trace beside the target well, calculating frequency division data of a plurality of frequency bands according to the time-frequency spectrum decomposition result, and determining an optimal frequency band based on the frequency division data;

[0045] Step 3: Select the best logging curve according to the correspondence between the logging curve and the thin reservoir;

[0046] Step 4: Calculate the seismic attribute and the correlation coefficient between the seismic attribute and the best logging curve based on the best frequency band;

[0047] Step 5: Calculate the thin reservoir attribute indicator according to the correlation coefficient;

[0048] Step 6: Predict the thin reservoir according to the thin reservoir attribute indicator.

[0049] The thin reservoir seismic prediction method of the present application adopts time-frequency analysis method to perform high-precision time-frequency spectrum decomposition on seismic data, analyzes the relationship between different frequency band data and thin reservoirs through well-seismic combination, finds the most sensitive advantage frequency band to thin reservoirs, then optimizes thin reservoir sensitive attributes on the advantage frequency band data and constructs thin reservoir indicator, and finally realizes high-precision seismic prediction of thin reservoirs. The method can be used for exploration and development of different types of thin reservoir oil and gas reservoirs such as clastic rock and shale, and can guide favorable area optimization, well site deployment and exploration and development scheme optimization.

[0050] Example 1

[0051] Figure 1 A flowchart of a thin reservoir seismic prediction method according to one embodiment of the present application is shown. As shown in the figure, the method includes steps 1-6.

[0052] Step 1: Determine the target well.

[0053] Step 1 specifically includes steps 11-13 as follows:

[0054] Step 11: Calculate the main frequency fp of the seismic data of the target layer section in the work area, and extract the seismic wavelet.

[0055] The main frequency of seismic data refers to the main frequency component of seismic signals. In seismic exploration, the frequency range of seismic signals is very wide, usually in the range of 2-90Hz. Different regions and different purposes of seismic exploration will have different seismic signal frequencies.

[0056] For the processing and analysis of seismic data, it is very important to understand the frequency components of seismic signals. Through the analysis of the frequency components of seismic signals, the physical properties, structure and other information of the underground rock layers can be inferred. Understanding the main frequency of seismic data is of great significance to the accuracy and reliability of seismic inversion results.

[0057] Seismic wavelets are secondary waves generated when seismic waves excite the internal structures of the Earth's crust during an earthquake. They are a crucial component of seismic waves, possessing high energy and destructive power. Seismic wavelets can significantly impact buildings and the natural environment during an earthquake. There are two main types of seismic wavelets: P-waves and S-waves. P-waves are seismic waves whose vibration direction is the same as the wave's propagation direction. P-waves propagate quickly and have relatively low destructive power, and are usually the main component of seismic wavelets. S-waves are seismic waves whose vibration direction is perpendicular to the wave's propagation direction. S-waves propagate slowly and have greater destructive power, and are usually a minor component of seismic wavelets.

[0058] Methods for extracting seismic wavelets from the dominant frequency of seismic data typically include the following steps:

[0059] Seismic data preprocessing: Seismic data is preprocessed, including data cleaning, format conversion, and other operations, to ensure data quality and consistency.

[0060] Calculate the dominant frequency of the seismic signal: Perform spectral analysis on the preprocessed seismic signal to calculate the dominant frequency of the seismic signal.

[0061] Determining the length of the seismic wavelet: The length of the seismic wavelet is determined based on the dominant frequency of the seismic signal and the length of the seismic record. Generally, the length of the seismic wavelet should be greater than the period of the seismic signal.

[0062] Seismic wavelet extraction: Seismic wavelets are extracted from seismic records using the dominant frequency of the seismic signal and a determined wavelet length. This can be achieved through filters or other signal processing methods.

[0063] Verification and optimization: The extracted seismic wavelets are verified and optimized to ensure their quality and reliability. This includes denoising and filtering operations on the wavelets to further improve their quality.

[0064] Extracting seismic wavelets from the dominant frequency of seismic data is a well-known technique in this field. It's important to note that this requires the comprehensive application of multiple techniques and methods, and also demands high quality from the seismic data and accuracy in the dominant frequency calculation. Therefore, adjustments and optimizations are necessary in practical applications based on specific circumstances.

[0065] Step 12: For each well, synthesize the well's seismic record based on the seismic wavelet and logging data, and calculate the correlation coefficient between the well's perimeter seismic trace and the synthesized seismic record.

[0066] In this embodiment, seismic records can be synthesized according to the following steps:

[0067] Acquiring seismic wavelets and logging data: Acquire seismic wavelets and logging data, which typically include waveform characteristics of seismic signals, time information of seismic events, and physical properties and structural information of underground rock formations, etc.

[0068] Data preprocessing: Preprocess seismic wavelets and logging data, including data cleaning, format conversion, etc. to ensure data quality and consistency.

[0069] Synthetic seismic record: Synthesize seismic wavelets and logging data into a seismic record. This usually requires matching seismic wavelets with waveform characteristics in logging data and adding seismic wavelets to the seismic record according to the matching results.

[0070] Data verification: Verify the synthesized seismic record to ensure its quality and reliability. This includes denoising, filtering, etc. to further improve the quality of the seismic record.

[0071] After synthesizing the seismic record, calculate the correlation coefficient between the well-side seismic trace and the synthesized seismic record, which can reflect the similarity between the well-side seismic trace and the synthesized seismic record. The larger the correlation coefficient, the higher the similarity between the two. The following are the general steps for calculating the correlation coefficient:

[0072] Obtain well-side seismic trace data and seismic record data;

[0073] Align the well-side seismic trace data and seismic record data to ensure consistency in time scale;

[0074] Calculate the covariance matrix of the well-side seismic trace data and the seismic record data;

[0075] Calculate the eigenvalues and eigenvectors of the covariance matrix;

[0076] Select the eigenvector corresponding to the largest eigenvalue as the principal component vector;

[0077] Calculate the modulus of the principal component vector, which is the correlation coefficient.

[0078] Step 13: Select the well with the largest correlation coefficient as the target well.

[0079] Step 2: Perform time-frequency spectrum decomposition on the well-side seismic trace of the target well, calculate the frequency division data of multiple frequency bands according to the results of time-frequency spectrum decomposition, and determine the best frequency band based on the frequency division data.

[0080] In step 2, perform time-frequency spectrum decomposition according to the following formula (1):

[0081]

[0082] where, a window function considering frequency, t represents time, τ represents the center time of the time window, f represents frequency, x(t) represents a well seismic trace signal, f0(τ) represents the dominant frequency at time τ, k represents a factor balancing time resolution and frequency resolution, Δf(τ) = f0(τ)-f, and X represents a time-frequency spectrum.

[0083] In the embodiment, the length of the window function can be adaptively determined according to the frequency band of the seismic data, and the accuracy is higher than that of a conventional time-frequency analysis method.

[0084] In the embodiment, a plurality of frequency bands are divided with the dominant frequency as the center, and the frequency ranges of the plurality of frequency bands are [f low , fp-M*a],…,[fp-M,fp+M],[fp+M~fp+3*M],...,[fp+M*b,f high ], where 2*M is the frequency band width, a and b are both odd numbers, and fp-M*a>f low , fp+M*b<f high , f low , f high are the lowest frequency and the highest frequency used for analysis. For example, the frequency ranges of the plurality of frequency bands can be [5, fp-5*a],…,[fp-5,fp+5],[fp+5~fp+15],...,[fp+5*b,2*fp], where a and b are both odd numbers, and fp-5a>5, fp+5b<2*fp.

[0085] Then, the frequency division data of the frequency band is calculated according to the result of the time-frequency spectrum decomposition in the plurality of frequency bands.

[0086] The time-frequency spectrum decomposition can represent the signal as a function of time and frequency. Through the time-frequency spectrum decomposition, we can obtain the energy distribution of the signal at different times and frequencies. Through the time-frequency spectrum decomposition, a spectrum diagram can be obtained, in which each pixel point represents the energy of the signal at a specific time and a specific frequency.

[0087] In order to calculate the frequency division data, the time-frequency spectrum diagram can be further processed. Specifically, the processing can be performed according to the following steps:

[0088] Determine the frequency division band: first, the range of the frequency division band to be used needs to be determined.

[0089] Statistical energy: for each frequency division band, we need to count the energy of the signal in the band. This can be achieved by accumulating the pixel points in the time-frequency spectrum diagram.

[0090] Calculate the frequency division data: according to the statistical energy value, the center frequency, bandwidth and energy of each frequency division band can be calculated. These data can be used as the frequency division data.

[0091] Finally, determine the optimal frequency band based on the frequency division data, including: comparing the frequency division data of multiple frequency bands with the logging interpretation results, and selecting the frequency band with the highest correlation between the frequency division data and the logging interpretation results as the optimal frequency band.

[0092] The logging interpretation results can be compared with the frequency division data to analyze the correlation between the signals at different frequency bands and the thin reservoir. The signal characteristics at different frequency bands can be compared with the key geological parameters such as the thickness, resistivity, and porosity of the thin reservoir to determine which frequency band is most sensitive to the characterization of the thin reservoir, and that frequency band can be determined as the optimal frequency band.

[0093] Step 3: Select the optimal logging curve according to the correspondence between the logging curve and the thin reservoir.

[0094] The logging curve can include velocity logging curve, porosity logging curve, gamma logging curve, natural potential logging curve, etc. By comparing the correspondence between these logging curves and the thin reservoir, the logging curve with the highest correspondence can be selected as the optimal logging curve.

[0095] The logging curves can be compared with the key geological parameters such as the thickness, resistivity, and porosity of the thin reservoir to determine which logging curves are more sensitive to the characterization of the thin reservoir.

[0096] By comparing and analyzing the parameters of the thin reservoir with the logging curves, the response characteristics of different logging curves to the thin reservoir can be observed. For example, for the characteristics of the thin reservoir such as the increase in resistivity or the decrease in acoustic velocity, the change rule of the corresponding logging curve can be analyzed.

[0097] Based on the comparison and analysis, the logging curves that are sensitive to the characterization of the thin reservoir are selected. These logging curves usually have high signal feature retention, low noise level, and good signal-to-noise separation. According to the application requirements, one optimal logging curve is selected in this embodiment.

[0098] Step 4: Calculate the seismic attributes and the correlation coefficient between the seismic attributes and the optimal logging curve based on the optimal frequency band.

[0099] Seismic attributes are physical characteristics and parameters used to describe seismic signals in seismic exploration, including amplitude, frequency, phase, and coherence attributes, etc. These attributes can provide information about the properties of underground rock layers, structural features, and reservoir parameters, which are helpful for geological exploration and identification and evaluation of oil and gas reservoirs.

[0100] The calculation of seismic attributes usually includes the following steps:

[0101] 1. Data preprocessing: Preprocess the seismic data, including noise removal, filtering, interpolation, etc., to improve the quality and consistency of the data.

[0102] 2. Determine seismic attributes: Determine the seismic attributes that need to be calculated, such as amplitude, frequency, phase, waveform, etc.

[0103] 3. Calculate seismic attributes: Calculate the attribute values in the seismic data according to the required seismic attributes using corresponding algorithms or techniques. For example, for amplitude attributes, the maximum amplitude value of the seismic signal can be calculated; for frequency attributes, the frequency components of the seismic signal can be calculated by Fourier transform, etc.

[0104] 4. Data post-processing: Post-process the calculated seismic attributes, such as normalizing or standardizing the attribute values, to facilitate analysis and application.

[0105] 5. Visualization or application: Visualize the calculated seismic attributes, such as making seismic attribute maps or seismic attribute profile maps, or apply the attributes to geological interpretation and resource assessment, etc.

[0106] Calculate the seismic attributes of the optimal frequency band and the correlation coefficients between the seismic attributes and the optimal well curve. Preferably, the correlation coefficients between each seismic attribute and the optimal well curve are sorted in descending order, and the three largest correlation coefficients are denoted as c1, c2, and c3, respectively. The seismic attributes corresponding to the correlation coefficients c1, c2, and c3 are denoted as A1, A2, and A3, respectively. Normalize the seismic attributes A1, A2, and A3, i.e. i A i i=1, 2, 3, where max(|A i |) represents the maximum value of the absolute values of the elements in the vector A i . i

[0107] Step 5: Calculate the thin reservoir attribute indicator according to the correlation coefficients.

[0108] Calculate the thin reservoir attribute indicator according to the following formula (2):

[0109] A=k1*A1+k2*A2+k3*A3 (2)

[0110] where A represents the thin reservoir attribute indicator,

[0111] Step 6: Perform thin reservoir prediction based on the thin reservoir attribute indicator.

[0112] After calculating the thin reservoir attribute indicator, thin reservoir prediction can be performed based on the thin reservoir attribute indicator. The thin reservoir attribute indicator is calculated by optimizing the thin reservoir sensitive attributes on the optimal frequency band data, thereby achieving high-precision seismic prediction of thin reservoirs.​

[0113] Preferably, the full-area seismic data is subjected to time-frequency spectrum decomposition, the frequency-decomposed data of the best frequency band is calculated for each trace, and the thin reservoir attribute indicator is calculated according to the foregoing steps, so as to realize full-area thin reservoir seismic prediction.

[0114] Example 2

[0115] Embodiment 2 provides a thin reservoir seismic prediction method, comprising the following steps:

[0116] Step 1: determining a target well;

[0117] Step 2: subjecting a seismic trace beside the target well to time-frequency spectrum decomposition, calculating frequency-decomposed data of a plurality of frequency bands according to a result of the time-frequency spectrum decomposition, and determining a best frequency band based on the frequency-decomposed data of the plurality of frequency bands;

[0118] Step 3: selecting a best well logging curve according to a corresponding relationship between the well logging curve and the thin reservoir;

[0119] Step 4: calculating a seismic attribute and a correlation coefficient between the seismic attribute and the best well logging curve based on the best frequency band;

[0120] Step 5: calculating a thin reservoir attribute indicator according to the correlation coefficient;

[0121] Step 6: performing thin reservoir prediction according to the thin reservoir attribute indicator.

[0122] In the embodiment, in Step 2, time-frequency spectrum decomposition is performed according to the following formula (1):

[0123]

[0124] wherein, represents a window function considering frequency, t represents time, τ represents a center time of a time window, f represents frequency, x(t) represents a trace beside well signal, f0(τ) represents a main frequency at time τ, k represents a factor balancing time resolution and frequency resolution, Δf(τ) = f0(τ) - f, and X represents time-frequency spectrum.

[0125] In the embodiment, in Step 2, a plurality of frequency bands are divided with the main frequency fp as the center, and the frequency ranges of the plurality of frequency bands are [5, fp-5*a], …, [fp-5, fp+5], [fp+5~fp+15], …, [fp+5*b, 2*fp], wherein a and b are both odd numbers, and fp-5a>5, fp+5b<2*fp.

[0126] In the embodiment, in Step 2, determining the best frequency band based on the frequency-decomposed data of the plurality of frequency bands comprises:

[0127] The frequency division data of multiple frequency bands are compared with the well logging interpretation results, and a frequency band with the highest correlation between the frequency division data and the well logging interpretation results is selected as the best frequency band.

[0128] In this embodiment, the well logging curves include multiple well logging curves such as a velocity logging curve, a porosity logging curve, a gamma logging curve, and a spontaneous potential logging curve.

[0129] In this embodiment, in step 4, the seismic attributes include amplitude, frequency, phase, and coherence attributes.

[0130] In this embodiment, step 4 further includes:

[0131] The correlation coefficients of the seismic attributes and the best well logging curve are sorted in descending order, and the three largest correlation coefficients are denoted as c1, c2, and c3, respectively.

[0132] The seismic attributes corresponding to the three largest correlation coefficients are denoted as A1, A2, and A3, and the seismic attributes A1, A2, and A3 are normalized, that is, i A i i=1, 2, 3, and max(|A i |) represents the maximum value of the absolute values of the elements in the vector A i . i

[0133] In this embodiment, in step 5, the thin reservoir attribute indicating factor is calculated according to the following formula (2):

[0134] A=k1*A1+k2*A2+k3*A3 (2)

[0135] wherein A represents the thin reservoir attribute indicating factor,

[0136] In this embodiment, the actual data of a certain work area is used for testing. The work area is a clastic rock reservoir, Figure 2 is a post-stack seismic data cross well profile of the work area, Figure 3 is a result of thin reservoir seismic prediction by using the thin reservoir seismic prediction method proposed in this embodiment. Figure 3 In the figure, blue and white represent smaller reservoir probabilities, and red, yellow, and green represent larger reservoir probabilities. The well curve is a sand body interpretation result. This embodiment shows that the thin sandstone predicted by the thin reservoir seismic prediction method is consistent with the well logging sand body interpretation result, and the method is effective.

[0137] Other detailed descriptions about this exemplary embodiment can be referred to the corresponding descriptions in the foregoing embodiments, which will not be repeated here. ​

[0138] Example 3

[0139] The embodiment provides a thin reservoir seismic prediction device, which comprises:

[0140] a target well determination unit configured to determine a target well;

[0141] a time-frequency processing unit configured to perform time-frequency spectrum decomposition on a well-side seismic trace of the target well, calculate frequency division data of a plurality of frequency bands according to a result of the time-frequency spectrum decomposition, and determine an optimal frequency band based on the frequency division data of the plurality of frequency bands;

[0142] a well logging curve processing unit configured to select an optimal well logging curve according to a corresponding relationship between a well logging curve and a thin reservoir;

[0143] a seismic attribute processing unit configured to calculate a seismic attribute and a correlation coefficient between the seismic attribute and the optimal well logging curve based on the optimal frequency band;

[0144] a thin reservoir attribute indication factor calculation unit configured to calculate a thin reservoir attribute indication factor according to the correlation coefficient;

[0145] a thin reservoir prediction unit configured to perform thin reservoir prediction according to the thin reservoir attribute indication factor.

[0146] Preferably, the determination of the target well comprises:

[0147] Step 11: calculating a main frequency fp of seismic data of a target layer section of a work area, and extracting a seismic wavelet;

[0148] Step 12: for each well, synthesizing a seismic record of the well according to the seismic wavelet and well logging data, and calculating a correlation coefficient between a well-side seismic trace of the well and the synthesized seismic record;

[0149] Step 13: selecting a well with the largest correlation coefficient as the target well.

[0150] Preferably, the time-frequency spectrum decomposition is performed according to the following formula (1):

[0151]

[0152] wherein, represents a window function considering frequency, t represents time, τ represents a center time of a time window, f represents frequency, x(t) represents a well-side seismic trace signal, f0(τ) represents a main frequency at time τ, k represents a factor balancing time resolution and frequency resolution, Δf(τ) = f0(τ)-f, and X represents a time-frequency spectrum.

[0153] Preferably, the plurality of frequency bands are divided with the main frequency fp as the center, and the frequency ranges of the plurality of frequency bands are [5, fp-5*a], …, [fp-5, fp+5], [fp+5~fp+15], …, [fp+5*b, 2*fp], where a and b are both odd numbers, and fp-5a>5, fp+5b<2*fp.

[0154] Preferably, the determining the optimal frequency band based on the frequency division data of the plurality of frequency bands comprises:

[0155] Comparing the frequency division data of the plurality of frequency bands with the well logging interpretation result, and selecting a frequency band with the highest correlation between the frequency division data and the well logging interpretation result as the optimal frequency band.

[0156] Preferably, the well logging curves comprise multiple well logging curves such as velocity logging curves, porosity logging curves, gamma logging curves, and spontaneous potential logging curves.

[0157] Preferably, the seismic attributes comprise amplitude, frequency, phase, and coherence attributes.

[0158] Preferably, the calculating the seismic attributes and the correlation coefficients between the seismic attributes and the optimal well logging curves based on the optimal frequency band further comprises:

[0159] Sorting the correlation coefficients between the seismic attributes and the optimal well logging curves in descending order, where the three largest correlation coefficients are respectively denoted as c1, c2, and c3.

[0160] Denoting the seismic attributes corresponding to the three largest correlation coefficients as A1, A2, and A3, and normalizing the seismic attributes A1, A2, and A3, i.e. i A i i i =1, 2, 3, and max(|A i |) represents the maximum value of the absolute values of the elements in the vector A i .

[0161] Preferably, the thin reservoir attribute indication factor is calculated according to the following formula (2):

[0162] A=k1*A1+k2*A2+k3*A3 (2)

[0163] where A represents the thin reservoir attribute indication factor,

[0164] Other detailed descriptions related to the present exemplary embodiment can refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0165] Example 4

[0166] The embodiment provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the thin reservoir seismic prediction method.

[0167] The computer readable storage medium includes, but is not limited to, an optical storage medium (for example, CD-ROM and DVD), a magneto-optical storage medium (for example, MO), a magnetic storage medium (for example, a magnetic tape or a mobile hard disk), a medium with a built-in rewritable nonvolatile memory (for example, a memory card), and a medium with a built-in ROM (for example, a ROM cartridge).

[0168] Other detailed descriptions about the present exemplary embodiment can refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0169] Example 5

[0170] The embodiment provides an electronic device, which comprises:

[0171] a memory, which stores executable instructions;

[0172] a processor, which runs the executable instructions in the memory to realize the thin reservoir seismic prediction method.

[0173] Other detailed descriptions about the present exemplary embodiment can refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0174] The above has described the embodiments of the present application, and the above descriptions are exemplary, not exhaustive, and are not limited to the disclosed embodiments. Many modifications and changes are obvious to those skilled in the art without departing from the scope and spirit of the described embodiments. The selection of the terms used herein is intended to best explain the principles of the embodiments, practical application or technical improvement to the technology in the market, or to enable other ordinary skilled persons in the art to understand the embodiments disclosed herein.

Claims

1. A thin reservoir seismic prediction method, characterized by, The method comprises: Step 1: determining a target well; Step 2: performing time-frequency spectrum decomposition on a well-side seismic trace of the target well, calculating frequency division data of a plurality of frequency bands according to a result of the time-frequency spectrum decomposition, and determining an optimal frequency band based on the frequency division data of the plurality of frequency bands; Step 3: selecting an optimal logging curve according to a corresponding relationship between the logging curve and a thin reservoir; Step 4: calculating a seismic attribute and a correlation coefficient between the seismic attribute and the optimal logging curve based on the optimal frequency band; Step 5: calculating a thin reservoir attribute indicating factor according to the correlation coefficient; Step 6: predicting a thin reservoir according to the thin reservoir attribute indicating factor.

2. The method of claim 1, wherein, The step 1 comprises: Step 11: calculating a main frequency fp of seismic data of a target layer of a work area, and extracting a seismic wavelet; Step 12: for each well, synthesizing a seismic record of the well according to the seismic wavelet and logging data, and calculating a correlation coefficient between a well-side seismic trace of the well and the synthesized seismic record; Step 13: selecting a well with the largest correlation coefficient as the target well.

3. The method of claim 1, wherein, In the step 2, time-frequency spectrum decomposition is performed according to the following formula (1): wherein, represents a window function that accounts for frequency, t represents time, τ represents a center time of the time window, f represents frequency, x(t) represents a near-borehole seismic trace signal, f0(τ) represents a dominant frequency at time τ, k represents a factor that balances time resolution and frequency resolution, Δf(τ) = f0(τ) - f, and X represents a time-frequency spectrum.

4. The method of claim 2, wherein, In the step 2, the multiple frequency bands are divided with the center of the main frequency fp, and the frequency range of the multiple frequency bands is [f low , fp-M*a], [fp-M*a+1, fp-M*a+2], …, [fp-M, fp+M], [fp+M+1, fp+M+2], …, [fp+M*b, f high ], where 2*M is the frequency band width, a and b are both odd numbers, and fp-M*a low < fp+M*b high , f low < fp-M high < fp+M are the lowest and highest frequencies for analysis.

5. The method of claim 4, wherein, In the step 2, the determination of the optimal frequency band based on the frequency division data of the plurality of frequency bands comprises: Comparing the frequency division data of the plurality of frequency bands with a logging interpretation result, and selecting a frequency band with the highest correlation between the frequency division data and the logging interpretation result as the optimal frequency band.

6. The method of claim 1, wherein, The logging curve comprises a plurality of velocity logging curves, porosity logging curves, gamma logging curves, and natural potential logging curves.

7. The method of claim 1, wherein, In the step 4, the seismic attribute comprises amplitude, frequency, phase, and coherence attributes.

8. The method of claim 7, wherein, The step 4 further comprises: Sorting the correlation coefficients between the seismic attribute and the optimal logging curve in descending order, wherein the three largest correlation coefficients are denoted as c1, c2, and c3, respectively; The seismic attributes corresponding to the three largest correlation coefficients are denoted as A1, A2, and A3. Seismic attributes A1, A2, and A3 are then normalized, i.e., A... i =A i / max(|A i |), i = 1, 2, 3, max(|A i |) represents vector A i The maximum absolute value of the elements in the set.

9. The method of claim 8, wherein, In the step 5, the thin reservoir attribute indicating factor is calculated according to the following formula (2): A=k1*A1+k2*A2+k3*A3 (2) wherein A represents a thin reservoir property indicator factor, 10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the thin reservoir seismic prediction method according to any one of claims 1-9.

11. An electronic device, comprising: The electronic device comprises: A memory storing executable instructions; A processor running the executable instructions in the memory to implement the thin reservoir seismic prediction method according to any one of claims 1-9.

Images