A method for identifying effective reservoirs of thin interbedded combinations based on resonance imaging principle

Abstract

The application discloses a kind of based on resonance imaging principle thin interbed combination effective reservoir identification method, steps are as follows: first, select typical well in study target area, divide lithology based on well logging data in target layer, determine the main frequency of target layer based on seismic data;Then calculate lithology combination analysis time window and forward simulation formation thickness, combine resonance imaging principle and divide target layer time window into four parts and determine the thickness of each part;Subsequently, identify typical combination category of target layer, extract each formation physical parameter and carry out forward simulation to generate synthetic seismic data set;After standardization processing to synthetic and actual seismic data, the frequency spectrum characteristics of synthetic data are analyzed by Mexican wavelet transform, and the discrimination standard is summarized;Finally, calculate the size of sliding time window, carry out time-frequency analysis and sliding comparison to actual seismic data, and determine effective reservoir.The application combines Mexican wavelet time-frequency analysis and resonance principle, improves the accuracy of effective reservoir identification in thin interbed area, and has wide adaptability and strong practical applicability.

Description

Technical Field

[0001] This invention belongs to the field of geophysical exploration technology, specifically, it relates to a method for effective reservoir identification based on the principle of resonance imaging for thin interbedded layers. Background Technology

[0002] As oil and gas exploration in my country continues to deepen, the focus of oilfield exploration and development is gradually shifting towards mid-to-deep formations. However, mid-to-deep exploration faces significantly higher development costs (compared to shallow formations). Existing mid-to-deep exploration indicates that the target drilling interval may be a dry layer, an effective reservoir, or a combination of both. Identifying the effective reservoir in the target interval is a key task in exploration and development, and is of great significance for reducing costs and increasing efficiency.

[0003] For the identification of effective reservoirs and dry layers, the identification based on seismic exploration is mainly based on time-frequency analysis. Liu Chuanhu, Liu Fugui and others used fast Fourier transform to convert seismic records into frequency domain energy spectra to predict reservoirs. The analysis model found that the time-frequency characteristic values ​​are affected by the formation thickness, formation combination and other factors. Based on this, they qualitatively predicted the oil-bearing properties of the reservoir and the inverted thickness, thus improving the accuracy and efficiency of exploration.

[0004] However, existing methods are mainly effective for thick reservoirs, but they are powerless for thin formation combinations. This is because in the case of thin formation combinations, some dry layers may exhibit high-frequency characteristics, while others may exhibit low-frequency characteristics. The same applies to effective reservoirs. In some combinations, oil layers may not exhibit low-frequency characteristics, but rather high-frequency characteristics. This makes it difficult to identify effective reservoirs and dry layers in areas with thin interbedded formations based on existing methods.

[0005] To address the shortcomings of existing methods, this paper proposes a method for identifying effective reservoirs and dry layers in thin interbedded formations based on the principle of resonance imaging. This method incorporates the principle of resonance imaging as a guide. Within a certain range, strata of a certain thickness will affect the imaging of a certain point. This effect may be resonance enhancement or resonance weakening, and different combinations have different effects. However, the spectral characteristics of a certain combination have certain patterns. Based on this, the study conducts typical stratigraphic combination analysis for the target layer and performs forward modeling. The forward modeling simulation is analyzed to obtain the discrimination criteria for effective reservoirs and dry layers under different combination categories. The actual seismic data is standardized, and the discrimination criteria of the forward modeling simulation are used to identify the effective reservoirs and dry layers in the target layer.

[0006] In view of this, the present invention is proposed. Summary of the Invention

[0007] To solve the above-mentioned technical problems, the basic concept of the technical solution adopted by the present invention is as follows:

[0008] A method for effective reservoir identification based on the principle of resonance imaging for thin interlayered formations includes the following steps:

[0009] (1) Select typical wells from the research target area and divide the lithology corresponding to different depths of the target layer based on relevant logging data.

[0010] (2) For the target layer where the seismic data is located, determine the dominant frequency f of the target layer based on the spectral analysis method. m .

[0011] (3) Use the statistical formula (Equation 1) to determine the analysis window W for lithological assemblage categories T (W T (The magnitude is comparable to the apparent principal period of the seismic wavelet), in milliseconds (ms).

[0012] Formula 1

[0013] Where a and b are the fitting coefficients;

[0014] (4) Determine the average velocity Va based on the sonic logging data of typical well sections, and based on Va and W T Determine the forward modeling stratigraphic thickness that can be used for correlation of the target section, W D , unit m.

[0015] (5) Determining W based on the principle of seismic imaging resonance D Within the thickness range, the thicknesses of the upper two parts, upper one part, lower one part, and lower two parts are respectively D2. T D1 T D1 B D2 B Its D2 T =D2 B D1 T =D1 B .

[0016] (6) For a typical well, start using W along the top of the target layer of the typical well. D The analysis was conducted, specifically focusing on the top layer P of the target layer. T Centered on [P] T -W D / 2,P T +W D The lithological assemblage category within [ / 2] is determined, and then D1 is slid in sequence. B Thickness, until the target segment does not meet W D Similar analyses were performed on all typical wells up to the thickness limit to ultimately determine the typical thin interbedded layer combination category {M} of the target interval. i |i is an integer, 0≤i≤N}.

[0017] (7) Combine the logging data such as sonic waves, density, and oil testing near the target well section to interpret the average velocity of the effective reservoir (oil layer, water layer), dry layer, overlying non-reservoir, intermediate non-reservoir, and underlying non-reservoir. , , , , ) and average density ( , , , , ).

[0018] (8) Thin interbedded layer combination categories {M} based on typical well interpretations i |i is an integer, 0≤i≤N} and the average velocity and average density of the corresponding effective reservoirs (oil layer, water layer), dry layer, overlying non-reservoir, intermediate non-reservoir, and underlying non-reservoir are used for forward modeling to obtain the corresponding synthetic seismic data set {S}. i |i is an integer, 0≤i≤L}.

[0019] (9) Find the corresponding maximum values ​​from the synthetic seismic data set {Si|i is an integer, 0≤i≤L} and the target layer of the actual seismic data. and Subsequently, a unified maximum amplitude benchmark S was established (this value is a pre-specified fixed value, and also serves as the reference maximum value for the standardization process of the synthetic seismic dataset and the actual seismic data), and standardization processing was performed on the two types of data respectively. Synthetic seismic dataset standardization: This process was completed using Equation 2, where... The amplitude represents the amplitude before standardization of the synthetic seismic dataset. The amplitude represents the standardized amplitude of the synthetic seismic dataset (0 ≤ i < K, where K is the total number of samples in the synthetic seismic dataset).

[0020] Standardization of actual seismic data: Processing is performed using Equation 3, where... This represents the amplitude of the actual earthquake data before standardization. This represents the standardized amplitude (0≤i<U, where U is the number of valid sample points in the actual seismic data).

[0021] Formula 2

[0022] Formula 3

[0023] (10) Using Mexican wavelet transform on the synthetic seismic dataset, time-frequency analysis data were obtained, and the spectral data at the center point of different combined models were extracted, denoted as {P}. i|i is an integer, 0≤i<T}, where T is the number of spectral data points extracted from the center point of different combination models. The spectral data is analyzed to summarize the spectral characteristics of different combination models.

[0024] (11) Determine the size of the typical wellbore sliding window, and combine it with the average velocity Va and thickness D1 of the typical wellbore. B Calculate the sliding window size W s (Equation 4), unit: ms.

[0025] Formula 4

[0026] (12) Extract the normalized seismic data near the typical well, and then obtain the time-frequency analysis data near the typical well based on the Mexican wavelet transform. Analyze the time-frequency analysis data of the target layer based on the sliding time window Ws. The analysis starts from the top of the target layer, extracts the spectral data at the top, and plots the amplitude spectrum. Compare the actual amplitude spectrum characteristics with the amplitude spectrum characteristics of the forward modeling to determine whether there is an effective reservoir within the thickness range of Ws below the target layer. After the analysis is completed, move the sliding time window down by Ws and determine whether there is an effective reservoir in the next thickness range. Repeat this process to complete the identification of the effective reservoir of the entire typical well.

[0027] Compared with the prior art, the present invention has the following advantages:

[0028] This invention combines Mexican wavelet time-frequency analysis with the resonance principle, solving the problem of difficult identification of effective reservoirs and dry layers in thin interbedded layers using traditional methods. High-precision identification: By proposing a method for identifying effective reservoirs in thin interbedded layers guided by resonance principle imaging, it significantly improves the accuracy of distinguishing effective reservoirs from dry layers and reduces ambiguity. Wide adaptability: Applicable to complex sand-mud interbedded geological conditions, it has a unique advantage in identifying thin reservoirs. Strong practical applicability: Verified by actual data, it highly matches the results of forward modeling, providing reliable technical support for oilfield exploration and development.

[0029] The specific embodiments of the present invention will now be described in further detail with reference to the accompanying drawings. Attached Figure Description

[0030] In the attached diagram:

[0031] Figure 1 The SP-GR-AC-R25-Den logging curves and lithology-reservoir (oil layer / dry layer) comprehensive interpretation diagrams for well A in the actual work area;

[0032] Figure 2 Spectral diagram of the target layer in the reservoir;

[0033] Figure 3 Geological model dimensions and target layer Segmentation and corresponding thickness diagram;

[0034] Figure 4 Geological model diagram of mud-encased sand category;

[0035] Figure 5 Geological model diagram of sand-bag mud category;

[0036] Figure 6 Geological model map of the entire range;

[0037] Figure 7 Seismic response diagrams for three models in the mud-encased sand category;

[0038] Figure 8 The energy analysis diagram in the frequency domain is studied using the first sliding time window;

[0039] Figure 9 The energy analysis diagram in the frequency domain is studied for the second sliding time window;

[0040] Figure 10 The energy analysis diagram in the frequency domain is studied for the third sliding time window;

[0041] Figure 11 This is a diagram showing the identification results of effective reservoirs and dry layers in the target layer. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments will be clearly and completely described below with reference to the accompanying drawings. The following embodiments are used to illustrate the present invention.

[0045] Practical example description:

[0046] 1. Select typical wells from the target area and, based on relevant logging data, delineate the lithology corresponding to different depths of the target interval, such as... Figure 1 show.

[0047] 2. For the target layer containing the seismic data, draw a spectrum diagram based on the spectral analysis method to determine the dominant frequency f of the target layer. m ,like Figure 2 The dominant frequency f of the target layer segment m It is 32Hz.

[0048] 3. Determine the analytical time window W for lithological assemblage categories using the statistical formula (Equation 1). T W was calculated T It takes 56ms.

[0049] 4. Based on the sonic logging data of typical wells, the average velocity V is calculated. a At approximately 4000 m / s, based on Va and W T Further determine the forward modeling stratigraphic thickness that can be used for correlation of the target section, and calculate W. D It is 112m.

[0050] 5. Determining W based on the principle of seismic imaging resonance D Within the thickness range, the thicknesses of the upper two parts, upper one part, lower one part, and lower two parts are as follows: Figure 3 The structure consists of four parts: the upper two sections (from the upper green line to the upper blue line), the upper one section (from the upper blue line to the red line), the lower one section (from the red line to the lower blue line), and the lower two sections (from the lower blue line to the lower green line). The thickness of each of these four sections is D2. T D1 T D1 B D2 B Its D2 T =D2 B =36m, D1 T =D1 B =20m.

[0051] 6. For a typical well, start utilizing W along the top of the target formation of the typical well. D The analysis was conducted, specifically focusing on the top layer P of the target layer. T Centered on [P] T -W D / 2,P T +W D The lithological assemblage categories within the range of / 2] are determined, and then D1 is slid in sequence. B Thickness, until the target segment does not meet W D Similar analyses were performed on all typical wells up to the thickness limit to ultimately determine the typical combination category of the target layer {M}. i |i is an integer, 0≤i≤6}. There are six typical combinations of thin interbedded formations in this well's target section: mud-covered sand, sand-covered mud, full interval, upper sand and lower mud, upper mud and lower sand, and intermittent type. Only the first three types are shown here, such as... Figure 4 , 5 6.

[0052] 7. Based on the logging data such as sonic waves, density, and oil testing near the target formation in typical wells, interpret the average velocity of the effective reservoir (oil layer, water layer), dry layer, overlying non-reservoir, intermediate non-reservoir, and underlying non-reservoir. , , , , ) and average density ( , , , , (as shown in Table 1).

[0053] 8. Combining typical well combination categories {M i |i is an integer, 0≤i≤6} and the average velocity and average density of the corresponding effective reservoir (oil layer, water layer), dry layer, overlying non-reservoir, intermediate non-reservoir, and underlying non-reservoir are used for forward modeling to obtain the corresponding synthetic seismic data set {S}. i |i is an integer, 0≤i≤L}. Here, L is 16 because each typical combination category is further subdivided into different models. Only the three models for the mud-covered sand category are shown here, such as... Figure 7 .

[0054] 9. From the synthesized record data {S i |i is an integer, 0≤i≤16} and the corresponding maximum values ​​of the target layer in the actual seismic data are respectively found. and A reference amplitude of 350 was set, and the synthetic records and actual seismic data were standardized.

[0055] 10. Using Mexican wavelet transform, time-frequency analysis data is obtained from the standardized synthetic seismic dataset. Spectral data at the center points of different combined models are extracted and denoted as {P}. i |i is an integer, 0≤i<T}. The spectral data is analyzed to summarize the spectral characteristics of each model in different combination categories, and the effective reservoir and dry layer distinction boundary values ​​corresponding to each model in different combination categories are clarified, as shown in Table 2.

[0056] 11. Determine the typical wellbore sliding time window size, combining it with the typical wellbore average velocity Va and thickness D1. B The sliding window size W is calculated using Equation 3. s It takes 10ms.

[0057] 12. Extract typical well-side standardized seismic data, then obtain typical well-side time-frequency analysis data based on Mexican wavelet transform, and analyze the target layer time-frequency analysis data based on a sliding time window (Ws). From the top to the bottom of the target layer in this well, a total of 13 sliding windows were used, resulting in four full-range categories, four upper sand / lower mud categories, one sand-mud category, and four upper mud / lower sand categories, sequentially from top to bottom. Specific models were then further subdivided.

[0058] By comparing the time-frequency analysis data of the center point with the data in Table 2, the D1 below the center point of each time window is determined. B Sand body types within the range, such as Figure 8 , 9 As shown in Figure 10, only the spectrum at the center point obtained from the first three time windows is displayed here. Figure 8For the first time window corresponding to the full range I model, it can be seen that the energy is greater than 50 near the frequency of 30Hz, so the lower part of this time window is determined to be a dry layer. Figure 9 , 10 These are the second and third time windows, respectively. Figure 8 The first time window is the same, so the lower part of both time windows is determined to be a dry layer.

[0059] By applying this method to actual data from the target layer, it was predicted that an effective reservoir exists in the lower part of the fifth slip time. The sandstone layers in the lower parts of the 1st, 2nd, 3rd, 4th, 6th, and 12th slip times are dry layers, while the lower parts of the 7th, 8th, 9th, 10th, 11th, and 13th slip times are classified as mudstone layers. Specific results are as follows: Figure 11 The predictions are largely consistent with the actual well logging interpretations.

[0060] Table 1 shows the velocity and density information for mudstone layers and different reservoirs.

[0061]

[0062] Table 2 shows the effective reservoir / dry layer discrimination boundaries for various model studies on the vicinity of frequency domain characteristics.

[0063]

Claims

1. A method for effective reservoir identification based on the principle of resonance imaging, characterized in that, Includes the following steps: Step S1: Select typical wells from the research target area and delineate the lithology corresponding to different depths of the target interval based on relevant logging data; Step S2: For the target layer containing the seismic data, determine the dominant frequency f of the target layer based on spectral analysis. m ; Step S3: Use the statistical formula in Equation 1 to determine the analysis window W for lithological combination categories. T， W T Its magnitude is comparable to the apparent principal period of the seismic wavelet, measured in milliseconds (ms). Formula 1 Step S4: Determine the average velocity Va based on the sonic logging data of the typical well target interval, and then use Va and W... T Determine the forward modeling stratigraphic thickness that can be used for correlation of the target section, W D The unit is meters (m). Step S5, determine W based on the principle of seismic imaging resonance. D Within the thickness range, the thicknesses of the upper two parts, upper one part, lower one part, and lower two parts are respectively D2. T D1 T D1 B D2 B Its D2 T =D2 B D1 T =D1 B ; Step S6: For a typical well, start using W along the top of the target formation of the typical well. D The analysis was conducted, specifically focusing on the top P layer of the target layer. T Centered on [P] T -W D / 2, P T +W D The lithological assemblage category within [ / 2] is determined, and then D1 is slid in sequence. B Thickness, until the target segment does not meet W D Similar analyses were performed on all typical wells up to the thickness limit to ultimately determine the typical thin interbedded layer combination category {M} of the target interval. i |i is an integer, 0≤i≤N}, where N is the number of combination categories; Step S7: Based on logging data such as sonic waves, density, and oil testing near the target well section, interpret the average velocity of the effective reservoir (oil layer, water layer), dry layer, overlying non-reservoir, intermediate non-reservoir, and underlying non-reservoir. , , , , ) and average density ( , , , , ); Step S8, combining the thin interbedded layer combination category {M} from typical well interpretations. i |i is an integer, 0≤i≤N} and the average velocity and average density of the corresponding effective reservoir, dry layer, overlying non-reservoir, intermediate non-reservoir, and underlying non-reservoir are used for forward modeling to obtain the corresponding synthetic seismic data set {Si|i is an integer, 0≤i≤L}, where L is the number of models included in all combination categories; Step S9, from the synthetic seismic data set {S i |i is an integer, 0≤i≤L} and the target segment of the actual seismic data are respectively found to have the corresponding maximum values. and Subsequently, a unified maximum amplitude benchmark S was set (this value is a pre-specified fixed value, and also serves as the reference maximum value in the standardization process of synthetic seismic data and actual seismic data), and the two types of data were standardized separately; Synthetic seismic data standardization: the process was completed using Equation 2, where This represents the amplitude of the synthetic seismic data before standardization. The amplitude represents the standardized amplitude of the synthetic seismic data (0 ≤ i < K, where K is the total number of samples in the synthetic seismic data); the actual seismic data is standardized using Equation 3, where... This represents the amplitude of the actual earthquake data before standardization. Represents the standardized amplitude (0 ≤ i < U, where U is the number of valid sample points in the actual seismic data); let the synthetic record under standardization be {S}. i ’ |i is an integer, 0≤i≤L}; Formula 2 Formula 3 Step S10: Using Mexican wavelet transform on the standardized synthetic seismic dataset, time-frequency analysis data is obtained, and the spectral data at the center point of different combined models is extracted, denoted as {P}. i |i is an integer, 0≤i<T}, where T is the number of spectral data points extracted from the center point of different combination models. The spectral data is analyzed to summarize the spectral characteristics of different combination models. Step S11: Determine the typical wellbore sliding time window size, combining the typical wellbore average velocity Va and thickness D1. B Calculate the sliding window size W s The unit is milliseconds (ms). Formula 4 Step S12: Extract the standardized seismic data near the typical well, and then obtain the time-frequency analysis data near the typical well based on the Mexican wavelet transform. Analyze the time-frequency analysis data of the target layer based on the sliding time window Ws. The analysis starts from the top of the target layer, extracts the spectral data at the top, and plots the amplitude spectrum. Compare the actual amplitude spectrum characteristics with the forward modeling to determine whether there is an effective reservoir within the thickness range of Ws below the target layer. After the analysis is completed, move the sliding time window down by Ws and determine whether there is an effective reservoir in the next thickness range. Repeat this process to complete the identification of the effective reservoir of the entire typical well.

2. The method for effective reservoir identification based on the principle of resonance imaging for thin interlayered formations according to claim 1, characterized in that, The logging data mentioned in step S1 includes sonic logging, density logging, GR logging, SP logging resistivity logging data, and lithology including effective reservoirs, dry layers, and non-reservoirs.

3. The method for effective reservoir identification based on the principle of resonance imaging for thin interlayered formations according to claim 1, characterized in that, The spectrum analysis method described in step S2 is power spectrum analysis, which determines the dominant frequency by identifying spectral peaks. .

4. The method for effective reservoir identification based on the principle of resonance imaging for thin interlayered formations according to claim 1, characterized in that, Step S11 analyzes the time-frequency analysis data of the target layer based on the sliding time window Ws. During the analysis, starting from the top of the target layer, the spectral data at the top is extracted and the amplitude spectrum is plotted. The actual amplitude spectrum characteristics are compared with the amplitude spectrum characteristics of the forward modeling to determine whether there is an effective reservoir within the thickness range of Ws below the target layer. After the analysis is completed, the sliding time window is moved down by Ws to determine whether there is an effective reservoir in the next thickness range. This process is repeated to complete the identification of the effective reservoir of the entire typical well.

5. The method for effective reservoir identification based on the principle of resonance imaging for thin interlayered formations according to claim 1, characterized in that, The forward modeling described in step S8 employs either the wave equation forward modeling method or the ray tracing method, with the source wavelet being the Ricker wavelet and the dominant frequency... Target segment main frequency.

6. The method for effective reservoir identification based on the principle of resonance imaging for thin interlayered formations according to claim 1, characterized in that, The reference amplitude value for the normalization process described in step S9 The range is 100-400, which allows the dynamic range of the amplitude of the standardized data to better distinguish the energy difference boundary between the effective reservoir and the dry layer.

7. The method for effective reservoir identification based on the principle of resonance imaging for thin interlayered formations according to claim 1, characterized in that, The scale range of the Mexican wavelet transform described in step S10 is 1-100, ensuring coverage of the wavelength range corresponding to the main frequency of the target layer segment.

8. The method for effective reservoir identification based on the principle of resonance imaging for thin interlayered formations according to claim 1, characterized in that, The spectral feature comparison in step S10 is based on a discrimination threshold. When the spectral energy value is greater than or equal to the discrimination threshold, it is determined to be an effective reservoir; otherwise, it is a dry reservoir or a non-reservoir.

Images