A method for constructing a three-dimensional velocity model of a formation.

Abstract

This invention relates to a method for constructing a three-dimensional formation velocity model, belonging to the field of three-dimensional seismic interpretation technology. First, this invention utilizes seismic data from the target area and the time-depth relationship of the pilot sections of vertical and horizontal wells to establish a well-seismic fusion three-dimensional velocity model. Then, based on the lithology encountered in the horizontal section, sonic logging curves, and the fitting relationship between sonic logging and lithology, the layer velocity at the horizontal section of the target layer in the target area is determined. This layer velocity is then interpolated and smoothed to obtain a horizontal section layer velocity model. Finally, this horizontal section layer velocity model is used to constrain the well-seismic fusion three-dimensional velocity model, resulting in the corresponding three-dimensional formation velocity model of this invention. This invention utilizes well logging information from the horizontal sections of horizontal wells to constrain and correct the three-dimensional velocity model, which not only improves the accuracy and stability of the lateral variations in the three-dimensional velocity model but also enhances the accuracy of low-amplitude structural interpretation.

Description

Technical Field

[0001] This invention relates to a method for constructing a three-dimensional velocity model of a formation, belonging to the field of three-dimensional seismic interpretation technology. Background Technology

[0002] Three-dimensional velocity models are key parameters in oil and gas exploration and development, directly affecting the accuracy of underground reservoir spatial distribution. Currently, the establishment of three-dimensional velocity models focuses on creating 3D seismic fusion velocity models under the constraints of seismic stacking velocity, logging velocity, and stratigraphic position. The logging data used in the analysis comes from the inclined sections (guide sections) of vertical, deviated, and horizontal wells. Horizontal sections cannot be calibrated using synthetic logging methods. This is because, in cases with horizontal sonic logging, although the layer velocity can be obtained from the sonic logging data, it only characterizes the lithological velocity of the drilled section. Directly using the velocity-constrained velocity model in areas with thin reservoirs and rapid lateral changes will cause anomalies and distortions in the layer velocity, resulting in a large error in the generated velocity model.

[0003] Chinese patent application CN105842736A discloses a method for constructing a formation velocity model. This method utilizes pre-stack depth migration velocity volumes, drilling time-depth curves, drilling seismic pseudovelocities, and a grid model of the work area to establish the formation velocity model. Firstly, the seismic velocity volume includes stacking velocity and migration velocity, with migration velocities including both time migration and depth migration velocities. Using pre-stack depth migration velocity to construct formation velocities results in a significantly narrower range. Secondly, horizontal well logging-while-drilling curves and geological logging information are used to determine the target layer where the horizontal well is located, and this target layer is calibrated on the formation velocity model to obtain the layering information of the target layer.

[0004] In the later stages of development, the seismic velocity spectrum is relatively clear. How to reconstruct the three-dimensional velocity model by utilizing the increasing amount of lithological information from horizontal well logging and well logging in the horizontal sections of horizontal wells, improve the accuracy of micro-structure description, and provide an accurate velocity model for fine oil and gas development is an urgent need for velocity models in the later stages of oil and gas field development. Summary of the Invention

[0005] The purpose of this invention is to provide a method for constructing a three-dimensional velocity model of a formation, so as to solve the problems of low accuracy and inability to reflect rapid lateral changes in the reservoir in current three-dimensional velocity models of formations.

[0006] To solve the above-mentioned technical problems, this invention provides a method for constructing a three-dimensional velocity model of a formation, the method comprising the following steps:

[0007] 1) Acquire seismic data, well logging data, and well logging data for the target area. Well data information includes the pilot sections of vertical and horizontal wells and the horizontal sections of horizontal wells in the target area.

[0008] 2) Construct a three-dimensional layer velocity model based on seismic data, and then correct the three-dimensional layer velocity model using well logging data from the pilot sections of vertical and horizontal wells to obtain a well-seismic fusion three-dimensional velocity model;

[0009] 3) Correct the acoustic transit time data of the horizontal section based on the acoustic transit time data of the horizontal well pilot section or the target layer of the surrounding vertical and inclined wells;

[0010] 4) Quantitative lithology inversion is performed using seismic and drilling data from the target area to obtain lithology inversion results for the target layer. Based on the relationship between sonic transit time and lithology sensitivity curve fitting formula, the inversion results of the horizontal segment within the target layer are converted into the layer velocity of the horizontal segment.

[0011] 5) Constrain the three-dimensional velocity model of well-seismic fusion based on the obtained layer velocity of the horizontal section to obtain the three-dimensional velocity model of the formation.

[0012] This invention first establishes a well-seismic fusion three-dimensional velocity model using seismic data from the target area and the time-depth relationships calibrated in the guide sections of vertical, deviated, and horizontal wells. Then, based on the lithology inversion of the horizontal section, the encountered lithology, sonic logging curves, and the fitting relationship between sonic logging and lithology, the layer velocity at the horizontal section of the target layer is determined. This layer velocity is then interpolated and smoothed to obtain a horizontal section layer velocity model. Finally, this horizontal section layer velocity model is used to constrain the well-seismic fusion three-dimensional velocity model, resulting in the corresponding formation three-dimensional velocity model of this invention. This invention utilizes well logging information from the horizontal section to constrain and correct the three-dimensional velocity model, which not only improves the accuracy and stability of the lateral variations in the three-dimensional velocity model but also enhances the accuracy of low-amplitude structural interpretation.

[0013] Furthermore, to ensure the accuracy of the horizontal segment acoustic time difference data, the correction process for the horizontal segment acoustic time difference data in step 3) is as follows:

[0014] A. Based on the logging data of vertical wells, horizontal well guide sections, and logging data in the target area, establish a sonic and lithological template;

[0015] B. Based on the acoustic transit time data of the horizontal well pilot section or the target formation of the surrounding vertical well, the acoustic transit time data of the horizontal section is corrected using acoustic waves and lithology templates. This ensures that the error value between the acoustic transit time measured in the horizontal section and the acoustic transit time measured in the target section of the vertical well or the horizontal well pilot section for the same lithology is controlled within a set percentage. The set percentage is 5%.

[0016] Furthermore, in order to ensure a one-to-one correspondence between acoustic transit time data and seismic data, the method also includes low-pass filtering and thinning of the corrected horizontal segment acoustic transit time data, so that each seismic sampling point corresponds to one acoustic transit time data.

[0017] Furthermore, the lithology-sensitive curves in step 3) are the natural gamma curve (clastic rocks) and the PE curve (carbonate rocks).

[0018] Furthermore, to ensure the accuracy of the vertical variation of the velocity data in the horizontal segment, the process for determining the layer velocity of the horizontal segment in step 4) is as follows:

[0019] a. Establish an acoustic transit time (AC)-lithology petrophysical template for the target section using acoustic transit time data from well logging and reservoir lithology sensitivity curves, and fit this template to obtain the relationship between acoustic transit time and reservoir lithology sensitivity curves;

[0020] b. Quantitative lithological inversion is performed using drilling and seismic data from the target area, and the lithological inversion results for the target layer are obtained based on the top and bottom interfaces of the target layer.

[0021] c. Based on the relationship between acoustic transit time and reservoir lithology sensitivity curve, the lithology inversion results of the target layer are converted into the layer velocity of the target layer in the horizontal segment. Attached Figure Description

[0022] Figure 1 This is a flowchart of the method for constructing the three-dimensional velocity model of the formation according to the present invention;

[0023] Figure 2 This is a target layer construction diagram obtained by using a velocity model without horizontal segments in the embodiment of the invention;

[0024] Figure 3 It is the target layer construction diagram obtained by establishing a velocity model using horizontal segments in the embodiment of the invention. Detailed Implementation

[0025] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0026] This invention addresses the problem of low lateral variation accuracy in current three-dimensional formation velocity models by providing a novel method for constructing such models. The method first obtains a layer velocity model using seismic data from the target area. Then, it corrects the layer velocity model based on the time-depth relationships calibrated in the guide sections of vertical, deviated, and horizontal wells within the target area, establishing a well-seismic fusion three-dimensional velocity model. Next, based on the lithology inversion of the horizontal section, the encountered lithology, sonic logging curves, and the fitting relationship between sonic logging and lithology, the layer velocity at the horizontal section of the target layer is determined. This velocity is then interpolated and smoothed to obtain a horizontal section layer velocity model. Finally, this horizontal section layer velocity model is used to constrain the well-seismic fusion three-dimensional velocity model, resulting in the corresponding three-dimensional formation velocity model of this invention. The implementation process of this method is as follows: Figure 1 As shown.

[0027] Taking a tight, low-permeability clastic gas field as an example, this gas field is a complex lithological and structural gas reservoir. Different structural traps exhibit significant variations in gas and water content, showing a distribution of water at the edge and bottom of the formation. The characterization of micro-structures has a significant impact on the deployment, design, and efficient production of horizontal well trajectories. In this example, the work area where the gas field is located already has 15 wells drilled, including 9 vertical wells and 6 horizontal wells (with pilot sections). The 3D seismic area is 80 square kilometers, and the accurate calibration is reliable. The specific implementation process of the 3D formation velocity model construction method of this invention is as follows.

[0028] 1. Acquire seismic data, well logging data, and well logging data for the target area.

[0029] The seismic data of this invention includes seismic processing and stacking velocity and stratigraphic data obtained from raw seismic data. The well logging data includes geological layer data and acoustic logging (AC) curves. The wells targeted by the well logging data include vertical wells, horizontal well pilot sections, and horizontal sections. In this embodiment, the well logging data includes logging data from 9 vertical wells, logging data from pilot sections of 6 horizontal wells, and logging data from 6 horizontal wells in the target area.

[0030] 2. Establish a three-dimensional velocity model for well-seismic fusion.

[0031] First, based on the seismic data obtained in step 1, the corresponding stacking velocity model is obtained. The stacking velocity model is dynamically corrected, and then the dynamically corrected stacking velocity model is modified, interpolated, and smoothed to form a three-dimensional root mean square velocity field. Through the Dix equation transformation, the stacking velocity model is converted into a three-dimensional layer velocity model. In order to improve the accuracy of the three-dimensional layer velocity model, this invention also needs to sample and process the three-dimensional layer velocity model profile and plane, and perform outlier point extraction and smoothing processing to continuously correct the three-dimensional layer velocity model.

[0032] Then, based on the relationship between drilling depth and the time-depth of the reflecting layer, a time-depth relationship curve is established at the well point using logging data (AC curves) from the pilot sections of vertical and horizontal wells in the target area. This relationship curve is then extended to the target area through linear interpolation to obtain a time-depth velocity model. This time-depth velocity model is used to constrain the three-dimensional layer velocity model to establish a well-seismic fusion velocity model, making the three-dimensional layer velocity model more closely approximate the actual formation conditions at the well point. The process of establishing the well-seismic fusion velocity model is extensively covered in published literature and related studies, and will not be described in detail here.

[0033] In this example, a synthetic record calibration was created based on the sonic transit time curves of 9 vertical wells and 6 horizontal well guide sections in the work area. A time-depth relationship curve was generated, and the time-depth relationship curve was used to establish a time-depth velocity model for the work area through linear interpolation. The time-depth velocity model was used to constrain the three-dimensional layer velocity model obtained from seismic data, so that the three-dimensional layer velocity model at the well point is closer to the actual formation.

[0034] 3. Correct the sonic transit time data of the horizontal section of the horizontal well.

[0035] Horizontal well drilling must ensure that the horizontal section remains within the target formation. Poor trajectory control can lead to penetration of the target formation into adjacent formations. Based on logging data from vertical and horizontal well pilot sections and lithological data in the target area, an acoustic and lithological template is established. The acoustic transit time of the horizontal section in the target area is corrected based on the established acoustic and lithological template, the acoustic characteristics of the pilot section of the horizontal well, or the target formation in surrounding vertical wells. This ensures that the error between the acoustic transit time measured in the horizontal section and that measured in the vertical well or pilot section for the same lithology is controlled within a set percentage. In this embodiment, the set percentage is 5%, but this percentage can be adjusted according to actual needs, generally keeping it below 10%. The corrected acoustic transit time data of the horizontal section is then low-pass filtered and thinned to ensure that each seismic sampling point corresponds to one acoustic logging (AC) data point. The low-pass filtering involves calculating the filtering wavelength based on the acquired surface grid, ensuring that 1-2 wavelengths are present in 3 consecutive surface grids. The low-pass filtering threshold range is calculated based on the velocities of the main lithological layers in the target formation.

[0036] This example establishes an acoustic and lithological petrophysical template based on well logging data from the pilot sections of 9 vertical wells and 6 horizontal wells in the target area, and uses this petrophysical template to correct the acoustic time difference in the horizontal sections of the 6 horizontal wells.

[0037] 4. Determine the formation velocity when the horizontal section of the horizontal well encounters the target formation.

[0038] Horizontal layer velocities can be obtained from acoustic transit time (AC) logging data, but the vertical sampling points for this velocity are too few, resulting in significant errors when directly used for velocity model constraint correction. Therefore, this invention determines the horizontal layer velocities through three-dimensional lithology inversion. The process is as follows: First, an AC-lithology petrophysical template for the target layer is established using AC logging data from vertical wells and pilot wells, along with reservoir lithology sensitivity curve logging data, and the relationship between AC and lithology sensitivity curves is fitted. Then, quantitative lithology inversion results are obtained using logging data and three-dimensional seismic data. The top and bottom interfaces of the target layer are used as control boundaries to output the lithology inversion results for the target layer. Finally, the horizontal layer inversion results within the target layer are converted into layer velocities using the fitting formula between AC and lithology sensitivity curves.

[0039] In this example, the reservoir lithology sensitive curves are the natural gamma (GR) curve and the photoelectric index (PE) curve. Different types of curves are used for different lithologies: clastic rocks use the GR curve, and carbonate rocks use the PE curve. Based on the acoustic transit time (AC) curve and natural gamma (GR) curve of 15 wells in the target area, corresponding AC-lithology petrophysical templates are established, and the relationship between the AC and GR curves is obtained through fitting, referred to as the AC-GR relationship. Based on the geostatistical stochastic simulation inversion method, quantitative GR inversion results are obtained using logging data (vertical and horizontal wells) and 3D seismic data. Using the top and bottom interfaces of the target layer as control boundaries, the lithology inversion results of the target layer are obtained, resulting in the inverted GR data. According to the AC-GR relationship, the inversion results of the horizontal section of the target layer are converted into layer velocities.

[0040] 5. Use the layer velocity of the target segment to constrain the three-dimensional velocity model obtained by well-seismic fusion in step 2, thereby reconstructing the three-dimensional velocity model of the formation.

[0041] The layer velocity data of the target layer obtained in step 4 are interpolated and smoothed to obtain the corresponding layer velocity model of the target layer. The obtained horizontal well section velocity model is used to constrain and correct the three-dimensional velocity model of well-seismic fusion in step 2.

[0042] In this example, the well seismic velocity model established by fusing 9 vertical wells and 6 pilot wells has significantly lower accuracy than the velocity model reconstructed by simultaneously involving the horizontal sections of 6 horizontal wells. Figure 2 , Figure 3 These are target layer structure diagrams obtained from velocity models built without horizontal segments and with horizontal segments, respectively. (Attached) Figure 3 The accuracy of the structural design was significantly improved, as verified using a newly drilled vertical well X2 that was not involved in the calculation. After the original three-dimensional velocity model was extended to a deeper depth, the top of the target sand body was 2428m. Using the newly reconstructed velocity model, the top of the sand body was 2378m. The actual depth at which the top of the sand body was encountered was 2382.5m, reducing the absolute error by 89.1%.

[0043] This invention utilizes the corrected sonic velocity of the horizontal section of a horizontal well to generate a velocity curve at intervals with seismic sampling points. By employing gamma-ray quantitative lithology inversion and sonic fitting formulas, the measured velocity of the horizontal section is extended to the target layer velocity, increasing the number of velocity control points and avoiding velocity distortion caused by directly using the horizontal section velocity. This method enables the constraint and correction of the three-dimensional velocity model using well logging information from the horizontal section, improving the accuracy and stability of the lateral variations in the three-dimensional velocity model and enhancing the accuracy of low-amplitude structural interpretation. It is beneficial for well location deployment based on lithological structures and structural oil and gas reservoir types, as well as for the design, adjustment, and optimization of horizontal well trajectories, thereby increasing the oil and gas reservoir encounter rate of productive wells.

Claims

1. A method for constructing a three-dimensional velocity model of a formation, characterized in that, The construction method includes the following steps: 1) Acquire seismic data, well logging data, and well logging data for the target area. The well logging data includes information on vertical wells, horizontal well pilot sections, and horizontal sections of horizontal wells in the target area. 2) Construct a three-dimensional layer velocity model based on seismic data, and then correct the three-dimensional layer velocity model using well logging data from the pilot sections of vertical and horizontal wells to obtain a well-seismic fusion three-dimensional velocity model; 3) Correct the acoustic transit time data of the horizontal section of the horizontal well based on the acoustic transit time data of the pilot section of the horizontal well or the target layer of the surrounding vertical well; the correction process is as follows: A. Based on the logging data of vertical wells, horizontal well guide sections, and logging data in the target area, establish a sonic and lithological template; B. Based on the acoustic transit time data of the horizontal well pilot section or the target layer of the surrounding vertical well, the acoustic transit time data of the horizontal section of the horizontal well is corrected using acoustic waves and lithological rock templates, so that the error value between the acoustic transit time measured in the horizontal section of the horizontal well and the acoustic transit time measured in the target section of the vertical well or the pilot section of the horizontal well of the same lithology is controlled within the set ratio. 4) Quantitative lithology inversion is performed using seismic and drilling data from the target area to obtain the lithology inversion results for the target interval. Based on the relationship between sonic transit time and the lithology sensitivity curve fitting formula, the inversion results of the horizontal segments of horizontal wells within the target interval are converted into the layer velocities of the horizontal segments of horizontal wells. The process for determining the layer velocities of the horizontal segments of horizontal wells is as follows: a. Establish an AC-lithological petrophysical template for the target interval using sonic transit time data from well logging data and reservoir lithology sensitivity curves, and fit this template to obtain the relationship between sonic transit time and reservoir lithology sensitivity curves; b. Quantitative lithological inversion is performed using drilling and seismic data from the target area, and the lithological inversion results for the target layer are obtained based on the top and bottom interfaces of the target layer. c. Based on the relationship between sonic transit time and reservoir lithology sensitivity curve, the lithology inversion results of the target interval are converted into the layer velocity of the target interval in the horizontal section of the horizontal well; 5) Constrain the three-dimensional velocity model of well-seismic fusion based on the layer velocity of the horizontal section of the horizontal well to obtain the three-dimensional velocity model of the formation.

2. The method for constructing a three-dimensional velocity model of a formation according to claim 1, characterized in that, The set ratio is 5%.

3. The method for constructing a three-dimensional velocity model of a formation according to claim 1, characterized in that, The method also includes low-pass filtering and thinning of the corrected horizontal segment acoustic time difference data so that each seismic sampling point corresponds to an acoustic time difference data.

4. The method for constructing a three-dimensional velocity model of a formation according to claim 1, characterized in that, The lithology-sensitive curves in step 3) are the natural gamma curve and the PE curve.

Images