Block stratum pressure prediction system and method based on three-dimensional seismic information

Abstract

The present application relates to oil and gas exploration and development technical field, especially to a block stratum pressure prediction system and method based on three-dimensional seismic information, comprising a data acquisition module, a depth domain conversion module, a three-dimensional gridding work area modeling module, a dynamic lithology volume generation module, a sparse survey line compaction modeling module, a three-dimensional pressure field calculation module and a dynamic regulation and control module. The three-dimensional grid model with well location markers is constructed to realize full space coverage and eliminate the limitations of two-dimensional data. The sparse survey line strategy controlled by structural features is used to process pure mudstone node data, which significantly reduces the calculation load. Based on the dynamic lithology probability threshold, reliable data sources are screened, and the lithology correction coefficient of sandstone area is combined to improve the pressure prediction accuracy. The real-time drilling DC index is used to close-loop verify the prediction results and trigger the regulation and control instructions. The present application effectively solves the technical bottlenecks of insufficient spatial coverage, low calculation efficiency and large prediction deviation in complex geological structure area.

Description

Technical Field

[0001] This invention relates to the field of oil and gas exploration and development technology, and in particular to a block formation pressure prediction system and method based on three-dimensional seismic information. Background Technology

[0002] 3D seismic exploration involves deploying a regular grid of sensor arrays on the surface of the exploration area and artificially generating seismic waves to propagate downwards. When the waves encounter subsurface interfaces of different densities or lithologies, some of the energy is reflected back to the surface and recorded by multiple receivers. After integrating high-density spatial sampling and time-delay information, processing and inversion techniques are used to reconstruct the three-dimensional geometry and physical property distribution of the geological body. This three-dimensional data volume eliminates the geometric limitations of two-dimensional profiles, significantly improving the analytical capability for the spatial evolution of complex reservoir structures or fault systems. It is widely used in oil and gas reservoir identification, mineral resource assessment, and regional geological model construction.

[0003] Existing formation pressure prediction methods rely on well logging data or 2D seismic data. The former only provides one-dimensional vertical information on discrete well locations, while the latter can only generate local 2D profiles, resulting in a severe lack of 3D spatial sampling density. When facing complex geological structures (such as fault cutting, lateral lithological abrupt changes, or lens development), the lack of continuous spatial data makes it difficult to establish a 3D pressure field model that conforms to geological laws. Furthermore, limited by data processing capabilities, existing technologies typically employ simplified global compaction curves or static lithological threshold values, failing to accurately characterize the velocity-pressure response relationship in heterogeneous formations. For example, in sandstone-mudstone interbedded blocks, directly fitting a compaction trend line without effectively removing sandstone data points will distort the slope of the normal pressure trend line; relying solely on sparse well data for pressure interpolation fails to identify the boundaries of overpressure storage boxes between wells, leading to well kicks or lost circulation risks during drilling due to pressure prediction deviations. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a block formation pressure prediction system and method based on three-dimensional seismic information, which solves the problems of limited three-dimensional spatial coverage, low computational efficiency, and insufficient prediction accuracy of existing formation pressure prediction methods.

[0005] To solve the above-mentioned technical problems, the specific technical solution of the present invention is as follows:

[0006] In a first aspect, the present invention provides a block formation pressure prediction system based on three-dimensional seismic information, comprising:

[0007] The data acquisition module acquires three-dimensional seismic impedance or velocity data volumes, wellbore acoustic time difference and formation layer data, performs data anomaly removal and unit unification operations, and outputs standardized seismic data volumes and depth-aligned well data.

[0008] The depth domain conversion module receives the standardized seismic data volume and wellbore acoustic time difference output by the data acquisition module, performs time-depth conversion to generate a depth domain acoustic velocity volume, and establishes a wave impedance to acoustic time difference conversion function.

[0009] The three-dimensional mesh chemical zone modeling module constructs a three-dimensional Cartesian mesh model with well location markers based on the depth domain acoustic velocity volume generated by the depth domain conversion module and the depth-aligned well data output by the data acquisition module.

[0010] The dynamic lithology scale generation module is based on the three-dimensional Cartesian grid model constructed by the three-dimensional grid chemical zone modeling module. It extracts the well-side acoustic velocity from the model, analyzes the probabilistic correlation between acoustic velocity and lithology, generates dynamic lithology boundary thresholds, and filters pure mudstone grid nodes.

[0011] The thinning survey line compaction modeling module, based on the depth domain acoustic velocity volume generated by the depth domain conversion module and the pure mudstone node markers selected by the dynamic lithology scale generation module, selects the thinning survey line under the control of structural features, fits a single-track compaction curve, and interpolates to generate a compaction parameter field for the entire work area.

[0012] The three-dimensional pressure field calculation module calculates pore pressure values ​​and performs lithological correction based on the compaction parameter field generated by the thinning survey line compaction modeling module, the depth domain acoustic velocity volume generated by the depth domain conversion module, and the dynamic lithology boundary threshold generated by the dynamic lithology volume generation module, and outputs three-dimensional pressure data volume and fluid potential distribution.

[0013] The dynamic control module, based on the three-dimensional pressure data volume output by the three-dimensional pressure field calculation module and combined with the real-time drilling DC index, matches the drilling trajectory pressure value with the measured data and triggers the mud density control command.

[0014] Furthermore, in the block formation pressure prediction system based on three-dimensional seismic information described in this invention, the data acquisition module is configured as follows:

[0015] Analyze the trace header information of seismic SGY files and perform amplitude statistics on anomalous seismic traces in seismic impedance or velocity data volumes;

[0016] Anomalies with amplitude values ​​exceeding ±3 standard deviations are marked as NULL values ​​to obtain the purified seismic data volume.

[0017] Unit conversion was performed on the wellbore acoustic time difference data to unify the time difference units from different sources to μs / ft; simultaneously, the formation layer data was aligned with the wellbore depth reference system and corrected to the depth datum of the work area, with the depth error controlled within 0.1%, to generate depth-aligned well data;

[0018] The system integrates and cleansed seismic data, sonic transit time data in a unified unit, and depth-aligned well data to output standardized seismic data and depth-aligned well data.

[0019] Furthermore, in the block formation pressure prediction system based on three-dimensional seismic information described in this invention, the depth domain conversion module is configured as follows:

[0020] The standardized seismic data volume output by the data acquisition module is received as input time-domain data;

[0021] The time-domain data is subjected to formation dip correction, and the layer velocity integration method is applied to generate the depth-domain acoustic velocity volume.

[0022] The seismic amplitude sequence of the well-side channel in the depth domain acoustic velocity volume is extracted and convolved with the acoustic time difference curve output by the data acquisition module to generate a synthetic record;

[0023] The synthetic record and the actual wellbore seismic amplitude envelope are compared by time-shift scanning to construct a wave impedance to acoustic time difference conversion function.

[0024] Furthermore, in the block formation pressure prediction system based on three-dimensional seismic information described in this invention, the three-dimensional grid chemical industrial zone modeling module is configured as follows:

[0025] Based on the depth domain acoustic velocity volume generated by the depth domain transformation module, a grid coordinate system is defined: the X-axis corresponds to the direction of the main seismic survey line, the Y-axis corresponds to the direction of the transverse survey line, and the Z-axis is the depth axis.

[0026] The depth-domain acoustic velocity volume is mapped to the grid nodes, and the depth-aligned well data output by the data acquisition module is called simultaneously;

[0027] Based on the well trajectory coordinates in the depth-aligned well data, calculate the coordinates of the intersection point between the well trajectory and the grid cell, and generate well location markers;

[0028] By combining seismic tectonic interpretation data, fault attribute labels are added to the fault zone grid cells, and cross-fault data interpolation is prohibited.

[0029] Furthermore, in the block formation pressure prediction system based on three-dimensional seismic information described in this invention, the dynamic lithology quantification module is configured as follows:

[0030] From the three-dimensional Cartesian grid model with well location markings constructed by the three-dimensional grid chemical zone modeling module, the sonic velocity values ​​of the well penetration grid nodes and the corresponding lithology labels are extracted.

[0031] The data is divided into segments based on vertical depth intervals, and the velocity and lithology probability distributions within each depth segment are calculated using a Bayesian classification algorithm.

[0032] A dynamic boundary curve is constructed with lithological probability >70% as the dividing point;

[0033] Based on the dynamic boundary curve, traverse all grid nodes in the work area, filter grid nodes with velocity values ​​lower than the boundary threshold corresponding to the current depth, and mark them as pure mudstone nodes.

[0034] Furthermore, in the block formation pressure prediction system based on three-dimensional seismic information described in this invention, the thinning survey line compaction modeling module is configured as follows:

[0035] Combining the fault attribute labels and seismic tectonic interpretation data in the three-dimensional grid chemical zone modeling module, one representative survey line is extracted every 50 lines along the main survey line direction to cover the structural high points, low points and fault transition zones of the work area.

[0036] Based on the pure mudstone node markers selected by the dynamic lithology scale generation module, the sonic transit time data corresponding to the pure mudstone grid nodes are extracted along the thinning survey line.

[0037] Fitting the logarithmic linear compaction equation of sonic transit time versus depth for pure mudstone;

[0038] The Kriging interpolation algorithm with structural orientation variation function constraints is used to extend the parameters of the single compaction curve to the entire work area, generating a three-dimensional compaction parameter field.

[0039] Furthermore, in the block formation pressure prediction system based on three-dimensional seismic information described in this invention, the three-dimensional pressure field calculation module is configured as follows:

[0040] The three-dimensional compaction parameter field generated by the thinning survey line compaction modeling module is invoked to calculate the theoretical acoustic transit time corresponding to each grid cell.

[0041] Obtain the depth domain acoustic velocity volume generated by the depth domain conversion module, and extract the actual acoustic time difference value from it;

[0042] Based on the deviation of the actual acoustic wave time difference from the theoretical acoustic wave time difference, the pore pressure value is generated.

[0043] For grid nodes in sandstone development areas, pressure anomaly amplitudes are reduced by using lithological probability distribution data output by the dynamic lithology quantification module.

[0044] Output three-dimensional pressure data volume and fluid potential distribution with lithological correction labels.

[0045] Furthermore, in the block formation pressure prediction system based on three-dimensional seismic information described in this invention, the dynamic control module is configured as follows:

[0046] Based on the three-dimensional Cartesian mesh model constructed by the three-dimensional mesh chemical zone modeling module, the mesh cells traversed by the drilling trajectory are located.

[0047] The predicted pore pressure value is extracted from the three-dimensional pressure data volume output by the three-dimensional pressure field calculation module;

[0048] Receive real-time drilling DC index data stream and convert it to generate an equivalent formation pressure curve;

[0049] By comparing the predicted pore pressure value with the real-time data of the equivalent formation pressure curve, when the pressure coefficient is >1.2 and the relative deviation is >10%, a mud density increment calculation command is triggered.

[0050] Furthermore, the block formation pressure prediction system based on three-dimensional seismic information described in this invention also includes a structural boundary processing submodule, which is integrated into the three-dimensional mesh chemical industrial zone modeling module. The structural boundary processing submodule is configured to: when a reverse fault footwall region is detected, trigger the triangulation method to refine the local mesh based on the fault attribute label.

[0051] Identify work area boundaries with no data grid cells, fill with NULL values ​​and disable parameter interpolation;

[0052] When calling depth domain acoustic velocity volume data, if the acoustic velocity of the salt rock layer is >4500m / s, the input of the corresponding time difference data of the salt rock layer will be automatically blocked.

[0053] Secondly, the block formation pressure prediction method based on three-dimensional seismic information provided by the present invention, applied to the aforementioned block formation pressure prediction system based on three-dimensional seismic information, includes:

[0054] Step 1: Acquire three-dimensional seismic impedance or velocity data volume, wellbore acoustic time difference and formation layer data, perform data anomaly removal and unit unification operation, and output standardized seismic data volume and depth-aligned well data.

[0055] Step 2: Receive the standardized seismic data volume and wellbore acoustic time difference output by the data acquisition module, perform time-depth conversion to generate depth domain acoustic velocity volume, and establish a wave impedance to acoustic time difference conversion function;

[0056] Step 3: Based on the depth domain acoustic velocity volume generated by the depth domain conversion module and the depth-aligned well data output by the data acquisition module, construct a three-dimensional Cartesian mesh model with well location markings.

[0057] Step 4: Based on the three-dimensional grid chemical zone modeling module, a three-dimensional Cartesian grid model is constructed, and the well-side acoustic velocity is extracted from the model. The probabilistic correlation between acoustic velocity and lithology is analyzed, a dynamic lithology boundary threshold is generated, and pure mudstone grid nodes are screened.

[0058] Step 5: Based on the depth domain acoustic velocity volume generated by the depth domain conversion module and the pure mudstone node markers selected by the dynamic lithology scale generation module, select the thinning survey line under the control of structural features, fit the single-channel compaction curve, and interpolate to generate the compaction parameter field of the entire work area.

[0059] Step 6: Based on the compaction parameter field generated by the thinning survey line compaction modeling module, the depth domain acoustic velocity volume generated by the depth domain conversion module, and the dynamic lithology boundary threshold generated by the dynamic lithology volume generation module, calculate the pore pressure value and perform lithology correction, and output the three-dimensional pressure data volume and fluid potential distribution.

[0060] Step 7: Based on the three-dimensional pressure data volume output by the three-dimensional pressure field calculation module, and combined with the real-time drilling DC index, match the drilling trajectory pressure value with the measured data, and trigger the mud density control command.

[0061] Beneficial effects of this invention;

[0062] This invention integrates comprehensive seismic and well data by constructing a three-dimensional gridded geological model, eliminating the spatial limitations of two-dimensional profiles; it employs a structural feature-controlled thinning survey line strategy to process pure mudstone node data, reducing computational load; it filters reliable data sources based on dynamic lithology probability thresholds and eliminates false overpressure misjudgments by combining lithology correction coefficients in sandstone development areas; and it systematically solves three major problems—spatial coverage blind spots, low computational efficiency, and insufficient accuracy—through real-time drilling data closed-loop verification of pressure prediction results. In applications in thin interbedded sandstone and mudstone blocks, the accuracy of three-dimensional pressure field identification is improved, the boundary delineation error of overpressure storage boxes in complex reverse fault zones is reduced, and real-time control response speed reaches the second level, significantly reducing the well kick accident rate. Attached Figure Description

[0063] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the drawings without creative effort.

[0064] Figure 1 A flowchart illustrating a block formation pressure prediction method based on three-dimensional seismic information provided in an embodiment of the present invention. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The technical solutions provided by various embodiments of this invention will be described in detail below with reference to the accompanying drawings. To better understand the objectives of this invention, it will be described in further detail below.

[0066] This invention provides a block formation pressure prediction system based on 3D seismic information, and details its implementation. The time-domain acoustic impedance data acquired during 3D seismic exploration is stored in SGY format, covering the main survey lines 50-200 and the transverse survey lines 100-400 within the work area. Simultaneously, sonic transit time curves (raw units μs / m) and Ng group formation layer data from three wells are acquired. The geographic coordinate system uses WGS84 projection, and the depth datum is set to mean sea level.

[0067] Data standardization phase: Seismic SGY file trace header information was analyzed. The amplitude value of trace 120 in Inline 73 was found to exceed the range of ±3 standard deviations of the average amplitude in the work area; this was marked as NULL and interpolated for compensation. The sonic transit time units of the three wells were uniformly converted to μs / ft to eliminate systematic errors from different logging series. Wellbore depth was corrected based on the difference between the core elevation and sea level (+5.2m), generating depth-aligned well data. The seismic data volume, standardized sonic transit time, and depth-corrected layered data were integrated and purified to output a standardized data stream.

[0068] Depth Domain Conversion and Mesh Modeling: A dip-corrected layer velocity integral method was used to convert time-domain data. Specifically, seismic interpretation horizons were introduced as control surfaces in the footwall region of the F1 fault, and layer velocity fields were constructed segmentally. A synthetic record was generated by convolving the seismic amplitude sequence from the Inline100 well bypass with the corrected acoustic transit time curve. A linear transformation function between acoustic impedance and acoustic transit time (correlation coefficient 0.85) was established through time-shift scanning. After generating the depth-domain acoustic velocity volume, a Cartesian mesh model with a horizontal grid spacing of 25m × 25m and a vertical stratification accuracy of 10m was constructed. Well locations were mapped to mesh nodes based on well trajectory coordinates. Fault attribute labels were added to the meshes on both sides of the F1 fault, and cross-fault interpolation was blocked.

[0069] Lithology identification and compaction modeling: The sonic velocity (3120 m / s) and lithology labels (85% sandstone) of the W1 well penetrating the grid were extracted from a 3D Cartesian grid model. Data segments were divided at 500 m vertical intervals. A Bayesian algorithm was applied to calculate the velocity-lithology probability distribution at depths of 3000-3500 m, determining the sandstone-mudstone boundary threshold to be 3150 m / s (mudstone probability 72%). Grid nodes with velocity values ​​below this threshold were marked as pure mudstone nodes. Combined with fault attribute labels, four survey lines (Inline50 / 100 / 150 / 200) were extracted to cover structural highs / lows. The sonic transit time of pure mudstone nodes was extracted along Inline100, and a log-linear compaction equation was fitted (slope k=0.28, R²=0.91). Kriging interpolation constrained by the structural strike variation function was used to generate a three-dimensional compaction parameter field with a slope of 0.22-0.25 on the footwall and 0.28-0.32 on the hanging wall.

[0070] Pressure Calculation and Real-Time Control: The theoretical sonic transit time is calculated based on the compaction parameter field, and the actual transit time value extracted from the depth domain velocity volume is compared. In sandstone development areas (probability > 80%), a lithology correction coefficient is applied (reducing the pressure anomaly amplitude by 15%), and a three-dimensional pressure data volume with correction labels is output. When a directional well is drilled to 3250m (Ng group), the predicted pore pressure coefficient of 1.28 is extracted from the grid model, and the real-time DC index is simultaneously converted into an equivalent pressure coefficient of 1.32. When the pressure coefficient is detected to be > 1.2 and the relative deviation (3%) continues to expand to the 10% threshold, a command is triggered to increase the mud density from 1.20 g / cm³ to 1.25 g / cm³. For high-velocity salt rock layers (velocity > 4500 m / s), their interference with the compaction model is automatically shielded.

[0071] Performance Verification: In a drilling area with a developed thrust fault, the 3D pressure field model identified the closed overpressure body in the footwall of the F1 fault, with an error of less than 5% compared to the actual drilling pressure test. The thinning survey line strategy improved computational efficiency by 8 times, and the pure mudstone screening mechanism reduced the slope error of the compaction trend line from 0.32 to 0.28, resulting in a decrease in the drilling kick rate.

[0072] In the application of 3D seismic exploration data, the data acquisition module first parses the trace header information of the SGY format file to identify the spatial distribution range and sampling parameters of the seismic impedance or velocity data volume. When performing amplitude statistical analysis on each seismic trace within the data volume, the standard deviation threshold method is used to detect outliers: when the amplitude value of a seismic trace exceeds the average amplitude of the work area by ±3 times the standard deviation, it is marked as NULL and weighted interpolation compensation is performed on adjacent traces. Wellbore acoustic transit time data, derived from logging curves of multiple wells, needs to be uniformly converted to the standard unit μs / ft to eliminate systematic biases between different logging series. Formation layering data and the wellbore depth reference system are aligned to the work area depth datum using linear offset correction. The correction amount is calculated based on the elevation difference between the wellhead core and the datum, with the depth error controlled within 0.1%. Finally, the seismic volume, standardized acoustic transit time, and depth-corrected well data are integrated and purified to form a standardized output with a unified spatial reference.

[0073] The depth domain conversion module receives the standardized seismic data volume and uses it as the time domain input source data. A layer velocity integration method is used to perform time-depth conversion. In fault-developed areas, seismic interpretation horizons are introduced as control surfaces for velocity modeling, and a segmented layer velocity field is constructed to address formation inversion distortion. The seismic amplitude sequence from the wellbore is extracted from the converted depth domain acoustic velocity volume and convolved with the standardized acoustic transit time curve to generate a synthetic seismic record. Time-shift scanning technology is used to compare the amplitude envelope morphology of the synthetic record with that of the actual seismic trace, dynamically adjusting the time offset and amplitude scaling factor to establish a quantitative conversion relationship between wave impedance and acoustic transit time. This process outputs a depth domain velocity volume to eliminate time propagation distortion and provides a physical basis for attribute conversion.

[0074] Based on the spatial distribution characteristics of acoustic velocity volumes in the depth domain, a Cartesian grid coordinate system is defined: the X-axis is parallel to the main seismic survey line, the Y-axis is parallel to the transverse survey line, and the Z-axis extends along the depth direction. A horizontal grid spacing of 25m × 25m and a vertical stratification accuracy of 10m are set to map velocity attribute values ​​to grid nodes. The spatial coordinates of well trajectories from depth-aligned well data are used, and cubic spline interpolation is employed to calculate the intersection points of the trajectories and grid cells, generating well location marker attributes. Combining seismic tectonic interpretation results, fault attribute labels are added to grid cells in fault polygon-covered areas, and interpolation isolation zones are set to prevent cross-fault data transmission. This grid model integrates seismic attributes, well location distribution, and tectonic boundary information, constructing a unified spatial data carrier for the entire work area.

[0075] The dynamic lithology quantification module extracts the sonic velocity values ​​and corresponding lithology labels of well penetration grid nodes from a 3D Cartesian grid model with well location markers. The lithology labels are derived from well logging interpretation data and have been associated with the corresponding nodes during the grid mapping stage. Data segments are divided into 500-meter vertical depth intervals. Within each depth segment, a Bayesian classification algorithm is applied: the frequency of mudstone and sandstone occurrences within different velocity ranges is statistically analyzed, and the conditional probability distribution of lithology under a given velocity condition is calculated. A dynamic boundary curve varying with depth is constructed using a lithology probability greater than 70%. All grid nodes in the entire work area are traversed, and the sonic velocity value of the current node is compared with the boundary threshold for its depth. Nodes with velocity values ​​below the threshold are marked as pure mudstone nodes. This dynamic threshold mechanism adapts to the variation of formation velocity with depth, avoiding lithology misjudgments caused by fixed threshold values.

[0076] The thinning survey line compaction modeling module first identifies key areas such as structural highs, lows, and fault transition zones in the work area based on fault attribute labels in the 3D Cartesian grid model and seismic tectonic interpretation data. One representative survey line is extracted every 50 lines along the main survey line direction to achieve full coverage of structural features. Based on the pure mudstone node markers output by the dynamic lithology quantification module, sonic transit time data of the corresponding grid is extracted along the thinning survey lines. The least squares method is used to fit the logarithmic linear relationship equation between sonic transit time and depth in pure mudstone to obtain single-channel compaction trend line parameters. The spatial correlation of geological bodies is quantified using the structural strike variogram, and the constrained Kriging interpolation algorithm extrapolates the single-channel parameters to the entire work area, generating a 3D compaction parameter field. This strategy improves computational efficiency while maintaining the continuity of geological patterns through structurally controlled data downsampling.

[0077] The 3D pressure field calculation module calls the 3D compaction parameter field generated by the thinning survey line compaction modeling module to calculate the theoretical acoustic transit time of each grid cell under normal compaction conditions. The actual acoustic transit time value is extracted from the depth domain acoustic velocity volume output by the depth domain conversion module. Based on the deviation between the actual and theoretical transit times, the compaction balance principle is applied to generate pore pressure values. For grid nodes in sandstone development areas, lithological probability distribution data output by the dynamic lithology volumetric module is referenced: when the sandstone probability exceeds a threshold, pressure anomaly amplitude correction is applied based on the probability value ratio (e.g., 85% sandstone probability corresponds to a 15% amplitude reduction) to eliminate false overpressure deviations caused by high-velocity sandstone. Finally, a 3D pressure data volume and fluid potential distribution map with lithological correction markers are output, where the correction markers record whether the pressure value has undergone lithological correction. This process achieves dynamic accuracy control in pressure calculation for heterogeneous strata.

[0078] The dynamic control module locates the grid cells traversed by the drilling trajectory based on a 3D Cartesian grid model and matches the corresponding grid positions in real time according to the spatial coordinates of the well trajectory. It extracts the predicted pore pressure value of the target grid from the 3D pressure data volume, generating a pressure profile curve that varies with depth. Simultaneously, it receives the DC exponential data stream transmitted from the drilling site and converts it using the Eaton model to generate an equivalent formation pressure curve. By comparing the real-time data points of the predicted pore pressure curve and the DC exponential conversion curve, when the pressure coefficient exceeds 1.2 and the relative deviation between the two is consistently greater than 10%, a mud density increment calculation command is triggered. This command dynamically adjusts the mud density value based on the overpressure amplitude, forming a closed-loop control chain from pressure prediction to engineering control. This module overcomes the limitations of traditional post-event control, achieving early warning of drilling risks.

[0079] The boundary processing submodule is integrated into the 3D mesh chemical zone modeling module. During the mesh construction phase, it monitors the footwall region of thrust faults in real time: when the fault attribute label indicates a dip angle > 30°, it automatically triggers a triangulation algorithm to refine the local mesh to a 10m spacing, accurately representing the fault morphology. After identifying mesh cells with no data coverage at the work area boundary, it fills in NULL values ​​and sets interpolation masking markers to prevent boundary distortion caused by parameter extrapolation. It uses depth-domain acoustic velocity volume data for stratigraphic scanning; when the average acoustic velocity of a certain stratigraphic layer is consistently > 4500m / s and horizontally continuous, it is identified as a salt rock layer, and its acoustic transit time input is automatically masked. This mechanism optimizes the mesh structure and data processing rules for special geological structures, ensuring the reliability of pressure field modeling in complex geological areas.

[0080] The optimized mesh model of the structural boundary treatment submodule provides a precise spatial positioning basis for dynamic control. Mesh refinement in the thrust fault zone reduces drilling trajectory positioning error to the 5-meter level, improving pressure extraction accuracy; NULL value filling at the boundaries prevents false pressure values ​​in data-free areas from interfering with control decisions; the salt rock layer shielding mechanism eliminates interference from high-speed anomalies on pressure conversion, ensuring the reliability of the DC index comparison benchmark. Real-time drilling pressure data from the dynamic control module verifies the rationality of the structural boundary treatment.

[0081] The following examples are based on typical application scenarios such as interbedded sandstone and mudstone blocks, thrust fault zones, and salt rock development areas, detailing the specific implementation process. In each example, the 3D seismic data coverage area is ≥200km², the seismic trace spacing is 25m×25m, and the vertical sampling interval is 2ms.

[0082] Example 1: Sandstone-Mudstone Interbedded Block: In a certain onshore work area, thin interbedded sandstone and mudstone of the Ng Formation are well-developed. The 3D seismic data volume contains Inline 200-500. The data acquisition module detected an amplitude anomaly (>3σ) in Inline 305, channel 80, marked NULL values, and interpolated for compensation. The original sonic transit time data of well W1 was in μs / m units, which was uniformly converted to μs / ft; it was corrected to sea level reference based on the core elevation +7.5m, with a depth error of 0.08%. The depth domain conversion module used the dip-corrected layer velocity integration method to establish a wave impedance-sonic transit time conversion function (correlation coefficient 0.82) in the well-side channel. 3D mesh modeling defined 10m vertical layers. When mapping the W1 well trajectory to the mesh nodes, a deviation of >5m between the well inclination segment and the mesh was found, triggering local mesh refinement. The dynamic lithology measurement module calculated the sandstone-mudstone boundary velocity to be 3150 m / s (mudstone probability 75%) in the 3000-3500 m depth range, screening out pure mudstone nodes accounting for 68% of the work area. One thinning survey line was extracted every 50 lines (6 lines in total), and the slope of the fitted compaction curve was k=0.29. The pressure calculation module applied an 18% correction to the sandstone area (probability > 85%), outputting a pressure coefficient of 1.25. At a drilling depth of 3250 m, the DC index showed a pressure coefficient of 1.32. A deviation > 10% triggered an increase in mud density from 1.18 g / cm³ to 1.23 g / cm³, mitigating the risk of well kick.

[0083] Example 2: Thrust Fault Zone: An F2 thrust fault was developed in a certain work area, with the footwall strata inverted. After the structural boundary processing submodule detected the footwall region, triangulation was used to refine the mesh from 25m to 10m. The 3D mesh modeling module added fault attribute tags to prevent cross-fault interpolation. The dynamic lithology quantification module corrected the boundary velocity to 3400m / s (80% mudstone probability) in the strongly compacted area of ​​the footwall (depth 4000m). Twelve thinning survey lines were extracted along the fault strike, and Kriging interpolation was used to generate a parameter field with a footwall compaction slope of 0.24 and a hanging wall slope of 0.31. 3D pressure field calculations showed a closed overpressure body in the footwall (pressure coefficient 1.45), with an error of <5% verified by actual drilling. The dynamic control module matched the pressure curve in real time when traversing the fault, adjusting the mud density to 1.40g / cm³ 20m in advance when encountering a high-pressure layer in the footwall.

[0084] Example 3: Salt-rock high-velocity layer interference zone: In the salt dome development area, the structural boundary processing submodule identified salt-rock layers with velocities >4800 m / s and automatically masked their sonic transit time data. The data acquisition module removed 23 abnormal seismic traces on the salt dome flanks. The depth domain conversion module used layer velocity modeling constrained by the salt-rock layer to reduce the depth error of the T4 reflection layer. The dynamic lithology quantification module excluded salt-rock interference when constructing the boundary curve in the subsalt strata (4500-5000m), setting the boundary velocity to 3800 m / s. Eight representative survey lines were extracted to avoid the salt dome boundary, and the compaction equation fitting R² > 0.9. The pressure calculation module output a subsalt overpressure body distribution map to guide well location optimization to avoid high-pressure areas, reducing the drilling accident rate.

[0085] Example 4: Efficiency Optimization in a Large-Scale Work Area: In a 300 km² work area, the thinning survey line compaction modeling module only processes 5% of the seismic trace data (existing methods require 100%). By dividing the area into 30 survey lines based on structural units, the Kriging interpolation time was reduced from 72 hours to 9 hours. The 3D mesh modeling module fills in NULL values ​​for data-free boundary areas to avoid extrapolation errors. The dynamic control module collects DC index data streams every 10 seconds and matches the grid pressure values ​​in real time. During drilling in the horizontal section of well P1, when the pressure deviation is detected to increase from 5% to 12%, the system triggers a mud density control command within 2 seconds, controlling the bottom hole pressure fluctuation within ±0.02 g / cm³.

[0086] This invention systematically addresses three major problems—limited 3D spatial coverage, low computational efficiency, and insufficient prediction accuracy—through the following technical approaches:

[0087] A 3D mesh-based chemical zone modeling module was constructed, fusing depth-domain acoustic velocity volumes with depth-aligned well data into a Cartesian mesh model with well location markers, achieving continuous spatial coverage across the entire work area. The mesh coordinate system (X / Y axes corresponding to the main seismic survey line / lateral survey line, Z axis for depth) eliminates the limitations of 2D profiles. A boundary processing submodule employs triangulation to locally refine the mesh in the footwall of thrust faults, blocking cross-fault interpolation and resolving data blind spots in fault-blocking areas. A salt-rock shielding mechanism automatically filters interfering data from layers with velocities >4500 m / s, ensuring spatial integrity in complex structural zones.

[0088] The thinning survey line compaction modeling module extracts 5%-10% representative survey lines based on structural characteristics (high points / low points / fault zones) (e.g., one line for every 50 main survey lines), processing only pure mudstone node sonic transit time data. Through Kriging interpolation constrained by the structural strike variogram, the parameters of a single compaction curve are extended to the entire work area, avoiding the resource consumption of traditional global computation. The dynamic lithology quantification module divides data by depth segment and applies a Bayesian classification algorithm to process the velocity-lithology relationship of each segment in parallel, reducing unnecessary computation by more than 75%.

[0089] The dynamic lithology volume generation module constructs a threshold curve with a lithology probability >70% as the dynamic dividing point, replacing the fixed lithology threshold. The 3D pressure field calculation module uses this threshold to screen pure mudstone nodes and establish a compaction model, eliminating interference from sandstone porosity effects. A correction coefficient related to lithology probability is applied to sandstone development areas (e.g., +18% correction when sandstone probability >85%) to resolve false overpressure misjudgments in high-velocity sandstone. The real-time control module compares the predicted values ​​of the 3D pressure data volume with the measured DC index curve. When the pressure coefficient >1.2 and the deviation >10%, mud density control is triggered, forming a closed-loop verification mechanism.

[0090] The above-mentioned technical approach systematically overcomes the shortcomings of existing methods through a logical chain of spatial global modeling, intelligent data thinning, dynamic parameter correction, and real-time feedback verification, achieving synergistic optimization of spatial coverage, efficiency, and accuracy.

Claims

1. A block formation pressure prediction system based on three-dimensional seismic information, characterized in that, include: The data acquisition module acquires three-dimensional seismic impedance or velocity data volumes, wellbore acoustic time difference and formation layer data, performs data anomaly removal and unit unification operations, and outputs standardized seismic data volumes and depth-aligned well data. The depth domain conversion module receives the standardized seismic data volume and wellbore acoustic time difference output by the data acquisition module, performs time-depth conversion to generate a depth domain acoustic velocity volume, and establishes a wave impedance to acoustic time difference conversion function. The three-dimensional mesh chemical zone modeling module constructs a three-dimensional Cartesian mesh model with well location markers based on the depth domain acoustic velocity volume generated by the depth domain conversion module and the depth-aligned well data output by the data acquisition module. The dynamic lithology scale generation module is based on the three-dimensional Cartesian grid model constructed by the three-dimensional grid chemical zone modeling module. It extracts the well-side acoustic velocity from the model, analyzes the probabilistic correlation between acoustic velocity and lithology, generates dynamic lithology boundary thresholds, and filters pure mudstone grid nodes. The dynamic lithology quantification generation module is configured as follows: From the three-dimensional Cartesian grid model with well location markings constructed by the three-dimensional grid chemical zone modeling module, the sonic velocity values ​​of the well penetration grid nodes and the corresponding lithology labels are extracted. The data is divided into segments based on vertical depth intervals, and the velocity and lithology probability distributions within each depth segment are calculated using a Bayesian classification algorithm. A dynamic boundary curve is constructed with lithological probability >70% as the dividing point; Based on the dynamic boundary curve, traverse all grid nodes in the work area, filter grid nodes whose velocity values ​​are lower than the boundary threshold corresponding to the current depth, and mark them as pure mudstone nodes. The thinning survey line compaction modeling module, based on the depth domain acoustic velocity volume generated by the depth domain conversion module and the pure mudstone node markers selected by the dynamic lithology scale generation module, selects the thinning survey line under the control of structural features, fits a single-track compaction curve, and interpolates to generate a compaction parameter field for the entire work area. The three-dimensional pressure field calculation module calculates pore pressure values ​​and performs lithological correction based on the compaction parameter field generated by the thinning survey line compaction modeling module, the depth domain acoustic velocity volume generated by the depth domain conversion module, and the dynamic lithology boundary threshold generated by the dynamic lithology volume generation module, and outputs three-dimensional pressure data volume and fluid potential distribution. The three-dimensional pressure field calculation module is configured as follows: The three-dimensional compaction parameter field generated by the thinning survey line compaction modeling module is invoked to calculate the theoretical acoustic transit time corresponding to each grid cell. Obtain the depth domain acoustic velocity volume generated by the depth domain conversion module, and extract the actual acoustic time difference value from it; Based on the deviation of the actual acoustic time difference from the theoretical acoustic time difference, the pore pressure value is generated. For grid nodes in sandstone development areas, the pressure anomaly amplitude is reduced by using the lithological probability distribution data output by the dynamic lithology quantification module; Output three-dimensional pressure data volume and fluid potential distribution with lithological correction labels; The dynamic control module, based on the three-dimensional pressure data volume output by the three-dimensional pressure field calculation module and combined with the real-time drilling DC index, matches the drilling trajectory pressure value with the measured data and triggers the mud density control command.

2. The block formation pressure prediction system based on three-dimensional seismic information according to claim 1, characterized in that, The data acquisition module is configured as follows: Analyze the trace header information of seismic SGY files and perform amplitude statistics on anomalous seismic traces in seismic impedance or velocity data volumes; Anomalies with amplitude values ​​exceeding ±3 standard deviations are marked as NULL values ​​to obtain the purified seismic data volume. Unit conversion was performed on the wellbore acoustic time difference data to unify the time difference units from different sources to μs / ft; simultaneously, the formation layer data was aligned with the wellbore depth reference system and corrected to the depth datum of the work area, with the depth error controlled within 0.1%, to generate depth-aligned well data; The system integrates and cleansed seismic data, sonic transit time data in a unified unit, and depth-aligned well data to output standardized seismic data and depth-aligned well data.

3. The block formation pressure prediction system based on three-dimensional seismic information according to claim 2, characterized in that, The depth domain conversion module is configured as follows: The standardized seismic data volume output by the data acquisition module is received as input time-domain data; The time-domain data is subjected to formation dip correction, and the layer velocity integration method is applied to generate the depth-domain acoustic velocity volume. The seismic amplitude sequence of the well-side channel in the depth domain acoustic velocity volume is extracted and convolved with the acoustic time difference curve output by the data acquisition module to generate a synthetic record; The synthetic record and the actual wellbore seismic amplitude envelope are compared by time-shift scanning to construct a wave impedance to acoustic time difference conversion function.

4. The block formation pressure prediction system based on three-dimensional seismic information according to claim 3, characterized in that, The 3D mesh chemical industrial zone modeling module is configured as follows: Based on the depth domain acoustic velocity volume generated by the depth domain transformation module, a grid coordinate system is defined: the X-axis corresponds to the direction of the main seismic survey line, the Y-axis corresponds to the direction of the transverse survey line, and the Z-axis is the depth axis. The depth-domain acoustic velocity volume is mapped to the grid nodes, and the depth-aligned well data output by the data acquisition module is called simultaneously; Based on the well trajectory coordinates in the depth-aligned well data, calculate the coordinates of the intersection point between the well trajectory and the grid cells, and generate well location markers; By combining seismic tectonic interpretation data, fault attribute labels are added to the fault zone grid cells, and cross-fault data interpolation is prohibited.

5. The block formation pressure prediction system based on three-dimensional seismic information according to claim 4, characterized in that, The thinning test line compaction modeling module is configured as follows: Combining the fault attribute labels and seismic tectonic interpretation data in the three-dimensional grid chemical zone modeling module, one representative survey line is extracted every 50 lines along the main survey line direction to cover the structural high points, low points and fault transition zones of the work area. Based on the pure mudstone node markers selected by the dynamic lithology scale generation module, the sonic transit time data corresponding to the pure mudstone grid nodes are extracted along the thinning survey line. Fitting the logarithmic linear compaction equation of sonic transit time versus depth for pure mudstone; By employing a Kriging interpolation algorithm constrained by a directional variogram, the parameters of a single compaction curve are extended to the entire work area, generating a three-dimensional compaction parameter field.

6. The block formation pressure prediction system based on three-dimensional seismic information according to claim 5, characterized in that, The dynamic control module is configured as follows: Based on the three-dimensional Cartesian mesh model constructed by the three-dimensional mesh chemical zone modeling module, the mesh cells traversed by the drilling trajectory are located. The predicted pore pressure value is extracted from the three-dimensional pressure data volume output by the three-dimensional pressure field calculation module; Receive real-time drilling DC index data stream and convert it to generate an equivalent formation pressure curve; By comparing the predicted pore pressure value with the real-time data of the equivalent formation pressure curve, when the pressure coefficient is >1.2 and the relative deviation is >10%, a mud density increment calculation command is triggered.

7. The block formation pressure prediction system based on three-dimensional seismic information according to claim 6, characterized in that, It also includes a boundary processing submodule, which is integrated into the 3D mesh chemical zone modeling module. The boundary processing submodule is configured to: when a reverse fault footwall region is detected, trigger the triangulation method to refine the local mesh based on the fault attribute label. Identify work area boundaries with no data grid cells, fill with NULL values ​​and disable parameter interpolation; When calling depth domain acoustic velocity volume data, if the acoustic velocity of the salt rock layer is >4500m / s, the input of the corresponding time difference data of the salt rock layer will be automatically blocked.

8. A method for predicting block formation pressure based on three-dimensional seismic information, applied to the block formation pressure prediction system based on three-dimensional seismic information as described in any one of claims 1 to 7, characterized in that, include: Step 1: Acquire three-dimensional seismic impedance or velocity data volume, wellbore acoustic time difference and formation layer data, perform data anomaly removal and unit unification operation, and output standardized seismic data volume and depth-aligned well data. Step 2: Receive the standardized seismic data volume and wellbore acoustic time difference output by the data acquisition module, perform time-depth conversion to generate depth domain acoustic velocity volume, and establish a wave impedance to acoustic time difference conversion function; Step 3: Based on the depth domain acoustic velocity volume generated by the depth domain conversion module and the depth-aligned well data output by the data acquisition module, construct a three-dimensional Cartesian mesh model with well location markings. Step 4: Based on the three-dimensional grid chemical zone modeling module, a three-dimensional Cartesian grid model is constructed, and the well-side acoustic velocity is extracted from the model. The probabilistic correlation between acoustic velocity and lithology is analyzed, a dynamic lithology boundary threshold is generated, and pure mudstone grid nodes are screened. Step 5: Based on the depth domain acoustic velocity volume generated by the depth domain conversion module and the pure mudstone node markers selected by the dynamic lithology scale generation module, select the thinning survey line under the control of structural features, fit the single-channel compaction curve, and interpolate to generate the compaction parameter field of the entire work area. Step 6: Based on the compaction parameter field generated by the thinning survey line compaction modeling module, the depth domain acoustic velocity volume generated by the depth domain conversion module, and the dynamic lithology boundary threshold generated by the dynamic lithology volume generation module, calculate the pore pressure value and perform lithology correction, and output the three-dimensional pressure data volume and fluid potential distribution. Step 7: Based on the three-dimensional pressure data volume output by the three-dimensional pressure field calculation module, and combined with the real-time drilling DC index, match the drilling trajectory pressure value with the measured data, and trigger the mud density control command.

Images