A geophysical prediction method for salt-type polyhalite

Abstract

The present invention provides a geophysical prediction method for halite-type polyhalite. The method first analyzes the deformation and displacement characteristics of potassium-rich gypsum-salt rock layers under complex tectonic transformations, clarifying the deformation, displacement, and aggregation characteristics of potassium-containing gypsum-salt rocks. The method also demonstrates a linear positive correlation between salt layer thickness and the scale of halite-type polyhalite development. Furthermore, research is conducted on fine-detailed seismic sequence characterization techniques based on a global isochronous stratigraphic framework, establishing a global isochronous framework model. Furthermore, combined with rock physics and sensitive parameter analysis of potassium-rich strata, the method extracts and applies sensitive parameters of potassium-rich strata primarily through seismic inversion. Based on Bayesian wave impedance stochastic inversion, a convolutional neural network is used to directly predict potassium content curves, thereby enabling the description and prediction of the spatial distribution characteristics of potassium-rich strata. The present invention is well-conceived and can effectively describe and predict the spatial distribution characteristics of potassium-rich strata, providing technical support for the search for favorable potassium-rich strata.

Description

Technical Field

[0001] The present invention relates to the technical field of geological exploration, and in particular to a geophysical prediction method for halite-type polyhalite. Background Art

[0002] Previous discoveries of solid potash deposits mostly consisted of thin layers of polyhalite (gypsum-type) associated with anhydrite. These deposits, buried at great depths, were unavailable for mining and considered "dead ore." While some polyhalite (rock-type) associated with halite has been discovered, and this type of potash deposit is more soluble than gypsum-type deposits, it has not been given much attention.

[0003] Halite-type polyhalite develops in the gypsum salt rock of the Jiasi and Jiasi 5 members. Due to tectonic adjustments and transformations, the gypsum salt rock undergoes plastic deformation, with large variations in lateral thickness. The thickness of the gypsum salt strata and the lateral differences in seismic wave groups are large, making detailed structural interpretation difficult. Conventional seismic inversion methods based on layered modeling are not applicable.

[0004] Therefore, in order to achieve a comprehensive analysis and distribution characteristics description of halite-type polyhalite, it is urgent to design a geophysical prediction method for halite-type polyhalite to provide a basis for resource evaluation and geological exploration work deployment, so as to improve the ore-finding rate of exploration projects and save exploration costs. Summary of the Invention

[0005] In response to the technical problems existing in the above-mentioned background technology, the present invention proposes a geophysical prediction method for halite-type polyhalite. The method has a reasonable conception. By establishing a potassium-rich rock formation description inversion technical solution based on seismic, geological and well logging data, with potassium-rich rock sedimentary phase analysis and fine inversion constraint model as prior guidance, and with phase-controlled high-precision inversion as the core, the method can effectively realize the comprehensive analysis and distribution characteristic description of potassium-rich rock formations, and provide technical support for finding favorable potassium-rich formations.

[0006] To solve the above technical problems, the present invention provides a geophysical prediction method for halite-type polyhalite. The method first analyzes the deformation and displacement characteristics of potassium-rich gypsum salt rock layers under complex tectonic transformation, clarifies the deformation and displacement and aggregation characteristics of potassium-containing gypsum salt rocks, and finds that the thickness of the salt layer is linearly positively correlated with the scale of halite-type polyhalite development. On this basis, research is conducted on the fine characterization technology of seismic sequences based on a global isochronous stratigraphic framework, and a global isochronous framework model is established. Combined with the rock physics and sensitive parameter analysis of potassium-rich strata, the sensitive parameters of potassium-rich strata are extracted and applied mainly through seismic inversion. Based on the Bayesian wave impedance stochastic inversion, a convolutional neural network is used to directly predict the potassium content curve, thereby realizing the description and prediction of the spatial distribution characteristics of potassium-rich strata.

[0007] The geophysical prediction method for halite-type polyhalite mainly comprises the following steps:

[0008] (1) Establishment of seismic spatial constraint model for inversion of potassium-rich gypsum-salt rock layers and potassium salt layers

[0009] (1.1) Analysis of deformation and displacement characteristics of potassium-rich gypsum-salt rock layers under complex tectonic transformation;

[0010] (1.2) Construction of inversion sequence framework for potassium-rich gypsum-salt rock layers;

[0011] (2) Analysis of rock physical parameters of salt rock type polyhalite

[0012] (2.1) Testing and analysis of rock physical parameters of rock samples containing polyhalite layers;

[0013] (2.2) Establishment and analysis of rock physics models for potassium-rich formations;

[0014] (3) Geophysical spatial characterization of potash-bearing rock layers

[0015] (3.1) Analysis of seismic rock physical characteristics of potassium-rich formations;

[0016] (3.2) Analysis of the dominant properties of polyhalite-containing layers;

[0017] (3.3) Spatial characterization of salt-bearing strata through Bayesian wave impedance inversion and potassium-rich strata analysis;

[0018] (4) Prediction of halite-type polyhalite

[0019] (4.1) Inversion of sensitivity factors of potassium-rich rock formations using a volumetric finite element learning machine;

[0020] (4.2) Analysis of sensitive factors of polyhalite and potassium-rich brine formations in Puguang area.

[0021] The geophysical prediction method for halite-type polyhalite, wherein the specific process of step (1.1) is: utilizing the principle of equilibrium profile recovery, i.e., the "four principles" of consistent volume, consistent layer length, consistent displacement, and consistent shortening of rocks before and after deformation, to carry out fracture recovery according to the principle of oblique shear stratum deformation.

[0022] The geophysical prediction method of halite type polyhalite, wherein the specific process of carrying out fracture recovery is as follows:

[0023] (1.1.1) Select seismic sections that are roughly consistent with the direction of tectonic movement, that is, select seismic sections that are perpendicular to the regional structural trend;

[0024] (1.1.2) Use drilling data to perform velocity fitting analysis, establish an average velocity model, perform velocity verification, unify the velocity expression to a linear expression, and convert the time domain profile to the depth domain;

[0025] (1.2.3) Based on the results of horizon calibration and in accordance with the principles of structural interpretation, interpret the key horizons and faults of the entire section and construct a reasonable structural model;

[0026] (1.2.4) Based on the structural analysis, the stable area of the syncline is selected as the pinning line, and the aforementioned "four principles" are used to restore and reconstruct each fault and fold in stages.

[0027] The geophysical prediction method for halite-type polyhalite, wherein the specific process of step (1.2) is as follows: first, a three-dimensional grid model is established based on seismic waveform data, and a fixed step size is set at polarity points, i.e., peaks, troughs, and zero phase points to form a node grid; then, the step size is defined, and the surface element formed by the local trace data within the step size is regarded as the most basic object unit, and a comprehensive correlation index is obtained after calculating multiple attributes between two surface elements formed by the local trace data within the step size; finally, a value function is calculated from a global perspective, and after comparing the calculation results of the value function, the overall node connection result with the smallest value function is output to obtain a globally isochronous three-dimensional stratigraphic model data volume.

[0028] The geophysical prediction method for halite-type polyhalite, wherein the specific process of step (2.1) is: selecting actual well core sampling to carry out rock physical parameter measurement, and analyzing the rock physical parameters and seismic elastic sensitivity parameters; collecting polyhalite-containing core data in the polyhalite-containing well section of the measured well to prepare rock samples; through rock sample analysis, the test rock samples are mainly gypsum, rock salt and halite-type polyhalite; according to the needs of rock physical parameter analysis, the rock sample porosity, velocity and density parameter laboratory tests are carried out on the test rock samples, and the physical parameter characteristics and laws of potassium-rich salt mines with different contents and different stratigraphic structures are analyzed.

[0029] The geophysical prediction method for halite-type polyhalite, wherein the specific process of step (2.2) is:

[0030] (2.2.1) Establishment of rock physics model for potassium-rich formations

[0031] According to the core analysis of potassium-rich formations, the rocks of potassium-rich formations are gypsum or salt rock, and the rock physical parameter calculation model based on the Kuster-Tauxetz equation (KT equation) is shown in the following equations (3-1), (3-2) and (3-3):

[0032] ;

[0033] ;

[0034] ;

[0035] In the above formulas (3-1), (3-2) and (3-3), represents the bulk modulus; represents the shear modulus; is the bulk modulus of the background medium; is the shear modulus of the background medium; is the bulk modulus of the i-th inclusion material; is the shear modulus of the i-th inclusion material; is the equivalent bulk modulus; is the equivalent shear modulus; the coefficient and Describes the effect of adding the i-th inclusion material to the background medium m. It is a parameter related to the composition and pore shape. All inclusions must be randomly distributed to make their effect isotropic.

[0036] Comparison and analysis of measured core P-wave and S-wave velocities and calculated values show an error of no more than 5%, proving the model is reliable. The KT equation is used to predict the P-wave and S-wave velocities of gypsum and salt rock with a relative error of less than 5%.

[0037] (2.2) Analysis of the influence of polyhalite on the sound velocity of salt rock and gypsum rock

[0038] Core and logging observations show that halite-type polyhalite is associated with halite. The established rock physics model is applied to analyze the factors affecting the rock physical characteristics of potassium-rich formations through changes in mineral content.

[0039] The geophysical prediction method for rock salt-type polyhalite, wherein the specific process of step (3.1) is: combining the results of rock physical lithology testing, analyzing the parameter characteristics of polyhalite and surrounding rocks, and establishing a rock physical model of a typical polyhalite-containing formation in combination with geological analysis, and analyzing the seismic response characteristics of the salt-type polyhalite formation through seismic forward modeling.

[0040] The geophysical prediction method for halite-type polyhalite, wherein the specific process of step (3.2) is: combining the extracted and optimized various seismic attributes with information such as the stratigraphic structure, rock physical properties, and reservoir oil and gas content of known wells, clarifying the geophysical significance of the available seismic attributes, and conducting detailed interpretation and inference to draw qualitative or quantitative conclusions about the reservoir.

[0041] The geophysical prediction method for halite-type polyhalite, wherein the specific process of step (3.3) is: using the Bayesian stochastic inversion method to obtain wave impedance information, that is, firstly, based on structural interpretation and fine logging calibration, using the isochronous grid and arc length attribute constraints of global optimization tracking to carry out initial model establishment research, and establish a priori parameter model that conforms to geological laws; on this basis, using Bayesian theory, seismic data, logging data and geostatistical information are integrated into the posterior probability distribution of formation model parameters, and a seismic inversion objective function is established and solved using the MCMC method, thereby finely characterizing the reservoir distribution; and the seismic inversion objective function is as follows:

[0042] ;

[0043] Among them, in the above formula (3-4), C is the covariance matrix, d is the seismic data, G is the kernel function, r is the reflection coefficient, is the low-frequency model, K is the integral operator, is the weight of seismic information, β is the weight of prior information, and μ is the weight of the model. By adjusting the three weight coefficients, the inversion effect of the well-connected profile was tested. When the β value gradually increased, the prior constraint effect of the arc length attribute in the inversion result was obvious, and the pisoloid and cluster characteristics of polyhalite were prominent. When the μ value increased, the vertical interlayers in the inversion result increased significantly, the thin layer resolution was improved, and the consistency with the drilling data was good.

[0044] Combined with the sensitive parameters of potassium-rich rock formations, the distribution characteristics of potassium-rich rock formations were preliminarily analyzed, and the wave impedance parameter hollowing method was used to interpret and analyze the parameters of potassium-rich formations, and the spatial parameter distribution characteristics of relatively potassium-rich formations were preliminarily divided.

[0045] The geophysical prediction method for halite-type polyhalite, wherein the specific process of step (4.1) is: using the convolutional limit learning method to combine low-level features to form more abstract high-level features, and then using the convolution operator to divide the receptive field to mine the potassium salt content profile within the local range of the potassium-rich layer.

[0046] The geophysical prediction method for halite-type polyhalite, wherein the specific process of step (4.2) is: based on the complex wrinkled stratum structural interpretation of the polyhalite distribution in the target layer section of Puguang area, the distribution characteristics of polyhalite and potassium-rich brine layer in Puguang area are predicted by applying the inversion results of the sensitive factors of the potassium-rich salt layer.

[0047] By adopting the above technical solution, the present invention has the following beneficial effects:

[0048] The present invention has a reasonable concept. By establishing a potassium-rich rock formation description and inversion technology solution based on seismic, geological and well logging data, with sedimentary phase analysis of potassium-rich rock formations and a fine inversion constraint model as a priori guidance, and with phase-controlled high-precision inversion as the core, the invention can effectively achieve comprehensive analysis and distribution feature description of potassium-rich rock formations, and provide technical support for finding favorable potassium-rich formations.

[0049] The present invention can effectively realize the comprehensive analysis and distribution characteristic description of halite-type polyhalite, provide a basis for resource evaluation and geological exploration work deployment, and achieve the purpose of improving the ore-finding rate of exploration projects and saving exploration costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 This is a flow chart of the seismic inversion technology for potassium-rich strata involved in the geophysical prediction method for halite-type polyhalite of the present invention;

[0052] Figure 2 This is a technical principle diagram of the global optimization formation model involved in the geophysical prediction method for halite-type polyhalite of the present invention;

[0053] Figure 3 Schematic diagram of the global isochronous stratigraphic framework of the Xuanhan-Daxian Jia 4th and 5th segments involved in the geophysical prediction method for halite-type polyhalite of the present invention;

[0054] Figure 4 It is a density and longitudinal wave velocity cross-plot involved in the geophysical prediction method of halite-type polyhalite of the present invention;

[0055] Figure 5 Crossplots of skeleton minerals and P-wave velocity of salt rock (left) and gypsum salt (right) involved in the geophysical prediction method of halite-type polyhalite of the present invention;

[0056] Figure 6 Schematic diagram of the changes in longitudinal wave velocity and pressure of salt rock and gypsum rock involved in the geophysical prediction method of halite-type polyhalite of the present invention;

[0057] Figure 7 This is a fitting diagram of the P-wave and S-wave velocities of salt rock and gypsum rock predicted by the KT equation involved in the geophysical prediction method for halite-type polyhalite of the present invention;

[0058] Figure 8This is a characteristic diagram of the logging curve response of the polyhalite section of a typical well involved in the geophysical prediction method for rock salt type polyhalite of the present invention;

[0059] Figure 9 This is a diagram of a multi-hidden layer extreme learning machine model involved in the geophysical prediction method for halite-type polyhalite of the present invention;

[0060] Figure 10 This is a flow chart of the sensitivity factor test of potassium-rich rock formations using a volumetric limit learning machine involved in the geophysical prediction method for halite-type polyhalite of the present invention. DETAILED DESCRIPTION

[0061] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0062] The present invention will be further explained below with reference to specific embodiments.

[0063] The geophysical prediction method for halite-type polyhalite provided in this embodiment first analyzes the deformation and displacement characteristics of potassium-rich gypsum-salt rock layers under complex tectonic transformation, clarifying the deformation, displacement and accumulation characteristics of potassium-containing gypsum-salt rocks, namely "early wing accumulation and later superimposed transformation", and finding that the salt layer thickness is linearly positively correlated with the scale of halite-type polyhalite development. On this basis, research is conducted on the fine characterization technology of seismic sequences based on a global isochronous stratigraphic framework, and a global isochronous framework model is established. Combined with the analysis of rock physics and sensitive parameters of potassium-rich strata, sensitive parameters of potassium-rich strata are extracted and applied mainly through seismic inversion. Based on Bayesian wave impedance stochastic inversion, a convolutional neural network is used to directly predict potassium content curves, thereby achieving a description and prediction of the spatial distribution characteristics of potassium-rich strata, and providing technical support for the search for favorable potassium-rich strata.

[0064] According to the research needs, a potassium-rich rock description inversion technology scheme (such as phase-controlled high-precision inversion) is established based on seismic, geological and well logging data, with potassium-rich rock sedimentary facies analysis and fine inversion constraint model as prior guidance, and phase-controlled high-precision inversion as the core. Figure 1 ), to achieve comprehensive analysis and distribution characteristics description of potassium-rich rock formations.

[0065] The geophysical prediction method of halite-type polyhalite of the present invention specifically comprises the following steps:

[0066] Establishment of inversion seismic spatial constraint model for S100 and potassium-rich gypsum-salt rock layers and potassium-salt layers

[0067] S110, Deformation and displacement characteristics of potassium-rich gypsum-salt rock layers under complex structural transformation

[0068] In light of the complex geological characteristics of the area, a seismic data quality assessment was conducted on the Jiashang 4th and 5th sections. Synthetic seismic records were systematically calibrated using logging data from exploration wells within the study area. Following a "starting from the well, moving from point to line, and from line to surface" approach, the team first interpreted the main and tie lines of the single well. The 3D seismic interpretation progressed from sparse to dense, and from coarse to fine, with a gradual intensification of the interpretation of each seismic reflection layer throughout the area. The core section interpretation grid was 20×20, and the intensified interpretation grid was 4×4.

[0069] In view of the complex characteristics of faults and plastic flow of gypsum salt under multi-stage tectonic transformation, the use of equilibrium profile restoration can reveal the characteristics of structural and gypsum salt rock evolution. The principle of equilibrium profile technology is actually a material conservation principle in structural geology. According to Dahlstrom's point of view, he proposed the most basic assumptions for the compilation of classical equilibrium profiles (restoration method) (Dahlstrom, 1969): (1) The principle of constant volume (area) before and after rock deformation; this principle takes into account that most structural deformation occurs after deposition, that is, when the stratum deforms, the sediments have been completely compacted and diagenetic, so when measuring whether the profile is balanced, the volume (area) loss caused by diagenetic compaction is ignored; among them, if the selected seismic profile is perpendicular to the structural strike, the volume conservation principle (three-dimensional space) can be converted into the area conservation principle (two-dimensional space). (2) The principle that the thickness of the strata remains unchanged before and after deformation. When the rock strata are affected by tectonic stress and form folds during deformation, the folds usually undergo tectonic deformation in the form of concentric folds, and there may only be shear along the layers between different strata. That is, in this case, the thickness of the rock strata does not change before and after deformation, and the length of the rock strata after the concentric folds are unfolded is the original length of the strata. (3) The displacement of the strata along the same fault does not change. If the displacement of the fault changes, there must be a sufficient reason to explain it.

[0070] Conventional balanced profile production requires reconstruction of strata and fault models based on seismic profile interpretation, and then step-by-step back-stripping based on the principles of layer length and area conservation, which is time-consuming and not intuitive enough. The innovative use of seismic profile structural evolution restoration methods based on dynamic recovery of structural equilibrium reveals the development of faults and the flow laws of gypsum-salt rocks. This method uses the principle of balanced profile restoration, that is, the "four principles" of consistent volume, consistent layer length, consistent displacement and consistent shortening of rocks before and after deformation, and performs fracture restoration based on the principle of oblique shear stratum deformation. First, select seismic profiles that are roughly consistent with the direction of tectonic movement, that is, select seismic profiles that are perpendicular to the regional structural trend. Second, use drilling to carry out velocity fitting analysis, establish an average velocity model, carry out velocity verification, and unify it to a linear expression (that is, V = V ０ +K ｔ), converting the time domain profile to the depth domain. Third, based on the results of horizon calibration and in accordance with the principles of structural interpretation, interpretation of key horizons and faults across the entire profile was conducted to construct a reasonable structural model. Fourth, based on structural analysis, stable synclines were selected as pinning lines, and the "four principles" were used to gradually restore and reconstruct each fault and fold in stages. Conventional methods for restoring equilibrium profiles of complex fault structures are time-consuming and cannot be directly restored based on seismic profiles. By innovatively restoring complex structural deformations through seismic profiles based on dynamic equilibrium reconstruction, the faults and their associated structures were restored and reconstructed in stages, clarifying the tectonic evolution time sequence of the potassium-rich gypsum-salt rock layers in the Jialing 4th and 5th Members.

[0071] Based on regional geological data and the structural evolution analysis based on seismic profiles in the Xuanhan-Daxian area, the laws of multi-stage tectonic activities of the potash-rich salt layers in the Jiasi and Jiasi-5 members were revealed.

[0072] ① In the early Indosinian-Yanshanian period, with the breakaway and orogeny of the Jiangnan Xuefeng basement, strong compression from the SE direction formed small-scale fault flats, fault slopes and early low-amplitude folds (yellow-green faults) developed in the middle deformation layer. At this time, the fault throw of the Jiasi-5th member was small and the fold amplitude was low, and the gypsum-salt rock showed the embryonic form of flow that adjusted laterally to the wing.

[0073] During the middle and late Yanshanian periods, compression from SE to NW during the Xuefeng intracontinental orogeny intensified. Simultaneously, blocked by the Central Sichuan Block, NW-SE reverse compression and thrusting occurred (purple faults), forming NE-dipping NW faults, backthrust faults (thrust structures), counterthrust faults, and thrust triangles. The Jiashen-5th Member gypsum-salt strata at the junction of the middle and upper deformation zones in this area underwent significant changes. Rapid compression and thinning occurred at the tops of several northeast-trending anticlines, such as the well-defined Leiyinpu-Maobachang and Shuangshimiao-Fenshuiling anticlines, followed by lateral flow and accumulation on their flanks, forming extremely thick gypsum-salt strata.

[0074] ③ In the Himalayan period, under the influence of the NE-SW tectonic stress of the Daba Mountains, on the one hand, northeast-distributed structures such as Qingxichang and Xuanhan East, which were involved in the Permian, were formed in the eastern part of Xuanhan-Daxian. On the other hand, on the basis of the main northeast-trending structure, cross-superposition occurred, and local small structures (blue faults) with northwest-trending faults were formed on the upper part of the gypsum-salt rock. In this stage, the Jiasi and Jiasi-5 members also produced local distribution, that is, bottom accumulation and thickening occurred in the positive structural parts of Leikoupo and Xujiahe.

[0075] The same law was also revealed by forward modeling of the tectonic model, that is, under the superposition and transformation of multi-stage and multi-boundary tectonic actions, the potassium-bearing gypsum salt rock has the deformation, displacement and aggregation characteristics of "early wing accumulation and later superposition and transformation", and it was found that the thickness of the salt layer is positively correlated with the scale of development of halite-type polyhalite.

[0076] Construction of inversion sequence framework for S120 and potassium-rich gypsum-salt rock layers

[0077] Classic evaporite sedimentary cycles reveal that as water concentration increases and decreases, limestone, dolomite, gypsum, halite, and finally potash are deposited in sequence. Analysis of single wells in the Xuanhan-Daxian region reveals three salinization cycles within the Jia 4th and 5th Members. The second cycle reached the stage of potash concentration and precipitation, depositing potash. Dolomite and limestone brine reservoirs developed in the lower portions of the second and third cycles, making them the most favorable intervals for three-dimensional exploration of solid-liquid two-phase potassium resources. Geological inversion based on well information and seismic data is a well-established technique in the petroleum industry. For example, acoustic impedance inversion based on seismic data requires establishing a stratigraphic framework and creating an initial model using interpolation of well curves. As described in the previous section, the gypsum-salt evolution of the Jia 4th and 5th Members exhibits rapid lateral variations. Without lateral constraints using seismic sequence information, the lateral variations of halite-type polyhalite cannot be captured.

[0078] To further improve the accuracy of geophysical inversion, research has been conducted on fine-grained seismic sequence characterization techniques based on a global isochronous stratigraphic framework. This inversion framework, based on a global three-dimensional sequence stratigraphic model, is more isochronous than traditional methods. By establishing a global isochronous stratigraphic model, constraints can be applied based on this model, allowing for fine characterization of target geological bodies.

[0079] Figure 2 It is a technical process for establishing a global optimized stratum model, which is mainly divided into the following steps: first, a three-dimensional grid model is established based on the seismic waveform data, and a fixed step size is set at the polarity (peak, trough and zero phase) to form a node grid; then the step size is defined, and the patch (surface element) formed by the local trace data within the step size is regarded as the most basic object unit, and a comprehensive correlation index is obtained after calculating the multiple attributes between the two units; finally, the value function is calculated from a global perspective, and after comparison, the overall node connection result with the smallest value function is output to obtain a globally isochronous three-dimensional stratum model data body.

[0080] Figure 3 It is a global isochronous stratigraphic framework for the Jia 4th and 5th sections of the Xuanhan-Daxian section. It tracks all the events in the section with the minimum cost function as the constraint, and the automatically tracked layers have a good matching relationship with the section. Figure 3 The top and bottom interfaces of the gypsum can be tracked with high accuracy. Furthermore, even in areas with severe wrinkling and complex internal structures, the grid information can be tracked based on the relative relationships between the top and bottom interfaces and the energy of the event axes. Because the interpretation is based on a 3D data volume, the interpretation results are more isochronous and accurate. Automatically tracked horizons can be extracted from the global isochronous grid model as needed, laying the foundation for seismic inversion based on isochronous analysis.

[0081] Analysis of rock physical parameters of S200 and salt rock type polyhalite

[0082] S210, Rock physical parameter testing and regularity analysis of rock samples containing polyhalite

[0083] The rock physical parameter characteristics of halite-type polyhalite layers determine their logging parameter characteristics and seismic reflection characteristics. By analyzing the rock physical characteristics of polyhalite-bearing rock layers, it is helpful to clarify the sensitive attribute parameters of polyhalite-bearing formations contained in logging and seismic information, and provide a basis for the description and prediction of polyhalite-bearing formations.

[0084] To accurately analyze the petrophysical characteristics of polyhalite-bearing layers, rock core sampling was conducted in northeastern Sichuan to measure rock physical parameters and analyze these parameters with seismic elastic sensitivity parameters. Over 30 polyhalite-bearing cores were collected from the polyhalite-bearing sections of the wells tested for rock sample preparation. Rock sample analysis revealed that the test samples primarily consisted of gypsum, rock salt, and halite-type polyhalite. Laboratory tests were conducted on these samples for parameters such as porosity, velocity, and density, tailored to the needs of rock analysis. The velocity, density, and elastic parameters of potash-rich salt deposits with varying content and stratigraphic structures were analyzed.

[0085] The experiment tested the acoustic wave velocities of salt rock and gypsum salt. The longitudinal wave velocity of salt rock ranged from 4400m / s to 4600m / s; the longitudinal wave velocity of gypsum salt ranged from 5400m / s to 5800m / s; and the longitudinal wave velocity of salt rock containing polyhalite ranged from 4800 to 5300m / s. Rock salt has a significantly lower longitudinal wave velocity and density than gypsum rock, while polyhalite is in between, with a velocity and density greater than rock salt but less than gypsum. Figure 4 shown.

[0086] For salt and gypsum rocks, the acoustic velocity is primarily controlled by the framework minerals. Outside the framework minerals, harder inclusions (magnesite) yield higher acoustic velocities, while softer inclusions (polyhalite) yield lower acoustic velocities. The acoustic velocity of gypsum rocks is primarily determined by the acoustic velocity of the framework, with the content and type of inclusions causing the acoustic velocity to fluctuate around the framework velocity. The elastic stiffness of salt rock framework minerals (rock salt), with polyhalite being the primary inclusion, is relatively low compared to other mineral components.

[0087] According to the two structural minerals of halite and gypsum, the longitudinal wave velocity analysis of salt rock skeleton minerals and gypsum rock skeleton minerals was carried out. Under the salt rock skeleton minerals, with the increase of polyhalite content, the longitudinal wave velocity increased linearly. When the contents of halite, anhydrite and polyhalite were 78%, 16% and 2% respectively, the velocity reached 4820m / s, indicating that when the salt rock contained polyhalite, the velocity increased significantly. Under the gypsum salt skeleton minerals, the thin section polyhalite identification results accounted for 92% and 55% respectively. With the increase of the proportion of halite type polyhalite content, the gypsum rock velocity decreased to about 5200-5500m / s. Figure 5 shown.

[0088] The sound velocity of gypsum and rock salt also has great differences in pressure variation. As the pressure increases from 5MPa to 30MPa, the sound velocity of salt increases by 100-200m / s, and the sound velocity of gypsum increases by 300-500m / s. Figure 6 As shown in Figure 2, an increase in pressure reduces the space between rock fractures, leading to an increase in velocity. It is speculated that the difference in velocity changes with pressure between gypsum and rock salt may be due to the different degrees of fracture development: gypsum salt contains more fractures, while rock salt contains fewer fractures.

[0089] S220, Establishment and Analysis of Rock Physics Model of Potassium-Rich Formation

[0090] S221. Establishment of rock physics model for potassium-rich formations. Based on the core analysis of potassium-rich formations, the formations are gypsum or salt rock with low porosity, random and sparse distribution of inclusions, and the inclusions (salt rock, gypsum, polyhalite) have bearing capacity and can be considered as spherical in geometry. A rock physics parameter calculation model based on the KT equation is formed, as shown in the following equations (3-1), (3-2), and (3-3):

[0091] ;

[0092] ;

[0093] ;

[0094] In the above formulas (3-1), (3-2) and (3-3), represents the bulk modulus; represents the shear modulus; is the bulk modulus of the background medium; is the shear modulus of the background medium; is the bulk modulus of the i-th inclusion material; is the shear modulus of the i-th inclusion material; is the equivalent bulk modulus; is the equivalent shear modulus; the coefficient and Describes the effect of adding the i-th inclusion material to the background medium m. It is a parameter related to the composition and pore shape. All inclusions must be randomly distributed to make their effect isotropic.

[0095] The error between the measured longitudinal and transverse wave velocities of the core and the calculated values is less than 5%, which proves that the model is relatively reliable. The relative error of the longitudinal and transverse wave velocities of gypsum and salt rock predicted by the KT equation is within 5%. Figure 7 shown.

[0096] Core and well logging observations of S222 and polyhalite on the acoustic velocity of salt and gypsum rocks indicate that halite-type polyhalite is associated with halite. Using an established rock physics model, we analyze the factors influencing the rock physical characteristics of potassium-rich formations through changes in mineral content. The initial P-wave velocity of the salt rock is 4580 m / s. As the polyhalite content increases to 20%, the velocity increases to 4800 m / s. When the polyhalite content reaches approximately 30%, the velocity increases to 4950 m / s. This indicates a clear positive correlation between polyhalite content and P-wave velocity. Conventional anhydrite-type polyhalite exhibits a decreasing P-wave velocity with increasing volume fraction, a behavior that is distinct from conventional anhydrite-associated polyhalite. Analysis of shear waves and impedance reveals a consistent pattern with P-wave behavior: an increase in relatively high-density, high-velocity polyhalite results in an overall increase in impedance.

[0097] S300, Geophysical Spatial Characterization of Potash-Bearing Rock Layers

[0098] S310, Seismic rock physical characteristics of potassium-rich formations

[0099] S311, Seismic forward modeling and seismic response characteristics analysis of polyhalite-bearing strata

[0100] Through the systematic study of the logging curves and rock physical measurement data in the previous section, it is revealed that the Jialingjiang Formation polyhalite-bearing strata have relatively low velocity, low density, low impedance, and high gamma characteristics, but other lithologies also have low velocity and low density characteristics. Combined with the coring results, the polyhalite-bearing cores have the characteristics of high K content, high acoustic wave delay, low impedance, and low density (such as Figure 8 ), and the rock physical properties are distributed between the reference properties (between the red and purple lines). The section below the reference properties is inferred to be mudstone.

[0101] Based on the genesis of polyhalite formations, the thickness of marine strata in northeastern Sichuan was analyzed. The Leikoupo and Jialingjiang Formations are relatively thick, making them prime targets for searching for potassium-rich strata. The Jialingjiang Formation was the primary target for this search. Combined with rock physics and lithologic testing results, the polyhalite and surrounding rock parameters were analyzed. A rock physics model of a typical polyhalite-bearing formation was developed based on geological analysis. Seismic forward modeling was used to analyze the seismic response characteristics of salt-type polyhalite formations. Using a salt-bearing formation as the background, the velocity was taken to be 4580 m / s. Assuming a salt-type polyhalite content of approximately 30%, the velocity was 4740 m / s. The velocity of carbonate rocks was 6360 m / s, and the velocity of gypsum-salt rocks was 5600 m / s. Forward modeling results from a typical seismic model of salt-type polyhalite formations indicate that polyhalite occurs in the middle of thicker salt layers, with a relatively high velocity slightly lower than that of anhydrite. This is reflected seismically as a relatively blank reflection, or a slight peak at the top and a weak trough at the bottom. There is no polyhalite layer, and the seismic response characteristics of the interlayers of polyhalite and gypsum are medium-strong peak reflection.

[0102] The target strata containing polyhalite in the Puguang area are mainly concentrated in the Jia 4-5 sections, and their sedimentary characteristics show multiple sets of thick layers of gypsum salt, which are interbedded deposits of salt gypsum basin and cloud gypsum flat of typical evaporative platform; as a regional slip layer, it is affected by regional compression stress and presents unequal thickness deformation characteristics; the top and bottom of the seismic are relatively continuous medium-strong reflections, and the interior is unstable amplitude and weak continuous reflection characteristics. Combined with the seismic forward modeling knowledge, the actual data of seismic characteristics of the polyhalite-containing strata in the target area were analyzed. From the perspective of seismic calibration of typical wells, Puguang 8 Well has a 27.1-meter-thick halite-type polyhalite in the middle of Jia 4-5 sections. The middle part of Puguang 8 Well is a thick layer of low-impedance salt rock, and the overall reflection is low-frequency and relatively weak. However, the overall salt rock in Qingxi 3 Well is not developed, and the polyhalite has a cumulative thickness of only 3.1 meters, and the thickest is only 1.6 meters. It is characterized by strong peaks and strong troughs, indicating that the lithology changes rapidly and lacks the development of thick-layered salt rock and associated polyhalite.

[0103] Puguang 3D seismic data shows a relatively high dominant frequency of 40 Hz in the Jialingjiang Formation. Spectral analysis indicates that polyhalite is generally found in thicker layers of rock salt. Due to the thicker and wider troughs of these layers, the corresponding locations on the seismic spectrum often exhibit low-frequency characteristics. High-frequency characteristics often reflect thin interbeds of rock salt and surrounding rock, which is unfavorable for the occurrence of thick layers of polyhalite.

[0104] S320, Analysis of the Dominant Properties of Polyhalite-Bearing Layers

[0105] Attribute prediction analysis combines the extracted and optimized seismic attributes with information such as the stratigraphic structure, rock properties, and reservoir oil and gas content of known wells, to clarify the geophysical significance of the available seismic attributes and conduct detailed interpretation and inference to draw qualitative or quantitative conclusions about the reservoir.

[0106] Forward modeling and cross-well attribute profile analysis revealed that polyhalite-bearing layers exhibit weak amplitude and low frequency characteristics, making them sensitive to amplitude and frequency attributes. A comparative analysis of a large number of seismic frequency and amplitude attributes extracted from the Jia 4th and 5th segments revealed that the arc length attribute, which reflects the combined changes in seismic wave amplitude and frequency, is highly sensitive to polyhalite-bearing layers and can qualitatively reflect the distribution characteristics of polyhalite.

[0107] The arc length attribute is the length of the waveform curve after expansion within the time window; it is a measure of reflection anomalies and lateral changes in reflection relationships. Arc length is defined as the waveform length of the seismic trace, which is a proportional measure of the variation range of all seismic traces within the time window. Imagine drawing a seismic trace curve with a waveform style of one trace, and then imagine a rope placed on the seismic trace and following each waveform fluctuation. The arc length of the seismic trace is the total length of the rope when it is stretched. The arc length attribute is similar to the inhomogeneity of reflections, which can distinguish between strong amplitude high frequency and strong amplitude low frequency, and weak amplitude high frequency and low frequency reflections. The calculation formula of the arc length attribute is as follows:

[0108] ;

[0109] Where a(i) is the amplitude at the i-th sampling point, T is the sampling period, which is the inverse of the frequency, and N is the number of sampling points in the time window. The low-frequency, weak-amplitude characteristics of polyhalite-bearing layers manifest as low-value anomalies in the arc length attribute.

[0110] Based on stacked arc length attribute maps from linked seismic sections from Wells Dawan-102, Puguang-11, and Puguang-5, the previous analysis shows that polyhalite primarily develops in large sets of salt rocks. These highly plastic salt rocks undergo local thickening and thinning due to subsequent tectonic compression. Therefore, identifying polyhalite-bearing layers requires first locating the large sets of salt rocks. Linked seismic sections reveal that the seismic response characteristics are relatively stable in the thinning areas of the Jia45 Member, with less salt rock and no polyhalite-bearing layers. In the thickening areas of the Jia45 Member, the seismic response characteristics are more complex, with moderately weak peaks developing in the middle. Potassium curves indicate that the low-value anomalies in the arc length attribute well characterize the location of polyhalite development and are in good agreement with drilling data.

[0111] Based on the stacked arc length attribute seismic profile of the original seismic data, it can be seen that the middle and lower parts of the thickened area of the Jiajing 4th and 5th members have obvious moderate-weak seismic reflection and low arc length characteristics, and the lateral distribution boundary is clear.

[0112] Based on the seismic response characteristics and arc length attributes of the polyhalite-containing layer, the arc length attribute seismic profile of the original seismic data was analyzed. According to the drilling calibration results, the arc length attribute threshold value was set to 4. The positions with a value less than 4 were considered to be polyhalite-containing salt rock layers, and the distribution map of the polyhalite-containing salt layer in the Jia 4th and 5th members was extracted.

[0113] In view of the actual characteristics of the marine potassium salt strata in northeastern Sichuan, the distribution characteristics of the polyhalite salt layers were preliminarily determined using the three-dimensional processed seismic data reprocessed in this project based on rock physical testing, seismic attribute analysis, well-seismic calibration, forward modeling and other methods.

[0114] S330, Spatial Characterization of Salt Strata - Bayesian Wave Impedance Inversion and Potassium-Rich Strata Analysis

[0115] Wave impedance inversion is one of the most commonly used and effective methods for quantitative reservoir prediction. It combines seismic data, well logging data, and geological interpretation, and fully utilizes the high vertical resolution of well logging data and the good lateral continuity of seismic data to "transform" the seismic data. However, due to the band-limited nature of seismic data, traditional deterministic inversion methods inevitably suffer from the defect of low resolution. Bayesian stochastic inversion describes the solution to the inverse problem as a posterior probability density, thereby clearly illustrating the relationship between the random inversion results. Compared with the single solution of deterministic inversion, the Bayesian probabilistic stochastic inversion solution is a posterior probability distribution that is difficult to express analytically, overcoming the problem that commonly used linearization methods affect the accuracy of the solution.

[0116] The Bayesian stochastic inversion method is used to obtain wave impedance information. First, based on structural interpretation and fine calibration of well logging, the isochronous grid and arc length attribute constraints of global optimization tracking are used to carry out initial model establishment research, and a priori parameter model that conforms to geological laws is established to effectively overcome the difficulties such as the rapid spatial variation of potassium-rich rock layers and the difficulty in establishing a fine initial model. On this basis, through Bayesian theory, seismic data, well logging data and geostatistical information are integrated into the posterior probability distribution of formation model parameters, and an innovative seismic inversion space-varying objective function is established. The MCMC method is used to solve it, and then the reservoir distribution is finely portrayed, improving the description and prediction ability of the inversion results for the reservoir. The inversion objective function is constructed as follows:

[0117] ;

[0118] Among them, in the above formula (3-4), C is the covariance matrix, d is the seismic data, G is the kernel function, and r is the reflection coefficient. is the low-frequency model, K is the integral operator, is the weight of the seismic information, β is the weight of the prior information, and μ is the weight of the model. By adjusting the three weight coefficients, the inversion performance of the well-connected profile was tested. As the β value gradually increased, the prior constraint on the arc length attribute in the inversion results became more effective, and the pea-shaped and clumping features of the polyhalite became more prominent. As the μ value increased, the inversion results showed a significant increase in vertical interlayers, improved thin-bed resolution, and good agreement with drilling data. Overall, the newly constructed objective function can effectively solve the problem of predicting the rapid spatial variation of potassium-rich strata containing polyhalite.

[0119] The inversion section shows clear features of the relatively stable sedimentary structures of the Leikoupo Formation and the wrinkled strata of the Jialingjiang Formation. The advantages of using the Bayesian stochastic inversion method to obtain wave impedance information are: first, it fully utilizes the information of the dominant frequency band of the seismic data; second, it reduces the inversion's dependence on the initial model.

[0120] Based on the sensitive parameters of potassium-rich rock formations, a preliminary analysis of their distribution characteristics was conducted. Wave impedance parameter hollowing was used to interpret and analyze the parameters of the potassium-rich formations, and a preliminary delineation of the spatial parameter distribution characteristics of relatively potassium-rich formations was achieved. The well logging curve is a K curve, indicating that potassium-rich layers have relatively low impedance characteristics, with the lowest impedance being found in salt rock formations.

[0121] S400, rock salt type polyhalite prediction

[0122] Statistical analysis of well logging potassium content and impedance curves for each layer reveals that potassium-rich formations generally exhibit a negative correlation, with higher potassium content and lower impedance. However, due to the influence of salt rock in the Jialingjiang Formation, the linear correlation between potassium content and impedance is limited, making it impossible to directly establish a relationship between impedance and potassium content. The convolutional limit learning function combines low-level features to form more abstract high-level features. The convolution operator is used to divide the receptive field, extracting useful information within the localized range of potassium-rich formations, and establish a correlation between the potassium content target curve, input impedance, and various seismic attributes, thereby improving prediction accuracy.

[0123] S410, Convolutional Limit Learning Machine Inversion of Sensitivity Factors of Potassium-Rich Rock Formations

[0124] Extreme Learning Machine is a single hidden layer feedforward neural network model for classification and regression. Unlike feedforward neural networks based on gradient descent, this model uses a fast, new learning mechanism that can randomly assign input weights and thresholds, and obtain the output layer connection weights through the least squares method. Compared with the BP algorithm, the ELM algorithm only needs to set the number of hidden layer nodes. During the algorithm implementation, there is no need to adjust the network input weights and hidden layer thresholds. It has the advantages of fast learning speed and good generalization performance. Because of its simple theory and easy implementation, the extreme learning machine has become a popular machine learning technology. Its structure is as follows: Figure 9 shown.

[0125] The extreme learning machine combines low-level features to form more abstract high-level features; the convolution operator is used to divide the receptive field, explore useful information within the local range of the potassium-rich layer, resolve the uncertainty of point-to-point calculations, and thus directly predict the potassium salt content. Figure 10 This is the flow chart of the active limit learning machine test of potassium-rich rock sensitivity factors.

[0126] Combined with the potash content profile obtained by the volume limit learning machine under the well line and the seismic inversion wave impedance results, and by comparing the potash content profile obtained by volume limit learning with the seismic inversion wave impedance profile and the multivariate regression profile, it can be concluded that the volume limit learning machine method can directly predict the potash content, and has a better consistency with the well, and can better show the distribution range of potash.

[0127] Analysis on the characteristics of sensitive factors of polyhalite and potassium-rich brine formations in S420 and Puguang areas

[0128] The Puguang area is generally located in the evaporative salt basin, where polyhalite is widely distributed. Based on the complex wrinkled stratigraphic structural interpretation of the polyhalite distribution in the target layer of the area, the results of the inversion of the sensitive factors of the potassium-rich salt layer were applied to carry out a prediction study on the distribution characteristics of polyhalite and potassium-rich brine layers in the area.

[0129] Combining the predictions of potash sensitivity factor profiles from Puguang 8 and Puguang 10 using a roll-up finite learning machine reveals good agreement between the predicted profiles and the wells. A thick layer of halite-type polyhalite develops in the central portion of Puguang 8, measuring 27.1 meters thick compared to the predicted 29.8 meters, achieving a 90% agreement. The thickness of the layer in Puguang 10 is 2.9 meters, while conventional inversion predictions have a thickness limit of 20 meters, making it impossible to effectively predict. However, this method also predicted the potash layer. Furthermore, the distribution of the potash sensitivity factor exhibits banded, dotted, and irregular clusters, consistent with previous geological understanding.

[0130] Combined with the regional inversion well cross-section, it can be seen that when the thickness of the potash layer varies greatly, the convolutional limit learning machine method is consistent with the well and can well present the planar distribution within the work area.

[0131] The sensitive factors of polyhalite-containing strata were used to predict the planar distribution characteristics of the polyhalite-containing strata in the Jialing River 4 and 5 sections. It was found that due to the strong wrinkling characteristics of the strata in the Jialing River 4-5 sections, the strata thickness varies greatly in the lateral direction. The parameter thickness analysis results of the potassium-rich strata are affected by both the stratum thickness and the enrichment degree of the potassium-rich layer. The cumulative thickness of the potassium-rich layer is greatly affected by the stratum thickness, which further reveals the distribution characteristics of the deformation and displacement of the potash layer under tectonic action.

[0132] The invention has a reasonable concept, can effectively describe and predict the spatial distribution characteristics of potassium-rich strata, and provide technical support for finding favorable potassium-rich strata.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A geophysical prediction method for halite-type polyhalite, characterized by: First, we analyzed the deformation and displacement characteristics of potassium-rich gypsum-salt rock layers under complex tectonic transformation, clarifying the deformation, displacement, and aggregation characteristics of potassium-containing gypsum-salt rocks. We found that the thickness of the salt layer is linearly positively correlated with the scale of halite-type polyhalite development. On this basis, we conducted research on the fine characterization of seismic sequences based on a global isochronous stratigraphic framework and established a global isochronous framework model. We then combined the rock physics and sensitive parameter analysis of potassium-rich strata, extracted and applied the sensitive parameters of potassium-rich strata mainly through seismic inversion, and used convolutional neural networks based on Bayesian wave impedance stochastic inversion to directly predict potassium content curves, thus achieving a description and prediction of the spatial distribution characteristics of potassium-rich strata. The prediction method mainly includes the following steps: (1) Establishment of a spatial constraint model for seismic inversion of potassium-rich gypsum-salt rock layers and potassium-salt layers (1.1) Analysis of deformation and displacement characteristics of potassium-rich gypsum-salt rock formations under complex tectonic transformation. The specific process involves: utilizing the principle of equilibrium profile restoration, i.e., the "four principles" of consistent volume, consistent layer length, consistent displacement, and consistent shortening before and after deformation, and conducting fracture restoration based on the principle of oblique shear formation deformation; (1.2) Construction of inversion sequence framework for potassium-rich gypsum-salt rock layers; (2) Analysis of rock physical parameters of halite-type polyhalite (2.1) Testing and analysis of rock physical parameters of rock samples containing polyhalite layers; (2.2) Establishment and analysis of rock physics models for potassium-rich formations; (3) Geophysical spatial characterization of potash-bearing rock layers (3.1) Analysis of seismic rock physical characteristics of potassium-rich formations; (3.2) Analysis of the dominant properties of polyhalite-containing layers; (3.3) Spatial characterization of salt-bearing strata through Bayesian wave impedance inversion and potassium-rich strata analysis; (4) Prediction of halite-type polyhalite (4.1) Inversion of sensitivity factors of potassium-rich rock formations using a volumetric finite element learning machine; (4.2) Analysis of sensitive factors of polyhalite and potassium-rich brine formations.

2. The geophysical prediction method for halite-type polyhalite according to claim 1, wherein: The specific process of carrying out fracture recovery is as follows: (1.1.1) Select seismic sections that are roughly consistent with the direction of tectonic movement, that is, select seismic sections that are perpendicular to the regional structural trend; (1.1.2) Use drilling data to perform velocity fitting analysis, establish an average velocity model, perform velocity verification, unify the velocity expression to a linear expression, and convert the time domain profile to the depth domain; (1.1.3) Based on the results of horizon calibration and in accordance with the principles of structural interpretation, interpretation of key horizons and faults across the entire profile is conducted to construct a reasonable structural model; (1.1.4) Based on the structural analysis, the stable area of the syncline is selected as the pinning line, and the aforementioned "four principles" are used to restore and reconstruct each fault and fold in stages.

3. The geophysical prediction method for halite-type polyhalite according to claim 1, wherein: The specific process of step (1.2) is as follows: first, a three-dimensional grid model is established based on the seismic waveform data, and a fixed step size is set at the polarity, i.e., the peak, trough and zero phase to form a node grid; then the step size is defined, and the surface element formed by the local trace data within the step size is regarded as the most basic object unit, and a comprehensive correlation index is obtained after calculating the various attributes between the two surface elements formed by the local trace data within the step size; finally, a value function is calculated from a global perspective, and after comparing the calculation results of the value function, the overall node connection result with the smallest value function is output to obtain a globally isochronous three-dimensional stratigraphic model data body.

4. The geophysical prediction method for halite-type polyhalite according to claim 1, wherein: The specific process of step (2.1) is as follows: selecting actual well core sampling to carry out rock physical parameter measurement, and analyzing the rock physical parameters and seismic elastic sensitivity parameters; collecting polyhalite-containing core data in the polyhalite-containing well section of the measured well to prepare rock samples; through rock sample analysis, the test rock samples are mainly gypsum, rock salt and rock salt type polyhalite. According to the needs of rock physical parameter analysis, laboratory tests of rock sample porosity, velocity and density parameters are carried out on the test rock samples, and the physical parameter characteristics and laws of potassium-rich salt mines with different contents and different stratigraphic structures are analyzed.

5. The geophysical prediction method for halite-type polyhalite according to claim 1, wherein: The specific process of step (2.2) is as follows: (2.2.1) Establishment of rock physics model for potassium-rich formations According to the core analysis of potassium-rich formations, the rocks of potassium-rich formations are gypsum or salt rock, and the rock physical parameter calculation model based on the Kuster-Tauxetz equation (KT equation) is shown in the following equations (3-1), (3-2) and (3-3): ； ； ； In the above formulas (3-1), (3-2) and (3-3), represents the bulk modulus; represents the shear modulus; is the bulk modulus of the background medium; is the shear modulus of the background medium; is the bulk modulus of the i-th inclusion material; is the shear modulus of the i-th inclusion material; is the equivalent bulk modulus; is the equivalent shear modulus; the coefficient and Describes the effect of adding the i-th inclusion material to the background medium m. It is a parameter related to the composition and pore shape. All inclusions must be randomly distributed to make their effect isotropic. Comparison and analysis of measured core P-wave and S-wave velocities and calculated values show an error of no more than 5%, proving the model is reliable. The KT equation is used to predict the P-wave and S-wave velocities of gypsum and salt rock with a relative error of less than 5%. (2.2.2) Analysis of the influence of polyhalite on the sound velocity of salt rock and gypsum rock Core and logging observations show that halite-type polyhalite is associated with halite. The established rock physics model is applied to analyze the factors affecting the rock physical characteristics of potassium-rich formations through changes in mineral content.

6. The geophysical prediction method for halite-type polyhalite according to claim 1, characterized in that: The specific process of step (3.1) is as follows: combining the results of rock physical and lithological tests, analyzing the parameter characteristics of polyhalite and surrounding rocks, establishing a rock physical model of a typical polyhalite-containing formation in combination with geological analysis, and analyzing the seismic response characteristics of the halite-type polyhalite formation through seismic forward modeling.

7. The geophysical prediction method for halite-type polyhalite according to claim 1, wherein: The specific process of step (3.2) is: combining the extracted and optimized seismic attributes with the information of known wells, clarifying the geophysical significance of the available seismic attributes, and conducting detailed interpretation and inference to draw qualitative or quantitative conclusions about the reservoir.

8. The geophysical prediction method for halite-type polyhalite according to claim 1, wherein: The specific process of step (3.3) is as follows: using the Bayesian stochastic inversion method to obtain wave impedance information, that is, firstly, based on structural interpretation and fine logging calibration, using the isochronous grid of global optimization tracking and arc length attribute constraints to carry out initial model establishment research, and establish a priori parameter model that conforms to geological laws; on this basis, using Bayesian theory, seismic data, logging data and geostatistical information are integrated into the posterior probability distribution of formation model parameters, and a seismic inversion objective function is established and solved using the MCMC method, thereby finely characterizing the reservoir distribution; and the seismic inversion objective function is as follows: ； Among them, in the above formula (3-4), C is the covariance matrix, d is the seismic data, G is the kernel function, r is the reflection coefficient, is the low-frequency model, K is the integral operator, is the weight of earthquake information, is the weight of the prior information, and μ is the weight of the model. By adjusting the three weight coefficients, the inversion effect of the well-connected profile was tested. When the β value gradually increased, the prior constraint effect of the arc length attribute in the inversion result was obvious, and the pisoloid and lumpy features of the polyhalite were prominent. When the μ value increased, the vertical interlayers in the inversion result increased significantly, the thin layer resolution was improved, and the agreement with the drilling data was good. Combined with the sensitive parameters of potassium-rich rock formations, the distribution characteristics of potassium-rich rock formations were preliminarily analyzed, and the wave impedance parameter hollowing method was used to interpret and analyze the parameters of potassium-rich formations, and the spatial parameter distribution characteristics of relatively potassium-rich formations were preliminarily divided.

9. The geophysical prediction method for halite-type polyhalite according to claim 1, wherein: The specific process of step (4.1) is as follows: using the convolutional limit learning machine method to combine low-level features to form more abstract high-level features, and then using the convolution operator to divide the receptive field, mining useful information within the local range of the potassium-rich layer, solving the uncertainty of point-to-point calculation, and thus directly predicting the potassium salt content.

10. The geophysical prediction method for halite-type polyhalite according to claim 1, characterized in that: The specific process of the step (4.2) is: based on the interpretation of the complex wrinkled stratum structure for the distribution of polyhalite in the target layer, the distribution characteristics of polyhalite and potassium-rich brine formations are predicted by applying the inversion results of the sensitivity factors of the potassium-rich salt layer.

Images