Nuclear magnetic resonance-based deepwater turbidite physical simulation sedimentary bottom type three-dimensional fine characterization method and system

Abstract

This invention belongs to the field of nuclear magnetic resonance (NMR) technology and discloses a three-dimensional fine characterization method for deep-water turbidite sedimentary bottom morphology based on NMR physical simulation. This invention solves the problem that traditional methods require destructive sample cross-sectional observation, failing to obtain a complete and continuous three-dimensional internal structure non-destructively, leading to severe loss of spatial information. It also addresses the issue that conventional surface scanning techniques cannot penetrate sediments, only obtaining top topography and being powerless to obtain crucial information such as internal sequence and physical property distribution. Furthermore, it overcomes the problem that existing imaging technologies are mostly used for macroscopic morphology recognition, making it difficult to perform non-destructive, quantitative, and three-dimensional spatial mapping of key rock physical parameters such as porosity and permeability. Finally, it addresses the problem that traditional methods cannot capture intermediate states of multi-stage sedimentary evolution, causing a spatiotemporal disconnect between sedimentary dynamics and the final configuration.

Description

Technical Field

[0001] This invention belongs to, but is not limited to, the field of nuclear magnetic resonance technology, and particularly relates to a method and system for three-dimensional fine characterization of sedimentary bottom types based on nuclear magnetic resonance physical simulation of deep-water turbidite deposits. Background Technology

[0002] Deep-water turbidites are key target reservoirs for global deep-water oil and gas exploration, and their internal sedimentary configuration and physical property distribution have a decisive influence on oil and gas enrichment and development. Physical simulation flume experiments are an important means to study their sedimentary processes and establish geological models. However, in simulating high-concentration turbidity current deposition, the experimental system is often in a highly turbid state, making it difficult to directly observe the evolution of the internal structure of the sedimentary body during the experiment.

[0003] Currently, most related studies employ endpoint observation methods, which involve obtaining the final sedimentary configuration through destructive methods such as drainage and slicing after the experiment. While this method can obtain the final geometry of the sedimentary body, it completely fails to capture information on the dynamic evolution during the sedimentation process, resulting in a lack of continuous and precise spatiotemporal correlation between the sedimentary dynamics and the final configuration. More importantly, even in the analysis of the final sedimentary body, traditional methods are mainly limited to macroscopic morphological descriptions, lacking a systematic and precise three-dimensional characterization capability for key physical properties such as internal pore structure characteristics, fluid occurrence state, and permeability heterogeneity.

[0004] Nuclear magnetic resonance (NMR) technology, due to its sensitivity to pore fluids and its ability to non-destructively acquire information such as porosity, pore size distribution, fluid saturation, and permeability from core-scale samples, has been widely used in reservoir rock physics experiments. However, currently, this technology is mostly used for static analysis of actual subsurface cores, and its integration with physical simulation sedimentary experiments for systematically reconstructing the three-dimensional physical properties of experimental sedimentary bodies is lacking. This deficiency creates a quantitative bridge based on a unified physical property scale between physical simulation experiments and subsurface reservoir prediction.

[0005] Therefore, there is an urgent need to develop a method for fine characterization of the internal structure of experimental sedimentary bodies based on non-destructive testing technologies such as nuclear magnetic resonance, so as to further reveal the three-dimensional spatial distribution law of its internal physical properties on the basis of obtaining the macroscopic configuration of the sedimentary body, thereby revealing the formation mechanism of deep-water turbidite reservoirs more completely and providing high-precision physical property parameter constraints for geological modeling.

[0006] In terms of precise quantitative characterization of underwater sediments, existing technologies often struggle to accurately measure and analyze the three-dimensional physical properties of sediments from physical simulation experiments. Current mainstream methods, such as laser particle size analysis and digital image analysis, can obtain the grain size distribution or surface features of sediments under specific conditions. However, these methods primarily target suspended or surface particles, and their physical principles prevent them from effectively acquiring key physical properties such as pore structure, fluid distribution, and permeability within the sediment body, let alone achieving precise measurements in three-dimensional space.

[0007] On the other hand, conventional core property analysis methods, such as helium porosimetry, liquid nitrogen adsorption, and high-pressure mercury intrusion, can provide accurate parameters such as porosity and pore size distribution. However, these methods are destructive measurements, which not only destroy the original structure and sedimentological information of the sample, but also fail to reconstruct the true spatial location of the target structural unit in the original sedimentary body. The data obtained are discrete and isolated.

[0008] To obtain information about the interior of sedimentary bodies, techniques such as X-ray computed tomography (CT) have been introduced into relevant research. However, existing studies mostly focus on overall scanning imaging of the final sedimentary body after the experiment, and the analysis results are mainly macroscopic density difference imaging, used to distinguish sedimentary facies and identify large-scale configurations. These techniques cannot directly and non-destructively obtain key, quantifiable rock physical property parameters such as the three-dimensional distribution of porosity, the three-dimensional model of pore structure, and the three-dimensional model of fluid saturation inside the sedimentary body. These parameters are the key bridge connecting macroscopic configuration and reservoir permeability.

[0009] Existing physical simulation experimental characterization methods suffer from a significant technical gap between obtaining the final morphological configuration of sedimentary bodies and providing the refined, three-dimensional petrophysical parameters required for supporting reservoir models and numerical simulations. This greatly limits the quantitative understanding of the formation and distribution of deep-water turbidite reservoirs. Therefore, there is an urgent need for a characterization method and technology that can non-destructively, precisely, and quantitatively reconstruct the three-dimensional physical properties of experimental sedimentary bodies.

[0010] Based on the above analysis, the urgent technical problems that need to be solved in the existing technology are:

[0011] (1) Traditional water tank experiments cannot achieve non-destructive detection of the internal structure of sediments. Slice dissection not only destroys the original bottom shape, but also can only obtain discrete two-dimensional profile information.

[0012] (2) A single three-dimensional surface scanner cannot penetrate water and sediment to obtain the internal sedimentary sequence and grain size variation. Summary of the Invention

[0013] To address the problems existing in the prior art, this invention provides a three-dimensional fine characterization method for the sedimentary bottom type of deep-water turbidite based on nuclear magnetic resonance physical simulation.

[0014] This invention is implemented as follows: A three-dimensional fine characterization method for deep-water turbidite sedimentary bottom types based on nuclear magnetic resonance physical simulation includes:

[0015] Step 1: Construct a non-magnetic experimental water tank and a nuclear magnetic resonance scanning system, so that the core observation section of the non-magnetic experimental water tank is within the combined influence of the static magnetic field, the triaxial gradient magnetic field, the radio frequency excitation field, and the signal receiving area.

[0016] Step 2: Inject background fluid into the non-magnetic experimental water tank, perform background calibration scanning on the background fluid to obtain the NMR signal reference under pure water background, and set the turbidity release parameters, turbidity trigger parameters and NMR scan timing parameters.

[0017] Step 3: Release turbidity mixture upstream of the non-magnetic experimental water tank through the material supply system, monitor the turbidity changes of the water body using the turbidity monitoring system, record the arrival time of turbidity when the turbidity exceeds the frontal detection threshold, and record the sedimentation stabilization time after the turbidity is below the background steady-state threshold and remains stable for a preset time.

[0018] Step 4: At the moment of sedimentation stabilization, in-situ nuclear magnetic resonance scanning is initiated. Hydrogen nuclei in the sedimentation system are excited by radio frequency pulses, and the observation area is spatially encoded in three dimensions by a triaxial gradient magnetic field. Spin echo signals are collected by a signal receiving coil to obtain the raw nuclear magnetic signal dataset at the corresponding moment.

[0019] Step 5: Using the original NMR signal dataset as input, perform three-dimensional voxel reconstruction to obtain the three-dimensional voxel matrix at the corresponding time point;

[0020] Step 6: Using the echo train of each voxel in the three-dimensional voxel matrix as input, and according to the relaxation relationship of the echo amplitude decaying with the acquisition time, output the lateral relaxation time distribution under the constraint of non-negative amplitude of the relaxation components.

[0021] Step 7: Using the transverse relaxation time distribution, the NMR signal benchmark under pure water background, and the overall porosity scale as inputs, output the porosity voxel field; using the logarithmic mean relaxation time corresponding to the transverse relaxation time distribution as input, output the particle size parameter voxel field according to the empirical mapping relationship between the logarithmic mean relaxation time and the median particle size of the sediment; based on the transverse relaxation time distribution, divide the free fluid index and the bound fluid index, and combine the porosity voxel field to output the permeability voxel field according to the empirical relationship of permeability.

[0022] Step 8: Perform multiple cyclic acquisitions in the order of turbidity release, turbidity judgment, in-situ scanning and three-dimensional voxel reconstruction. Use the three-dimensional voxel matrix of adjacent phases as input to identify voxel differences, output single-phase sedimentary sequences, and spatially superimpose each single-phase sedimentary sequence in chronological order to obtain a four-dimensional configuration model of the deep-water turbidite sedimentary bottom.

[0023] Step 9: Using the four-dimensional configuration model, porosity voxel field, particle size parameter voxel field, and permeability voxel field as inputs, output three-dimensional visualization results, three-dimensional geomorphological results of sedimentary bottom type, and quantitative evaluation results of microphysical properties.

[0024] Furthermore, the construction of the non-magnetic experimental water tank and the nuclear magnetic resonance scanning system includes:

[0025] A non-magnetic deposition simulation water tank with a length of 2m, a width of 1m, and a height of 0.5m was constructed. The non-magnetic deposition simulation water tank was formed by splicing non-magnetic glass composite material with a thickness of 3cm, and the joints were sealed with modified epoxy resin.

[0026] Two planar superconducting magnet plates are arranged in parallel on the upper and lower sides of the core observation section of the non-magnetic sedimentation simulation tank to create a static magnetic field environment with a central magnetic field strength of 0.5T and a magnetic field uniformity better than 50ppm in the core observation section.

[0027] A triaxial gradient magnetic field coil is arranged outside the non-magnetic sedimentation simulation tank. The triaxial gradient magnetic field coils correspond to the X, Y and Z directions respectively, and the maximum linear gradient field strength is 20 mT per meter.

[0028] A radio frequency pulse transmitter was configured in the observation section of the non-magnetic sedimentation simulation tank to match the radio frequency center frequency with the Larmor resonance frequency of the hydrogen nucleus under a static magnetic field of 0.5T.

[0029] A signal receiving coil is placed close to the outer wall of a non-magnetic deposition simulation tank so that the signal receiving coil receives the spin echo signal generated by the transverse relaxation of hydrogen nuclei during the interval of radio frequency pulses.

[0030] A 200L slurry pump was connected to the upstream of the non-magnetic deposition simulation tank, and a turbid mixture was delivered to the tank through a wear-resistant hose.

[0031] Furthermore, the background calibration scan includes:

[0032] Low-mineralized water to a depth of 400 mm was injected into the non-magnetic experimental water tank as background fluid.

[0033] The background fluid is pre-scanned, and the free induction attenuation signal is acquired to determine the longitudinal relaxation time and transverse relaxation time of the background fluid.

[0034] The NMR signal amplitude of the background fluid was calibrated as a pure water signal reference.

[0035] The pure water signal reference is input into the porosity inversion stage to make the initial amplitude of each voxel echo proportional to the pure water signal reference, and outputs the porosity value of the corresponding spatial location under the overall porosity scale constraint.

[0036] Furthermore, the release and triggering control of the turbid mixture includes:

[0037] Add 100-mesh quartz sand, 125-mesh calcite and 500-mesh kaolin to the mixing chamber and prepare the solid phase composition according to the following mass fractions: quartz sand 15%, calcite 30% and kaolin 55%.

[0038] A turbid mixture with a volume concentration of 25% and an allowable deviation of 2% was prepared by stirring continuously at a stirring speed of 300 rpm for 20 min.

[0039] A turbid mixture with a single volume of 50L is released upstream of the non-magnetic experimental water tank at a flow rate of 5L per second using a slurry pump.

[0040] Turbidity changes were monitored using a turbidimeter with a range of 0 to 4000 NTU. The arrival time of turbidity flow was recorded when the turbidity rose by more than 50 NTU, and the deposition stabilization time was recorded after the turbidity dropped to less than 60 NTU and remained stable for more than 5 minutes. The deposition stabilization time was used as the trigger time for in-situ nuclear magnetic resonance scanning.

[0041] Furthermore, the in-situ nuclear magnetic resonance scanning includes:

[0042] A layer-selective gradient is applied along the Z-direction when a 90-degree radio frequency pulse is emitted to define the target imaging layer.

[0043] A phase encoding gradient is applied in the Y direction at the leading edge of spin echo generation, so that hydrogen nuclei with different Y coordinates can obtain different phase information;

[0044] During the spin echo acquisition window, a frequency coding gradient is applied along the X direction so that hydrogen nuclei with different X coordinates obtain different frequency information.

[0045] Using the static magnetic field strength, the gradient magnetic field strength in the X direction, the gradient magnetic field strength in the Y direction, the gradient magnetic field strength in the Z direction, and the spatial coordinates as inputs, according to the relationship between the hydrogen nucleus Larmor resonance frequency and the magnetic field strength and spatial position, a spin echo signal carrying three-dimensional spatial position information is output, and this spin echo signal is input into the three-dimensional voxel reconstruction stage.

[0046] Furthermore, the three-dimensional voxel reconstruction, multi-scale property inversion, and four-dimensional configuration analysis include:

[0047] The raw NMR signal datasets from each period are transformed from the three-dimensional frequency domain to the spatial domain, and the corresponding three-dimensional voxel matrix is ​​output. Each voxel corresponds to a physical space with a side length of 1.95 mm.

[0048] Using the echo trains within each voxel as input, the transverse relaxation time distribution is output under the non-negative constraint of the relaxation spectrum amplitude by employing singular value decomposition combined with non-negative least squares solution.

[0049] The initial amplitude of the echo is obtained by using the signal integral of the transverse relaxation time distribution as input, and the porosity voxel field is output by combining the pure water signal reference and the overall porosity scale.

[0050] The logarithmic mean relaxation time is calculated using the transverse relaxation time distribution as input, and the particle size parameter voxel field is output according to the empirical proportionality coefficient and empirical exponential relationship between the logarithmic mean relaxation time and the median particle size.

[0051] Using the porosity voxel field, the free fluid index and the bound fluid index obtained by dividing the transverse relaxation time cutoff value of 33ms as inputs, the permeability voxel field is output according to the Timur-Coates empirical relationship.

[0052] Boolean difference operations are performed on the three-dimensional voxel matrices of adjacent phases as input to output the new single-phase depositional sequence within the corresponding time interval. The single-phase depositional sequences are then spatially aligned and superimposed in chronological order to output a four-dimensional configuration model.

[0053] Another objective of this invention is to provide a three-dimensional fine characterization system for deep-water turbidite physical simulation sedimentary bottom based on nuclear magnetic resonance, comprising a non-magnetic sedimentation simulation tank, a static magnetic field generator, a triaxial gradient magnetic field coil, a radio frequency pulse transmitter, a signal receiving coil, a material supply system, a turbidity monitoring system, a data acquisition system, and a computer;

[0054] The non-magnetic sedimentation simulation tank is used to contain background fluid and turbidity mixture, and provides an observation space for turbidity movement, settling and deposition;

[0055] The static magnetic field generating device is set on the upper and lower sides of the core observation section of the non-magnetic deposition simulation tank, and is used to form a static magnetic field in the core observation section to cause hydrogen nuclei in the observation area to undergo orientation polarization.

[0056] The triaxial gradient magnetic field coils are arranged around the non-magnetic deposition simulation tank to generate spatially encoded gradients in the X, Y, and Z directions based on the static magnetic field.

[0057] The radio frequency pulse transmitter is used to transmit radio frequency pulses that match the Larmor resonance frequency of the hydrogen nucleus to the core observation segment;

[0058] The signal receiving coil is located near the outer wall of the non-magnetic deposition simulation tank and is used to receive the spin echo signal generated by the transverse relaxation of hydrogen nuclei during the interval of radio frequency pulses.

[0059] The material supply system is connected upstream of the non-magnetic deposition simulation tank and is used to release turbidity mixture into the non-magnetic deposition simulation tank.

[0060] The turbidity monitoring system is used to monitor the turbidity of the water in the non-magnetic sedimentation simulation tank and output turbidity arrival signal and sedimentation stability signal to the data acquisition system.

[0061] The data acquisition system is used to trigger the coordinated operation of the radio frequency pulse transmitter, the triaxial gradient magnetic field coil, and the signal receiving coil based on the deposition stability signal, and to acquire the raw NMR signal dataset.

[0062] The computer is used to take the raw NMR signal dataset as input and output three-dimensional voxel matrix, porosity voxel field, particle size parameter voxel field, permeability voxel field, four-dimensional configuration model, three-dimensional visualization results, sedimentary bottom three-dimensional geomorphological results, and quantitative evaluation results of microphysical properties.

[0063] Furthermore, the non-magnetic deposition simulation tank is a tank with a length of 2m, a width of 1m, and a height of 0.5m. The tank is formed by splicing non-magnetic glass composite material with a thickness of 3cm, and the joints are sealed with modified epoxy resin. The material supply system includes a slurry pump with a volume of 200L and a wear-resistant hose. The slurry pump is connected to the upstream of the non-magnetic deposition simulation tank through the wear-resistant hose.

[0064] Furthermore, the static magnetic field generating device includes two planar superconducting magnet plates, which are respectively arranged in parallel on the upper and lower sides of the core observation section of the non-magnetic deposition simulation tank. The static magnetic field generating device forms a static magnetic field environment with a central magnetic field strength of 0.5T and a magnetic field uniformity better than 50ppm in the core observation section; the maximum linear gradient field strength of the triaxial gradient magnetic field coil is 20mT per meter.

[0065] Furthermore, the computer includes a three-dimensional voxel reconstruction unit, a relaxation spectrum inversion unit, a physical property parameter inversion unit, a sequence analysis unit, and a result output unit;

[0066] The three-dimensional voxel reconstruction unit is used to take the original nuclear magnetic resonance signal dataset as input, and output a three-dimensional voxel matrix after three-dimensional frequency domain to spatial domain conversion.

[0067] The relaxation spectrum inversion unit is used to output the lateral relaxation time distribution under the non-negative constraint of relaxation spectrum amplitude, taking the echo train of each voxel in the three-dimensional voxel matrix as input.

[0068] The physical property parameter inversion unit is used to take the transverse relaxation time distribution, pure water signal reference, overall porosity scale, free fluid index and bound fluid index as inputs, and output porosity voxel field, particle size parameter voxel field and permeability voxel field according to porosity scale relationship, particle size empirical mapping relationship and permeability empirical relationship.

[0069] The sequence analysis unit is used to take the three-dimensional voxel matrix of adjacent phases as input, identify the single-phase depositional sequence through voxel differences, and spatially superimpose the single-phase depositional sequences in chronological order to output a four-dimensional configuration model.

[0070] The output unit is used to take a four-dimensional configuration model, a porosity voxel field, and a permeability voxel field as inputs, and output three-dimensional visualization results, three-dimensional geomorphological results of sedimentary bottom type, and quantitative evaluation results of microphysical properties.

[0071] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows:

[0072] This invention solves the problem that traditional methods require destroying samples for cross-sectional observation, which makes it impossible to obtain a complete and continuous three-dimensional internal structure without damage, resulting in a serious loss of spatial information.

[0073] This invention addresses the problem that conventional surface scanning techniques cannot penetrate sediments, can only obtain the top topography, and are powerless to obtain key information such as internal stratification and physical property distribution.

[0074] This invention addresses the problem that existing imaging technologies are mostly used for macroscopic morphology recognition, making it difficult to perform non-destructive, quantitative, and three-dimensional spatial mapping of key rock physical parameters such as porosity and permeability.

[0075] This invention solves the problem that traditional methods cannot capture the intermediate states of multi-stage sedimentary evolution, resulting in a spatiotemporal disconnect between the sedimentary dynamics process and the final configuration.

[0076] This invention addresses the problem that data on macroscopic morphology, microstructure, and physical properties obtained through different technologies differ in scale, coordinate system, and principle, lacking a unified framework for effective registration and collaborative interpretation, resulting in fragmented understanding.

[0077] This invention designs a physical simulation tank suitable for nuclear magnetic resonance (NMR) measurement environments. The main body of the tank is made of a non-magnetic, high-strength, and radio-frequency transparent composite glass material. The tank is equipped with a triaxial gradient magnetic field system, which can apply controllable excitation and spatially encoded magnetic fields to the sediments inside the tank. Based on this, a four-dimensional workflow of in-situ static placement, dynamic triggering, temporal evolution, and multi-scale fusion is proposed.

[0078] This invention enables simultaneous multi-scale property inversion: using T2 spectral technology, quantitative characterization of macroscopic topography and microscopic properties (porosity, particle size) is achieved simultaneously within the same workflow.

[0079] This invention enables true four-dimensional dynamic monitoring: through a mechanism of triggered scanning after sedimentation stabilization and in-situ scanning, it achieves for the first time non-destructive capture of turbidity current deposition processes (such as scouring, filling, and overflow), solving the problem of spatiotemporal separation between sedimentation dynamics and final configuration.

[0080] This invention is non-destructive and highly accurate: it avoids the damage to the original sedimentary structure caused by slicing and dissection, and the voxel resolution can reach the millimeter level, providing high-fidelity physical experimental data for geological modeling of deep-water oil and gas reservoirs.

[0081] (1) The expected benefits and commercial value of the technical solution of this invention after transformation are as follows: This invention directly empowers the reservoir geological modeling and reservoir numerical simulation links in deepwater oil and gas exploration and development. In view of the core pain points of traditional modeling, such as low spatial resolution and large uncertainty caused by relying on well logging data interpolation or core extrapolation, this invention can perfectly restore the sedimentary process of deepwater turbidite bodies under controlled experimental conditions, and obtain the three-dimensional spatial distribution of its internal physical parameters in situ and simultaneously, thereby providing a high-fidelity training dataset and verification benchmark for geological modeling and numerical simulation, which has significant industrial transformation value.

[0082] (2) The technical solution of this invention fills a technical gap in the industry both domestically and internationally: This invention breaks through the limitations of traditional characterization methods (macro-topographic photography, CT density imaging, and endpoint slice sampling) that cannot simultaneously satisfy non-destructive, three-dimensional, and quantitative physical property requirements, and overcomes the technical bottleneck that conventional nuclear magnetic resonance equipment can only detect centimeter-level plunger cores due to coil size limitations. This invention achieves, for the first time both domestically and internationally, three major leaps in spatial scale (expanding from centimeter-level cores to meter-level flumes), temporal dimension (expanding from endpoint static measurement to multi-phase dynamic tracking), and information dimension (simultaneous inversion of three independent physical property parameters using a single echo), filling the technical gap of lacking a unified experimental characterization method between macro-sedimentary configurations and micro-rock physical properties.

[0083] (3) The technical solution of this invention solves a technical problem that people have long desired to solve but have never been able to achieve: This invention successfully solves a long-standing geological problem in the fields of deep-water sedimentology and reservoir geology, namely, quantitatively characterizing the strong heterogeneity of reservoir properties that can differ by 1-2 orders of magnitude between different sedimentary configuration units (channels, natural dikes, etc.) without destroying the original structure of the sedimentary body. Through the combination of large-size planar magnets + triaxial gradient coding + voxel-level T2 spectrum inversion, the three core sub-problems of seeing the whole picture (covering a range of 2m×1m), seeing the details (1.95mm voxel resolution), and seeing the accuracy (directly outputting physical properties rather than inferring them indirectly by density) are simultaneously overcome in the same set of equipment and the same non-destructive scan, and a quantitative mapping relationship between sedimentary configuration, physical property distribution, and seepage capacity is established.

[0084] (4) The technical solution of the present invention overcomes technical bias: The present invention completely overturns the technical bias in the industry that nuclear magnetic resonance imaging equipment must work in a highly shielded, narrow and closed cavity, and cannot perform full-field imaging of large-scale, sediment-containing and dynamically changing water tanks. The present invention resolves this contradiction through non-obvious innovative means: First, it uses a dual-plane parallel plate superconducting magnet to transform the effective magnetic field from a closed cavity to an open region; second, it introduces a distributed radio frequency transmitting and receiving array to overcome the problem of non-uniform radio frequency field in large space; third, it pioneers an intermittent workflow of deposition-static-triggered scanning-recovery experiment, and successfully resolves the inherent conflict between the static requirements of nuclear magnetic resonance scanning and the dynamic characteristics of deposition experiments through time decoupling. Attached Figure Description

[0085] Figure 1 This is a flowchart of a three-dimensional fine characterization method for deep-water turbidite physical simulation of sedimentary bottom type based on nuclear magnetic resonance, provided in an embodiment of the present invention.

[0086] Figure 2 This is a structural block diagram of a three-dimensional fine characterization system for deep-water turbidite physical simulation of sedimentary bottom types based on nuclear magnetic resonance, provided in an embodiment of the present invention.

[0087] Figure 3 This is a flowchart of a three-dimensional fine characterization method for deep-water turbidite physical simulation of sedimentary bottom type based on nuclear magnetic resonance, provided in an embodiment of the present invention.

[0088] Figure 4 This is a structural diagram of the dual-plane static magnetic field generator provided in an embodiment of the present invention.

[0089] Figure 5 This is a comparison diagram of the T2 relaxation distribution of different depositional configuration units provided in the embodiments of the present invention.

[0090] Figure 6 This is a spatiotemporal evolution diagram of multi-stage deposition porosity and permeability provided in an embodiment of the present invention.

[0091] Figure 7 This is a comparison chart of heterogeneity evaluation parameters provided in the embodiments of the present invention.

[0092] Figure 8 This is a 3D porosity volume drawing and digital terrain simulation registration map provided in the embodiments of the present invention.

[0093] Figure 4 In the middle: 1. Superconducting magnet plate; 2. Gradient magnetic field coil X; 3. Signal receiving coil; 4. Deposition substrate; 5. Material supply system; 6. Gradient magnetic field coil Y; 7. Non-magnetic composite glass; 8. Superconducting magnet plate; 9. Gradient magnetic field coil Z; 10. High-precision turbidimeter; 11. Signal receiving coil; 12. Electromagnetic pulse transmitter. Detailed Implementation

[0094] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0095] Nuclear magnetic resonance (NMR): A physical interpretation method that quantitatively obtains reservoir porosity, pore structure, fluid saturation, and mobile fluid distribution based on the relaxation response of hydrogen nuclei in formation fluids under the action of an applied static magnetic field and radio frequency field.

[0096] T2 distribution spectrum: The transverse relaxation time distribution spectrum of porous media fluid obtained by nuclear magnetic resonance relaxation inversion. Its T2 value is positively correlated with the pore size, which characterizes the occurrence characteristics of fluid in pores of different sizes.

[0097] Configurational unit: The basic constituent unit of sedimentary bodies in deep-water turbidite systems that has a clear sedimentary origin, a defined geometric shape, and continuous physical properties, such as turbidite channel complexes, single channels, lobes, natural dikes, and overflow sedimentary bodies.

[0098] Voxel: The smallest volumetric unit in a three-dimensional digital model that represents the distribution of spatial properties. It is the basic computational unit for constructing three-dimensional reservoir configuration models and performing spatial characterization of pore structure and fluid.

[0099] like Figure 1 As shown in the figure, the present invention provides a three-dimensional fine characterization method for deep-water turbidite sedimentary bottom based on nuclear magnetic resonance physical simulation, which includes the following steps:

[0100] S101, a non-magnetic experimental water tank and nuclear magnetic resonance scanning system were built so that the core observation section of the non-magnetic experimental water tank was within the combined influence of the static magnetic field, the triaxial gradient magnetic field, the radio frequency excitation field and the signal receiving area.

[0101] S102, Inject background fluid into the non-magnetic experimental water tank, perform background calibration scanning on the background fluid, obtain the NMR signal reference under pure water background, and set the turbidity release parameters, turbidity trigger parameters and NMR scan timing parameters;

[0102] S103, a turbidity mixture is released upstream of the non-magnetic experimental water tank through the material supply system, and the turbidity change of the water body is monitored by the turbidity monitoring system. When the turbidity exceeds the frontal identification threshold, the arrival time of the turbidity is recorded. When the turbidity is lower than the background steady-state threshold and remains stable for a preset time, the sedimentation stabilization time is recorded.

[0103] S104, at the moment of sedimentation stabilization, in-situ nuclear magnetic resonance scanning is initiated. Hydrogen nuclei in the sedimentation system are excited by radio frequency pulses, and the observation area is spatially encoded in three dimensions by a triaxial gradient magnetic field. Spin echo signals are collected by a signal receiving coil to obtain the raw nuclear magnetic signal dataset at the corresponding moment.

[0104] S105, using the original NMR signal dataset as input, performs three-dimensional voxel reconstruction to obtain the three-dimensional voxel matrix at the corresponding time point;

[0105] S106 takes the echo train of each voxel in the three-dimensional voxel matrix as input, and outputs the lateral relaxation time distribution according to the relaxation relationship of the echo amplitude decaying with the acquisition time, under the constraint of the non-negativity of the relaxation component amplitude.

[0106] S107 takes the transverse relaxation time distribution, the NMR signal benchmark under pure water background, and the overall porosity scale as inputs and outputs a porosity voxel field; takes the logarithmic mean relaxation time corresponding to the transverse relaxation time distribution as input and outputs a grain size parameter voxel field according to the empirical mapping relationship between the logarithmic mean relaxation time and the median grain size of the sediment; and divides the free fluid index and bound fluid index based on the transverse relaxation time distribution, and combines the porosity voxel field to output a permeability voxel field according to the empirical relationship of permeability.

[0107] S108, in the order of turbidity release, turbidity judgment, in-situ scanning and three-dimensional voxel reconstruction, multiple cyclic acquisitions were carried out. The three-dimensional voxel matrix of adjacent phases was used as input to identify voxel differences, output single-phase sedimentary sequences, and the single-phase sedimentary sequences were spatially superimposed in chronological order to obtain a four-dimensional configuration model of the deep-water turbidite sedimentary bottom.

[0108] S109 takes a four-dimensional configuration model, porosity voxel field, particle size parameter voxel field, and permeability voxel field as inputs and outputs three-dimensional visualization results, three-dimensional geomorphological results of sedimentary bottom type, and quantitative evaluation results of microphysical properties.

[0109] The construction of the non-magnetic experimental water tank and nuclear magnetic resonance scanning system provided in this embodiment of the invention includes:

[0110] A non-magnetic deposition simulation water tank with a length of 2m, a width of 1m, and a height of 0.5m was constructed. The non-magnetic deposition simulation water tank was formed by splicing non-magnetic glass composite material with a thickness of 3cm, and the joints were sealed with modified epoxy resin.

[0111] Two planar superconducting magnet plates are arranged in parallel on the upper and lower sides of the core observation section of the non-magnetic sedimentation simulation tank to create a static magnetic field environment with a central magnetic field strength of 0.5T and a magnetic field uniformity better than 50ppm in the core observation section.

[0112] A triaxial gradient magnetic field coil is arranged outside the non-magnetic sedimentation simulation tank. The triaxial gradient magnetic field coils correspond to the X, Y and Z directions respectively, and the maximum linear gradient field strength is 20 mT per meter.

[0113] A radio frequency pulse transmitter was configured in the observation section of the non-magnetic sedimentation simulation tank to match the radio frequency center frequency with the Larmor resonance frequency of the hydrogen nucleus under a static magnetic field of 0.5T.

[0114] A signal receiving coil is placed close to the outer wall of a non-magnetic deposition simulation tank so that the signal receiving coil receives the spin echo signal generated by the transverse relaxation of hydrogen nuclei during the interval of radio frequency pulses.

[0115] A 200L slurry pump was connected to the upstream of the non-magnetic deposition simulation tank, and a turbid mixture was delivered to the tank through a wear-resistant hose.

[0116] The background calibration scan provided in this embodiment of the invention includes:

[0117] Low-mineralized water to a depth of 400 mm was injected into the non-magnetic experimental water tank as background fluid.

[0118] The background fluid is pre-scanned, and the free induction attenuation signal is acquired to determine the longitudinal relaxation time and transverse relaxation time of the background fluid.

[0119] The NMR signal amplitude of the background fluid was calibrated as a pure water signal reference.

[0120] The pure water signal reference is input into the porosity inversion stage to make the initial amplitude of each voxel echo proportional to the pure water signal reference, and outputs the porosity value of the corresponding spatial location under the overall porosity scale constraint.

[0121] The release and triggering control of turbid mixtures provided in this embodiment of the invention includes:

[0122] Add 100-mesh quartz sand, 125-mesh calcite and 500-mesh kaolin to the mixing chamber and prepare the solid phase composition according to the following mass fractions: quartz sand 15%, calcite 30% and kaolin 55%.

[0123] A turbid mixture with a volume concentration of 25% and an allowable deviation of 2% was prepared by stirring continuously at a stirring speed of 300 rpm for 20 min.

[0124] A turbid mixture with a single volume of 50L is released upstream of the non-magnetic experimental water tank at a flow rate of 5L per second using a slurry pump.

[0125] Turbidity changes were monitored using a turbidimeter with a range of 0 to 4000 NTU. The arrival time of turbidity flow was recorded when the turbidity rose by more than 50 NTU, and the deposition stabilization time was recorded after the turbidity dropped to less than 60 NTU and remained stable for more than 5 minutes. The deposition stabilization time was used as the trigger time for in-situ nuclear magnetic resonance scanning.

[0126] The in-situ nuclear magnetic resonance (NMR) scan provided in this embodiment of the invention includes:

[0127] A layer-selective gradient is applied along the Z-direction when a 90-degree radio frequency pulse is emitted to define the target imaging layer.

[0128] A phase encoding gradient is applied in the Y direction at the leading edge of spin echo generation, so that hydrogen nuclei with different Y coordinates can obtain different phase information;

[0129] During the spin echo acquisition window, a frequency coding gradient is applied along the X direction so that hydrogen nuclei with different X coordinates obtain different frequency information.

[0130] Using the static magnetic field strength, the gradient magnetic field strength in the X direction, the gradient magnetic field strength in the Y direction, the gradient magnetic field strength in the Z direction, and the spatial coordinates as inputs, according to the relationship between the hydrogen nucleus Larmor resonance frequency and the magnetic field strength and spatial position, a spin echo signal carrying three-dimensional spatial position information is output, and this spin echo signal is input into the three-dimensional voxel reconstruction stage.

[0131] The three-dimensional voxel reconstruction, multi-scale property inversion, and four-dimensional configuration analysis provided in this invention include:

[0132] The raw NMR signal datasets from each period are transformed from the three-dimensional frequency domain to the spatial domain, and the corresponding three-dimensional voxel matrix is ​​output. Each voxel corresponds to a physical space with a side length of 1.95 mm.

[0133] Using the echo trains within each voxel as input, the transverse relaxation time distribution is output under the non-negative constraint of the relaxation spectrum amplitude by employing singular value decomposition combined with non-negative least squares solution.

[0134] The initial amplitude of the echo is obtained by using the signal integral of the transverse relaxation time distribution as input, and the porosity voxel field is output by combining the pure water signal reference and the overall porosity scale.

[0135] The logarithmic mean relaxation time is calculated using the transverse relaxation time distribution as input, and the particle size parameter voxel field is output according to the empirical proportionality coefficient and empirical exponential relationship between the logarithmic mean relaxation time and the median particle size.

[0136] Using the porosity voxel field, the free fluid index and the bound fluid index obtained by dividing the transverse relaxation time cutoff value of 33ms as inputs, the permeability voxel field is output according to the Timur-Coates empirical relationship.

[0137] Boolean difference operations are performed on the three-dimensional voxel matrices of adjacent phases as input to output the new single-phase depositional sequence within the corresponding time interval. The single-phase depositional sequences are then spatially aligned and superimposed in chronological order to output a four-dimensional configuration model.

[0138] The present invention provides a method for three-dimensional fine characterization of sedimentary bottom types based on nuclear magnetic resonance physical simulation of deep-water turbidite deposits, comprising the following steps:

[0139] (1) Construction of a non-magnetic experimental water tank and nuclear magnetic resonance scanning system;

[0140] (2) Pre-experiment preparation and baseline calibration;

[0141] (3) Intermittent NMR dynamic acquisition based on electromagnetic pulse excitation and coil reception;

[0142] (4) Multiscale property inversion and four-dimensional configuration analysis;

[0143] (5) 3D visualization, configuration analysis and output of results.

[0144] The non-magnetic experimental water tank and nuclear magnetic resonance scanning system provided in this embodiment of the invention are as follows:

[0145] Step 1: Construct a non-magnetic sedimentation simulation tank; the main body dimensions are set as length L=2m, width W=1m, and height H=0.5m; the tank body is made of 3cm thick glass composite material with high wave transmittance and no magnetism, and the joints are sealed with modified epoxy resin to ensure that it can withstand an internal water pressure of 0.02MPa and does not produce magnetic field distortion.

[0146] Step 2: Construct a dual-plane static magnetic field generating device; two large planar superconducting magnet plates are arranged in parallel on the upper and lower sides of the core observation section of the water tank to provide a highly stable magnetic field; the central magnetic field strength is set to B0=0.5T, and a static magnetic field environment with a magnetic field uniformity better than 50ppm is established inside the water tank to complete the polarization of fluid hydrogen protons.

[0147] Step 3: Deploy the triaxial gradient magnetic field; install the gradient magnetic field coil system; arrange gradient coils in the X (left-right), Y (front-back), and Z (up-down) directions respectively, and set the maximum linear gradient field strength to [value missing]. , used to implement spatial encoding;

[0148] Step 4: Install an electromagnetic pulse transmitter; configure a radio frequency (RF) pulse transmitter in the observation section of the non-magnetic water tank; based on the static magnetic field strength of B0=0.5T, set the RF center frequency to match the Larmor resonance frequency of hydrogen nuclei (approximately 21.3MHz), and use it to transmit specific excitation and re-aggregation electromagnetic pulses (90° and 180° pulses in the CPMG sequence) to the deposition system inside the water tank to achieve resonant excitation of hydrogen nuclei in the fluid;

[0149] Step 5: Deploy the signal receiving coil; Deploy a high-sensitivity radio frequency receiving coil close to the outer wall of the water tank, specifically for capturing the weak spin echo signal generated during the transverse relaxation of hydrogen nuclei during the pulse interval, and transmitting the signal to the data acquisition system;

[0150] Step 6: Set up the material supply system; connect a 200L slurry pump and transport the turbid mixture to the upstream of the tank through a wear-resistant hose.

[0151] The experimental preparation and background calibration provided in this embodiment of the invention are as follows:

[0152] Step 7: Inject background fluid; Inject low-mineralized water with a depth of h=400mm into the water tank as background fluid to ensure that the radio frequency field can effectively penetrate the detection area of ​​the water tank; In order to simulate the dynamic viscosity environment of real seawater, add an appropriate amount of low magnetic viscosity enhancer to the background fluid, while suppressing the paramagnetic effect on the surface of clay minerals.

[0153] Step 8: Deploy a turbidity monitoring system; suspend a high-precision turbidity meter (range 0-4000 NTU) in the middle of the water tank to monitor changes in water turbidity in real time, and set the trigger threshold to 50 NTU;

[0154] Step 9: Perform background calibration scan; use nuclear magnetic resonance system to pre-scan the pure water background, acquire free induction decay (FID) signal, measure the longitudinal relaxation time of about 2500ms and the transverse relaxation time of about 2000ms under this environment, and calibrate the amplitude of this signal as 100% water signal reference S0.

[0155] Step 10: Preset the scanning timing parameters in the control system; set the single turbidity release volume to V. inj =50L; Set NMR scan sequence parameters: repetition wait time T R =8000ms, echo time T E =0.5ms, number of echoes NECH=6000, scan resolution set to 1024×512×256 voxels.

[0156] Step 11: Prepare the experimental turbidity mixture; weigh out 100-mesh quartz sand (15%), 125-mesh calcite (30%), and 500-mesh kaolin (55%) by mass fraction, and add them to the mixing chamber; set the stirring speed to 300 rpm and stir continuously for 20 minutes to ensure uniform solid-liquid mixing, preparing a mixture with a volume concentration of C. v =25%±2% high-concentration non-Newtonian fluid (turbidity).

[0157] The present invention provides an intermittent NMR dynamic acquisition method based on electromagnetic pulse excitation and coil reception:

[0158] Step 12: Start the first batch of turbidity release (T1); start the high-pressure slurry pump to release the first batch (T1) of experimental turbidity upstream of the tank at a flow rate of Q=5L / s;

[0159] Step 13: Dynamic monitoring and trigger condition judgment; use the turbidity meter on the side wall of the tank to monitor the arrival signal of the turbidity front; when the turbidity at the monitoring point rises above the threshold (>50 NTU), the system automatically marks it as time t0; when the turbidity meter value drops to the preset background steady-state threshold (<60 NTU), and this state remains stable for more than 5 minutes, the system determines that the batch of sediment and suspended matter has completely stopped, and this time is marked as time t1;

[0160] Step 14: Perform three-dimensional orthogonal gradient spatial coding; layer selection coding: at the instant of transmitting a 90° RF pulse, apply a short layer selection gradient along the Z-axis. To lock onto the target imaging layer, it is then quickly shut down; Phase encoding: A microsecond-level phase encoding gradient is applied along the Y-axis in a very short time before the echo signal is generated. Assigning specific phases to spin nuclei at different Y coordinates, then turning them off; Frequency encoding (signal readout): Continuously applying a frequency-encoded gradient along the X-axis throughout the entire acquisition window for spin echo generation. This linearly varying magnetic field encodes the spatial X-coordinate information into the instantaneous frequency of the spin echo signal; the Larmor resonance frequency of any spatial point (x, y, z). Follow the following spatial coding equation:

[0161]

[0162] in, The gyromagnetic ratio of the hydrogen proton; , , These represent the triaxial gradient magnetic field strengths; (x, y, z) are the spatial coordinates. The Larmor angular frequency (unit: rad / s or MHz);

[0163] Step 15: Trigger in-situ nuclear magnetic resonance scanning; the control system sends a start command to the radio frequency transmitter to activate the CPMG pulse sequence and continuously acquire the echo train generated by the CPMG sequence until the preset 6000 echo samples are completed; the final dataset is the three-dimensional k-space data at time t1, which contains the density and relaxation information of all hydrogen protons inside the sediment at that time.

[0164] Step 16: Resume the experiment and inject the next batch; restart the slurry pump to inject the second batch (T2) of turbidity (volume 50L), and the new sediment will be superimposed on the T1 bottom.

[0165] Step 17: Loop data acquisition; Execute steps 12 to 16 repeatedly; According to the experimental design, perform n=5 cycles of injection and scanning, and sequentially acquire t1, t2, ..., t n The raw NMR signal dataset at time points D1, D2, ..., D n .

[0166] The multi-scale property inversion and four-dimensional configuration analysis provided by the embodiments of the present invention:

[0167] Step 18: 3D voxel reconstruction; For the collected K-space data D from each period... i Applying a three-dimensional fast Fourier transform (3D-FFT) to convert it from the frequency domain to the spatial domain, reconstructing it into a three-dimensional voxel matrix V. i (x, y, z), each voxel represents a side length of 1.95 × 1.95 × 1.95 mm. 3 The physical space;

[0168] Step 19: Single voxel T2 spectrum inversion; using singular value decomposition (SVD) combined with nonnegative least squares (NNLS) algorithm, perform multi-exponential fitting on the echo train S(t) within each voxel to solve the transverse relaxation time distribution spectrum f(T2):

[0169]

[0170] Where S(t) represents the amplitude of the spin echo signal that decays over time; t is the echo time (acquisition time). The distribution function of the transverse relaxation time T2 (i.e., the signal amplitude corresponding to each T2 component);

[0171] Step 20: Porosity ( 3D inversion; calculation of the initial amplitude of the echo within each voxel. (i.e., the signal integral at t→0), and compared with the pure water signal S0 calibrated at the background, to obtain the porosity value at that location:

[0172]

[0173] in, For actual measurement Porosity, used for calibration; For spatial location Porosity at that location; The initial signal amplitude extrapolated to zero time from the echo train within the voxel;

[0174] Step 21: Particle size parameter inversion; Based on the positive correlation between T2 value and pore throat radius r (T2∝r), calculate the logarithmic mean relaxation time T of the voxel. 2lm Establish T 2lm Compared with the median grain size d of sediments 50 Empirical mapping equation:

[0175]

[0176] in, denoted as median particle size at spatial location; k is the empirical proportionality coefficient; α is the empirical exponent.

[0177] Step 22: Three-dimensional estimation of permeability; applying the Timur-Coates empirical formula, combined with porosity... and T2 cutoff value ( ) Divide the fluid into movable and bound fluid ratios, and calculate the permeability K of each voxel:

[0178]

[0179] Where C is the empirical coefficient. The free fluid index, The bound fluid index; Permeability at a spatial location, in mD or μm 2 m is the porosity index; n is the empirical index, which is 2.

[0180] Step 23: Sequence dynamic stripping; Since solid sediment particles produce virtually no NMR response, turbidite deposits at the bottom, after displacing the original background water, will cause a decrease in the total NMR signal amplitude in that area; using Boolean difference operations of a three-dimensional matrix, t n-1 voxel matrix at time t minus t n The time-matrix precisely extracts the newly accumulated single-stage sedimentary sequences within that time interval. :

[0181]

[0182] Step 24: Multi-stage sequence overlay; combine all single-stage sedimentary layers. Spatial alignment and stacking based on time series data were used to construct a complete four-dimensional evolution model of deep-water turbidite deposits. .

[0183] The three-dimensional visualization, configuration analysis, and result output provided by the embodiments of this invention are as follows:

[0184] Step 25: 3D visualization rendering; import 4D model The ray-projection volume is drawn using a computer (equipped with a GPU); the opacity transfer function is set; for porosity contour plots: the porosity is set. >35% of the area is red (high porosity). Areas with less than 15% porosity are blue (low porosity), with a linear transition in the middle; for permeability distribution: a logarithmic scale is used for coloring. (Hypertonic, red) (Hypotency, blue);

[0185] Step 26: Generation of 3D geomorphological map of sedimentary bottom type; extraction of t n The top surface of the deposit at time (Z) top The coordinate data of the surface are used to generate a digital elevation model (DEM); the surface roughness (R) is calculated. s The system outputs a 3D terrain rendering map with contour lines (1cm interval) to visually display the waterway incision depth and spillway extent, along with slope variation.

[0186] Step 27: Quantitative evaluation of microscopic property heterogeneity; randomly select N=1000 sampling points on the three-dimensional voxel model, and statistically analyze the frequency distribution histogram of porosity and permeability; calculate the coefficient of variation (C). v ), grade difference (R) and breakthrough coefficient (T) k The degree of heterogeneity within a sedimentary body is quantitatively evaluated using the following formula:

[0187]

[0188] in Standard deviation; K is the average value; max K represents the maximum permeability. min This represents the minimum penetration rate.

[0189] like Figure 2 As shown in the figure, the embodiment of the present invention provides a three-dimensional fine characterization system for deep-water turbidite physical simulation sedimentary bottom based on nuclear magnetic resonance, including a non-magnetic sedimentation simulation tank, a static magnetic field generator, a triaxial gradient magnetic field coil, a radio frequency pulse transmitter, a signal receiving coil, a material supply system, a turbidity monitoring system, a data acquisition system, and a computer;

[0190] The non-magnetic sedimentation simulation tank is used to contain background fluid and turbidity mixture, and provides an observation space for turbidity movement, settling and deposition;

[0191] The static magnetic field generating device is set on the upper and lower sides of the core observation section of the non-magnetic deposition simulation tank, and is used to form a static magnetic field in the core observation section to cause hydrogen nuclei in the observation area to undergo orientation polarization.

[0192] The triaxial gradient magnetic field coils are arranged around the non-magnetic deposition simulation tank to generate spatially encoded gradients in the X, Y, and Z directions based on the static magnetic field.

[0193] The radio frequency pulse transmitter is used to transmit radio frequency pulses that match the Larmor resonance frequency of the hydrogen nucleus to the core observation segment;

[0194] The signal receiving coil is located near the outer wall of the non-magnetic deposition simulation tank and is used to receive the spin echo signal generated by the transverse relaxation of hydrogen nuclei during the interval of radio frequency pulses.

[0195] The material supply system is connected upstream of the non-magnetic deposition simulation tank and is used to release turbidity mixture into the non-magnetic deposition simulation tank.

[0196] The turbidity monitoring system is used to monitor the turbidity of the water in the non-magnetic sedimentation simulation tank and output turbidity arrival signal and sedimentation stability signal to the data acquisition system.

[0197] The data acquisition system is used to trigger the coordinated operation of the radio frequency pulse transmitter, the triaxial gradient magnetic field coil, and the signal receiving coil based on the deposition stability signal, and to acquire the raw NMR signal dataset.

[0198] The computer is used to take the raw NMR signal dataset as input and output three-dimensional voxel matrix, porosity voxel field, particle size parameter voxel field, permeability voxel field, four-dimensional configuration model, three-dimensional visualization results, sedimentary bottom three-dimensional geomorphological results, and quantitative evaluation results of microphysical properties.

[0199] The non-magnetic deposition simulation tank provided in this embodiment of the invention is a tank body with a length of 2m, a width of 1m, and a height of 0.5m. The tank body is formed by splicing non-magnetic glass composite material with a thickness of 3cm, and the joints are sealed with modified epoxy resin. The material supply system includes a slurry pump with a volume of 200L and a wear-resistant hose. The slurry pump is connected to the upstream of the non-magnetic deposition simulation tank through the wear-resistant hose.

[0200] The static magnetic field generating device provided in this embodiment of the invention includes two planar superconducting magnet plates, which are respectively arranged in parallel on the upper and lower sides of the core observation section of the non-magnetic deposition simulation tank. The static magnetic field generating device forms a static magnetic field environment in the core observation section with a central magnetic field strength of 0.5T and a magnetic field uniformity better than 50ppm. The maximum linear gradient field strength of the triaxial gradient magnetic field coil is 20mT per meter.

[0201] The computer provided in this embodiment of the invention includes a three-dimensional voxel reconstruction unit, a relaxation spectrum inversion unit, a physical property parameter inversion unit, a sequence analysis unit, and a result output unit;

[0202] The three-dimensional voxel reconstruction unit is used to take the original nuclear magnetic resonance signal dataset as input, and output a three-dimensional voxel matrix after three-dimensional frequency domain to spatial domain conversion.

[0203] The relaxation spectrum inversion unit is used to output the lateral relaxation time distribution under the non-negative constraint of relaxation spectrum amplitude, taking the echo train of each voxel in the three-dimensional voxel matrix as input.

[0204] The physical property parameter inversion unit is used to take the transverse relaxation time distribution, pure water signal reference, overall porosity scale, free fluid index and bound fluid index as inputs, and output porosity voxel field, particle size parameter voxel field and permeability voxel field according to porosity scale relationship, particle size empirical mapping relationship and permeability empirical relationship.

[0205] The sequence analysis unit is used to take the three-dimensional voxel matrix of adjacent phases as input, identify the single-phase depositional sequence through voxel differences, and spatially superimpose the single-phase depositional sequences in chronological order to output a four-dimensional configuration model.

[0206] The output unit is used to take a four-dimensional configuration model, a porosity voxel field, and a permeability voxel field as inputs, and output three-dimensional visualization results, three-dimensional geomorphological results of sedimentary bottom type, and quantitative evaluation results of microphysical properties.

[0207] Specific implementation of the present invention:

[0208] like Figure 3 I. Construction of Non-Magnetic Experimental Water Tank and Nuclear Magnetic Resonance Scanning System

[0209] Step 1: Construct a non-magnetic sedimentation simulation tank. The main dimensions are set as follows: length L=2m, width W=1m, and height H=0.5m. The tank body is constructed from 3cm thick glass composite material with high wave transmittance and no magnetic field. The joints are sealed with modified epoxy resin to ensure that it can withstand an internal water pressure of 0.02MPa without generating magnetic field distortion.

[0210] Step 2: Construct a dual-plane static magnetic field generator. Two large planar superconducting magnet plates are arranged parallel to each other on the upper and lower sides of the core observation section of the water tank to provide a highly stable magnetic field. The central magnetic field strength is set to B0 = 0.5T, and a static magnetic field environment with a magnetic field uniformity better than 50ppm is established inside the water tank to complete the polarization of fluid hydrogen protons.

[0211] Step 3: Deploy the triaxial gradient magnetic field. Install the gradient magnetic field coil system. Arrange gradient coils in the X (left-right), Y (front-back), and Z (up-down) directions respectively, and set the maximum linear gradient field strength to [value missing]. , used to implement spatial encoding.

[0212] Step 4: Install the electromagnetic pulse transmitter. Configure a radio frequency (RF) pulse transmitter in the observation section of the non-magnetic tank. Based on the static magnetic field strength of B0 = 0.5T, set the RF center frequency to match the Larmor resonance frequency of hydrogen nuclei (approximately 21.3MHz). This is used to transmit specific excitation and re-aggregation electromagnetic pulses (90° and 180° pulses in the CPMG sequence) into the deposition system inside the tank to achieve resonant excitation of hydrogen nuclei in the fluid.

[0213] Step 5: Deploy the signal receiving coil. A high-sensitivity radio frequency receiving coil is deployed close to the outer wall of the water tank. This coil is specifically designed to capture the weak spin echo signal generated during the transverse relaxation of hydrogen nuclei during the pulse interval and transmit the signal to the data acquisition system.

[0214] Step 6: Set up the material supply system. Connect a 200L slurry pump and transport the turbid mixture to the upstream of the tank via a wear-resistant hose.

[0215] like Figure 4 The dual-plane static magnetic field generating device provided in this embodiment of the invention includes: a superconducting magnet plate 1, a gradient magnetic field coil X 2, a signal receiving coil 3, a deposition substrate 4, a material supply system 5, a gradient magnetic field coil Y 6, a non-magnetic composite glass 7, a superconducting magnet plate 8, a gradient magnetic field coil Z 9, a high-precision turbidimeter 10, a signal receiving coil 11, and an electromagnetic pulse transmitter 12.

[0216] II. Pre-experimental preparation and baseline calibration

[0217] Step 7: Inject background fluid. Inject low-mineralized water to a depth of h=400mm into the water tank as background fluid to ensure that the radio frequency field effectively penetrates the detection area of ​​the water tank. To simulate the dynamic viscosity environment of real seawater, add an appropriate amount of low-magnetic viscosity enhancer to the background fluid, while suppressing the paramagnetic effect of clay mineral surfaces.

[0218] Step 8: Deploy the turbidity monitoring system. A high-precision turbidity meter (range 0-4000 NTU) is suspended in the middle of the water tank to monitor changes in water turbidity in real time. A threshold of 50 NTU is set for triggering turbidity flow during an increase in turbidity; a threshold of 60 NTU is set for triggering the end of turbidity flow deposition during a decrease in turbidity.

[0219] Step 9: Perform background calibration scan. Use an NMR system to pre-scan the pure water background, acquire free induction decay (FID) signals, and measure the longitudinal relaxation time of approximately 2500 ms and the transverse relaxation time of approximately 2000 ms under this environment. Then, calibrate the amplitude of this signal as the 100% water signal reference S0.

[0220] Step 10: Preset the scan timing parameters in the control system. Set the single turbidity release volume to V.inj =50L; Set NMR scan sequence parameters: repetition wait time T R =8000ms, echo time T E =0.5ms, number of echoes NECH=6000, scan resolution set to 1024×512×256 voxels.

[0221] Step 11: Prepare the experimental turbidity mixture. Weigh out 15% (100-mesh quartz sand), 30% (125-mesh calcite), and 55% (500-mesh kaolin) by mass fraction and add them to the mixing chamber. Set the stirring speed to 300 rpm and stir continuously for 20 minutes to ensure uniform solid-liquid mixing, preparing a mixture with a volume concentration of C. v =25%±2% high-concentration non-Newtonian fluid (turbidity).

[0222] III. Intermittent NMR Dynamic Acquisition Based on Electromagnetic Pulse Excitation and Coil Reception

[0223] Step 12: Initiate the first batch of turbidity release (T1). Start the high-pressure slurry pump to release the first batch (T1) of experimental turbidity upstream of the tank at a flow rate of Q=5L / s.

[0224] Step 13: Dynamic Monitoring and Trigger Condition Determination. The arrival signal of the turbidity front is monitored using a turbidity meter on the sidewall of the tank. When the turbidity at the monitoring point rises above the threshold (>50 NTU), the system automatically marks this as time t0. When the turbidity meter reading drops to the preset background steady-state threshold (<60 NTU), and this state remains stable for more than 5 minutes, the system determines that the batch of sediment and suspended matter has completely settled, and this is marked as time t1.

[0225] Step 14: Perform 3D orthogonal gradient spatial coding. Layer selection coding: At the instant of transmitting a 90° RF pulse, apply a short layer selection gradient along the Z-axis. To lock onto the target imaging layer, it is then quickly shut down; Phase encoding: A microsecond-level phase encoding gradient is applied along the Y-axis in a very short time before the echo signal is generated. Assigning specific phases to spin nuclei at different Y coordinates, then turning them off; Frequency encoding (signal readout): Continuously applying a frequency-encoded gradient along the X-axis throughout the entire acquisition window for spin echo generation. This linearly varying magnetic field encodes the spatial X-coordinate information into the instantaneous frequency of the spin echo signal. The Larmor resonance frequency of any spatial point (x, y, z) is... Follow the following spatial coding equation:

[0226]

[0227] in, The gyromagnetic ratio of the hydrogen proton; , , These represent the triaxial gradient magnetic field strengths; (x, y, z) are the spatial coordinates. This is the Larmor angular frequency (unit: rad / s or MHz).

[0228] Step 15: Trigger in-situ nuclear magnetic resonance (NMR) scanning. The control system sends a start command to the radio frequency transmitter, activating the CPMG pulse sequence and continuously acquiring the echo train generated by the CPMG sequence until the preset 6000 echo samples are completed. The final dataset obtained is the three-dimensional k-space data at time t1, which contains the density and relaxation information of all hydrogen protons inside the sediment at that time.

[0229] Step 16: Resume the experiment and inject the next batch. Restart the slurry pump to inject the second batch (T2) of turbidity (50L volume), and the new sediment will be superimposed on the T1 substrate.

[0230] Step 17: Loop Data Acquisition. Repeat steps 12 to 16. According to the experimental design, perform n=5 cycles of injection and scanning, sequentially acquiring t1, t2, ..., t... n The raw NMR signal dataset at time points D1, D2, ..., D n .

[0231] IV. Multiscale property inversion and four-dimensional configuration analysis

[0232] Step 18: 3D Voxel Reconstruction. This involves reconstructing the collected K-space data from various periods. i Applying a three-dimensional fast Fourier transform (3D-FFT) to convert it from the frequency domain to the spatial domain, reconstructing it into a three-dimensional voxel matrix V. i (x, y, z), each voxel represents a side length of 1.95 × 1.95 × 1.95 mm. 3 The physical space.

[0233] Step 19: Single voxel T2 spectrum inversion. Using singular value decomposition (SVD) combined with nonnegative least squares (NNLS) algorithm, multi-exponential fitting is performed on the echo train S(t) within each voxel to solve for the transverse relaxation time distribution spectrum f(T2):

[0234]

[0235] Where S(t) represents the amplitude of the spin echo signal that decays over time; t is the echo time (acquisition time). lateral relaxation time The distribution function (i.e., the signal amplitude corresponding to each T2 component).

[0236] Step 20: Porosity ( 3D inversion. Calculate the initial amplitude of the echo within each voxel. (i.e., the signal integral at t→0), and compared with the pure water signal S0 calibrated at the background, to obtain the porosity value at that location:

[0237]

[0238] in, For actual measurement Porosity, used for calibration; For spatial location Porosity at that location; The initial signal amplitude is extrapolated to zero time from the echo train within the voxel.

[0239] Step 21: Particle size parameter inversion. Based on the positive correlation between the T2 value and the pore throat radius r (T2∝r), calculate the logarithmic mean relaxation time T of the voxel. 2lm Establish T 2lm Compared with the median grain size d of sediments 50 Empirical mapping equation:

[0240]

[0241] in, denoted as median particle size at spatial location; k is the empirical proportionality coefficient; and α is the empirical index.

[0242] Step 22: Three-dimensional permeability estimation. Apply the Timur-Coates empirical formula, combined with porosity... and T2 cutoff value ( ) Divide the fluid into movable and bound fluid ratios, and calculate the permeability K of each voxel:

[0243]

[0244] Where C is the empirical coefficient. The free fluid index, The bound fluid index; Permeability at a spatial location, in mD or μm 2 m is the porosity index; n is the empirical index, which is 2.

[0245] Step 23: Sequence Dynamic Stripping. Since solid sediment particles produce virtually no NMR response, the accumulation of turbidite at the bottom, displacing the original background water, leads to a decrease in the total NMR signal amplitude in that region. Using Boolean difference operations on a three-dimensional matrix, t... n-1 voxel matrix at time t minus t n The time-matrix precisely extracts the newly accumulated single-stage sedimentary sequences within that time interval. :

[0246]

[0247] Step 24: Multi-stage sequence overlay. Overlay all single-stage deposition layers. Spatial alignment and stacking based on time series data were used to construct a complete four-dimensional evolution model of deep-water turbidite deposits. .

[0248] V. 3D Visualization, Configuration Analysis, and Output of Results

[0249] Step 25: 3D visualization rendering. Import the 4D model. The ray-projection volume is drawn using a computer (equipped with a GPU). The opacity transfer function is set: For porosity contour plots: the porosity is set. >35% of the area is red (high porosity). Areas with less than 15% porosity are blue (low porosity), with a linear transition in the middle; for permeability distribution: a logarithmic scale is used for coloring. (Hypertonic, red) (Hypotency, blue).

[0250] Step 26: Generation of 3D topographic map of sedimentary bottom type. Extract t n The top surface of the deposit at time (Z) top The coordinate data of the surface are used to generate a digital elevation model (DEM). The surface roughness (R) is then calculated. s The system outputs a 3D terrain rendering map with contour lines (1cm interval) to visually display the waterway incision depth and overflow range, along with slope variation.

[0251] Step 27: Quantitative evaluation of microscopic property heterogeneity. Randomly select N=1000 sampling points on the three-dimensional voxel model and statistically analyze the frequency distribution histograms of porosity and permeability. Calculate the coefficient of variation (C0). v ), grade difference (R) and breakthrough coefficient (T) k This method quantitatively evaluates the degree of heterogeneity within a sedimentary body. The calculation formula is as follows:

[0252]

[0253] in Standard deviation; K is the average value; max K represents the maximum permeability. min This represents the minimum penetration rate.

[0254] The working principle of this invention is as follows: Figure 3 and Figure 4As shown, this method uses a non-magnetic experimental flume as the main space for simulating the physical composition of deep-water turbidity current sediments, and a nuclear magnetic resonance (NMR) scanning system as a non-destructive means of probing the sedimentary bottom. It combines the turbidity current sedimentation experiment with in-situ NMR imaging to achieve continuous three-dimensional characterization of the external morphology, internal bedding, and physical property distribution of the sedimentary body. The non-magnetic composite glass flume provides a simulated channel for turbidity current movement, diffusion, settling, and deposition, with the sedimentary bottom formed in the core observation section at the bottom of the flume. Because the flume is made of non-magnetic materials, it reduces interference from the static magnetic field, gradient magnetic field, and radio frequency signals, ensuring that the NMR response primarily originates from hydrogen nuclei in the water, the turbidity mixture, and the pore water of the sedimentary body.

[0255] Before the experiment, superconducting magnet plates were placed on the upper and lower sides of the water tank to form a stable static magnetic field covering the observation section, causing hydrogen nuclei in the observation area to undergo orientation polarization under the influence of the applied magnetic field. Then, gradient magnetic field coils X, Y, and Z were arranged around the water tank, corresponding to the length, width, and height of the tank, respectively. These were used to superimpose spatially encoded gradients onto the static magnetic field, giving NMR signals from different spatial locations distinguishable frequency, phase, or relaxation characteristics. Simultaneously, an electromagnetic pulse transmitter was placed on the side of the water tank, and a signal receiving coil was positioned near the outer wall to complete radio frequency excitation and echo reception. After the system was built, scanning timing parameters were further set, a temperature monitoring system was deployed, and a background calibration scan was performed to eliminate the influence of the non-magnetic water tank, background fluid, and system noise on subsequent data.

[0256] During the experiment, the material supply system transports the pre-prepared turbidity mixture to the upstream of the tank. Under the combined effects of gravity, inertia, and water resistance, the turbidity propagates along the bottom of the tank, undergoing particle sorting, suspended sedimentation, and bottom deposition as its velocity gradually decreases, forming... Figure 4 The sedimentary bottom type is shown. After the turbidity flow is released, the water body is usually in a disturbed and high turbidity state. At this time, the turbidity change is monitored by a high-precision turbidity meter and the body is allowed to stand still. When the turbidity drops to the preset condition, it indicates that the main suspended particles have completed sedimentation and the sediment structure tends to be stable. At this time, in-situ nuclear magnetic resonance scanning is started to avoid the violent movement of particles affecting the stability of the echo signal and the imaging accuracy.

[0257] During the scanning process, the electromagnetic pulse transmitter emits radio frequency pulses into the observation area according to a preset nuclear magnetic resonance sequence, causing the hydrogen nuclei, which have been polarized by the static magnetic field, to deviate from their equilibrium state and generate a transverse relaxation response. A triaxial gradient magnetic field is applied sequentially or in combination according to a set time sequence to perform three-dimensional orthogonal spatial encoding of the observation area. The signal receiving coil captures the spin echo signal generated during the transverse relaxation of the hydrogen nuclei during the pulse intervals and transmits it to the data acquisition system. Since transmission, reception, and spatial encoding are all completed through external devices to the water tank, the entire process does not require cutting, coring, or dissecting the sediment, thus preserving the original spatial structure, sequence continuity, and pore structure integrity of the sedimentary substrate.

[0258] The data acquisition system performs three-dimensional voxel reconstruction based on the intensity, attenuation characteristics, and spatial coding information of the echo signals, obtaining continuous internal images of the sedimentary body in the length, width, and height directions. Further, by combining T2 relaxation spectrum inversion, grain size parameter inversion, three-dimensional porosity inversion, and three-dimensional permeability estimation, the system obtains the internal bedding interfaces, pore water distribution, physical property differences, and sedimentary configuration characteristics of the sedimentary body. For multi-phase turbidity current experiments, the system cyclically executes the process of turbidity current release—static settling—turbidity assessment—in-situ scanning—three-dimensional reconstruction, and overlays the scanning results of each phase in chronological order. Through sequence dynamic stripping and multi-phase sequence overlay, the system establishes the correspondence between the sedimentary dynamic process and the final sedimentary configuration. The final output includes a three-dimensional visualization rendering, a three-dimensional geomorphological map of the sedimentary base, and quantitative evaluation results of microscopic physical properties, thus achieving a non-destructive, continuous, dynamic, and refined three-dimensional characterization of the sedimentary base in deep-water turbidite physical simulation.

[0259] Evidence related to the technical effects obtained by the embodiments of the present invention.

[0260] like Figure 5 The figure shows a comparison of the nuclear magnetic resonance T2 relaxation time distributions of different sedimentary configuration units obtained in an embodiment of the present invention. Through non-destructive in-situ scanning and inversion, the system accurately captured the differences in pore structure among different microfacies. The figure shows that the T2 spectrum of the central region of the turbidite channel (mainly coarse-grained sediments) exhibits a distinct single right-hand peak, with the peak value concentrated around 100 ms, reflecting its high-quality reservoir characteristics of high macroporosity and abundant free fluid. In contrast, the T2 spectrum of the overflow / natural dike region (mainly fine-grained sediments) is significantly shifted to the left, with a significant increase in the proportion of capillary-bound fluid to the left of the T2 cutoff value (33 ms). This result demonstrates that the present invention can achieve accurate and non-destructive quantitative characterization of the internal millimeter-scale voxel pore structure without destroying the original sedimentary bottom type of the physically simulated flume.

[0261] like Figure 6The figure shows a quantitative evaluation of porosity, permeability, and newly added sedimentary volume after five depositional phases (T1-T5) in an embodiment of the present invention. Utilizing the unique intermittent nuclear magnetic resonance dynamic acquisition and Boolean difference set dynamic stripping technology of this invention, the spatiotemporal evolution process was systematically and completely recorded. With the increase in injection phases, influenced by the shear force of the water flow from the previous bottom micro-topography, the newly added sedimentary volume in a single phase gradually increases. Simultaneously, due to the compaction of sediments from multiple phases and the mixing and filling of fine-grained materials, the average porosity and average permeability of the sedimentary layer both show a decreasing evolutionary pattern in each phase. This figure intuitively verifies the core advantages of this invention in true four-dimensional dynamic monitoring, successfully filling the technical gap that traditional endpoint destructive experiments cannot capture intermediate dynamic evolution processes.

[0262] like Figure 7 The figure shows a comparison of the heterogeneity parameters between the traditional two-dimensional slice discrete sampling method and the three-dimensional voxel model full-field evaluation method of this invention under the same physically simulated deposition substrate. As can be seen from the figure, the coefficient of variation (C) calculated based on the three-dimensional voxel model of this invention containing N=1000 random sampling points is significantly higher. v ), permeability gradient (R) and breakthrough coefficient (T) k The results are significantly higher than those of traditional two-dimensional slicing methods. This comparison shows that traditional destructive slicing methods, due to the severe loss of spatial information, are prone to missing extreme anomalies with extremely high or low permeability within the sedimentary body, thus underestimating the degree of reservoir heterogeneity. In contrast, the full-field non-destructive three-dimensional physical property mapping provided by this invention can faithfully reproduce the true strong heterogeneity within the sedimentary body, providing a more accurate parameter verification benchmark for numerical simulation of deepwater reservoirs.

[0263] like Figure 8 As shown, this is a collaborative registration diagram of the sedimentary bottom three-dimensional porosity volume mapping and digital geomorphological simulation output by an embodiment of the present invention. The three-dimensional porosity volume mapping cloud map on the left intuitively demonstrates the spatial heterogeneity of the internal properties of the sedimentary body. The warm-toned areas clearly delineate the meandering and interconnected morphology of the high-porosity channel center, while the cool-toned areas correspond to the low-porosity muddy overflow area. On the right, the digital elevation model (DEM) based on the same spatial coordinate system finely depicts the surface scour channels (channel incision) and sedimentary ripples, among other geomorphological features. This achievement successfully realizes the perfect nesting and collaborative interpretation of macroscopic three-dimensional geomorphology and microscopic three-dimensional properties within the same physical space, completely solving the problem of fragmented understanding caused by differences in scale and principles between different observation techniques in the past.

[0264] In the description of this invention, unless otherwise stated, "multiple" means two or more; the terms "upper," "lower," "left," "right," "inner," "outer," "front end," "rear end," "head," and "tail," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0265] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A three-dimensional fine characterization method for sedimentary bottom types in deep-water turbidite physics simulation based on nuclear magnetic resonance, characterized in that, Includes the following steps: Step 1: Construct a non-magnetic experimental water tank and a nuclear magnetic resonance scanning system, so that the core observation section of the non-magnetic experimental water tank is within the combined influence of the static magnetic field, the triaxial gradient magnetic field, the radio frequency excitation field, and the signal receiving area. Step 2: Inject background fluid into the non-magnetic experimental water tank, perform background calibration scanning on the background fluid to obtain the NMR signal reference under pure water background, and set the turbidity release parameters, turbidity trigger parameters and NMR scan timing parameters. Step 3: Release turbidity mixture upstream of the non-magnetic experimental water tank through the material supply system, monitor the turbidity changes of the water body using the turbidity monitoring system, record the arrival time of turbidity when the turbidity exceeds the frontal detection threshold, and record the sedimentation stabilization time after the turbidity is below the background steady-state threshold and remains stable for a preset time. Step 4: At the moment of sedimentation stabilization, in-situ nuclear magnetic resonance scanning is initiated. Hydrogen nuclei in the sedimentation system are excited by radio frequency pulses, and the observation area is spatially encoded in three dimensions by a triaxial gradient magnetic field. Spin echo signals are collected by a signal receiving coil to obtain the raw nuclear magnetic signal dataset at the corresponding moment. Step 5: Using the original NMR signal dataset as input, perform three-dimensional voxel reconstruction to obtain the three-dimensional voxel matrix at the corresponding time point; Step 6: Using the echo train of each voxel in the three-dimensional voxel matrix as input, and according to the relaxation relationship of the echo amplitude decaying with the acquisition time, output the lateral relaxation time distribution under the constraint of non-negative amplitude of the relaxation components. Step 7: Using the transverse relaxation time distribution, the NMR signal reference under pure water background, and the overall porosity scale as inputs, output the porosity voxel field; using the logarithmic mean relaxation time corresponding to the transverse relaxation time distribution as inputs, output the grain size parameter voxel field according to the empirical mapping relationship between the logarithmic mean relaxation time and the median grain size of the sediment. Based on the transverse relaxation time distribution, the free fluid index and the bound fluid index are divided, and combined with the porosity voxel field, the permeability voxel field is output according to the empirical relationship of permeability. Step 8: Perform multiple cyclic acquisitions in the order of turbidity release, turbidity judgment, in-situ scanning and three-dimensional voxel reconstruction. Use the three-dimensional voxel matrix of adjacent phases as input to identify voxel differences, output single-phase sedimentary sequences, and spatially superimpose each single-phase sedimentary sequence in chronological order to obtain a four-dimensional configuration model of the deep-water turbidite sedimentary bottom. Step 9: Using the four-dimensional configuration model, porosity voxel field, particle size parameter voxel field, and permeability voxel field as inputs, output three-dimensional visualization results, three-dimensional geomorphological results of sedimentary bottom type, and quantitative evaluation results of microphysical properties.

2. The method for three-dimensional fine characterization of sedimentary bottom type based on nuclear magnetic resonance physical simulation of deep-water turbidite deposits as described in claim 1, characterized in that, The construction of the non-magnetic experimental water tank and nuclear magnetic resonance scanning system includes: A non-magnetic deposition simulation water tank with a length of 2m, a width of 1m, and a height of 0.5m was constructed. The non-magnetic deposition simulation water tank was formed by splicing non-magnetic glass composite material with a thickness of 3cm, and the joints were sealed with modified epoxy resin. Two planar superconducting magnet plates are arranged in parallel on the upper and lower sides of the core observation section of the non-magnetic sedimentation simulation tank to create a static magnetic field environment with a central magnetic field strength of 0.5T and a magnetic field uniformity better than 50ppm in the core observation section. A triaxial gradient magnetic field coil is arranged outside the non-magnetic sedimentation simulation tank. The triaxial gradient magnetic field coils correspond to the X, Y and Z directions respectively, and the maximum linear gradient field strength is 20 mT per meter. A radio frequency pulse transmitter was configured in the observation section of the non-magnetic sedimentation simulation tank to match the radio frequency center frequency with the Larmor resonance frequency of the hydrogen nucleus under a static magnetic field of 0.5T. A signal receiving coil is placed close to the outer wall of a non-magnetic deposition simulation tank so that the signal receiving coil receives the spin echo signal generated by the transverse relaxation of hydrogen nuclei during the interval of radio frequency pulses. A 200L slurry pump was connected to the upstream of the non-magnetic deposition simulation tank, and a turbid mixture was delivered to the tank through a wear-resistant hose.

3. The method for three-dimensional fine characterization of sedimentary bottom type based on nuclear magnetic resonance physical simulation of deep-water turbidite deposits as described in claim 1, characterized in that, The background calibration scan includes: Low-mineralized water to a depth of 400 mm was injected into the non-magnetic experimental water tank as background fluid. The background fluid is pre-scanned, and the free induction attenuation signal is acquired to determine the longitudinal relaxation time and transverse relaxation time of the background fluid. The NMR signal amplitude of the background fluid was calibrated as a pure water signal reference. The pure water signal reference is input into the porosity inversion stage to make the initial amplitude of each voxel echo proportional to the pure water signal reference, and outputs the porosity value of the corresponding spatial location under the overall porosity scale constraint.

4. The method for three-dimensional fine characterization of sedimentary bottom type based on nuclear magnetic resonance physical simulation of deep-water turbidite deposits as described in claim 1, characterized in that, The release and triggering control of the turbid mixture includes: Add 100-mesh quartz sand, 125-mesh calcite and 500-mesh kaolin to the mixing chamber and prepare the solid phase composition according to the following mass fractions: quartz sand 15%, calcite 30% and kaolin 55%. A turbid mixture with a volume concentration of 25% and an allowable deviation of 2% was prepared by stirring continuously at a stirring speed of 300 rpm for 20 min. A turbid mixture with a single volume of 50L is released upstream of the non-magnetic experimental water tank at a flow rate of 5L per second using a slurry pump. Turbidity changes were monitored using a turbidimeter with a range of 0 to 4000 NTU. The arrival time of turbidity flow was recorded when the turbidity rose by more than 50 NTU, and the deposition stabilization time was recorded after the turbidity dropped to less than 60 NTU and remained stable for more than 5 minutes. The deposition stabilization time was used as the trigger time for in-situ nuclear magnetic resonance scanning.

5. The method for three-dimensional fine characterization of sedimentary bottom type based on nuclear magnetic resonance physical simulation of deep-water turbidite deposits as described in claim 1, characterized in that, The in-situ nuclear magnetic resonance scanning includes: A layer-selective gradient is applied along the Z-direction when a 90-degree radio frequency pulse is emitted to define the target imaging layer. A phase encoding gradient is applied in the Y direction at the leading edge of spin echo generation, so that hydrogen nuclei with different Y coordinates can obtain different phase information; During the spin echo acquisition window, a frequency coding gradient is applied along the X direction so that hydrogen nuclei with different X coordinates obtain different frequency information. Using the static magnetic field strength, the gradient magnetic field strength in the X direction, the gradient magnetic field strength in the Y direction, the gradient magnetic field strength in the Z direction, and the spatial coordinates as inputs, according to the relationship between the hydrogen nucleus Larmor resonance frequency and the magnetic field strength and spatial position, a spin echo signal carrying three-dimensional spatial position information is output, and this spin echo signal is input into the three-dimensional voxel reconstruction stage.

6. The method for three-dimensional fine characterization of sedimentary bottom type based on nuclear magnetic resonance physical simulation of deep-water turbidite deposits as described in claim 1, characterized in that, The three-dimensional voxel reconstruction, multi-scale property inversion, and four-dimensional configuration analysis include: The raw NMR signal datasets from each period are transformed from the three-dimensional frequency domain to the spatial domain, and the corresponding three-dimensional voxel matrix is ​​output. Each voxel corresponds to a physical space with a side length of 1.95 mm. Using the echo trains within each voxel as input, the transverse relaxation time distribution is output under the non-negative constraint of the relaxation spectrum amplitude by employing singular value decomposition combined with non-negative least squares solution. The initial amplitude of the echo is obtained by using the signal integral of the transverse relaxation time distribution as input, and the porosity voxel field is output by combining the pure water signal reference and the overall porosity scale. The logarithmic mean relaxation time is calculated using the transverse relaxation time distribution as input, and the particle size parameter voxel field is output according to the empirical proportionality coefficient and empirical exponential relationship between the logarithmic mean relaxation time and the median particle size. Using the porosity voxel field, the free fluid index and the bound fluid index obtained by dividing the transverse relaxation time cutoff value of 33ms as inputs, the permeability voxel field is output according to the Timur-Coates empirical relationship. Boolean difference operations are performed on the three-dimensional voxel matrices of adjacent phases as input to output the new single-phase depositional sequence within the corresponding time interval. The single-phase depositional sequences are then spatially aligned and superimposed in chronological order to output a four-dimensional configuration model.

7. A three-dimensional fine characterization system for deep-water turbidite physical simulation of sedimentary bottom types based on nuclear magnetic resonance, implementing the method as described in any one of claims 1-6, characterized in that, It includes a non-magnetic sedimentation simulation tank, a static magnetic field generator, a triaxial gradient magnetic field coil, a radio frequency pulse transmitter, a signal receiving coil, a material supply system, a turbidity monitoring system, a data acquisition system, and a computer; The non-magnetic sedimentation simulation tank is used to contain background fluid and turbidity mixture, and provides an observation space for turbidity movement, settling and deposition; The static magnetic field generating device is set on the upper and lower sides of the core observation section of the non-magnetic deposition simulation tank, and is used to form a static magnetic field in the core observation section to cause hydrogen nuclei in the observation area to undergo orientation polarization. The triaxial gradient magnetic field coils are arranged around the non-magnetic deposition simulation tank to generate spatially encoded gradients in the X, Y, and Z directions based on the static magnetic field. The radio frequency pulse transmitter is used to transmit radio frequency pulses that match the Larmor resonance frequency of the hydrogen nucleus to the core observation segment; The signal receiving coil is located near the outer wall of the non-magnetic deposition simulation tank and is used to receive the spin echo signal generated by the transverse relaxation of hydrogen nuclei during the interval of radio frequency pulses. The material supply system is connected upstream of the non-magnetic deposition simulation tank and is used to release turbidity mixture into the non-magnetic deposition simulation tank. The turbidity monitoring system is used to monitor the turbidity of the water in the non-magnetic sedimentation simulation tank and output turbidity arrival signal and sedimentation stability signal to the data acquisition system. The data acquisition system is used to trigger the coordinated operation of the radio frequency pulse transmitter, the triaxial gradient magnetic field coil, and the signal receiving coil based on the deposition stability signal, and to acquire the raw NMR signal dataset. The computer is used to take the raw NMR signal dataset as input and output three-dimensional voxel matrix, porosity voxel field, particle size parameter voxel field, permeability voxel field, four-dimensional configuration model, three-dimensional visualization results, sedimentary bottom three-dimensional geomorphological results, and quantitative evaluation results of microphysical properties.

8. The three-dimensional fine characterization system for deep-water turbidite physical simulation of sedimentary bottom type based on nuclear magnetic resonance as described in claim 7, characterized in that, The non-magnetic deposition simulation tank is a tank with a length of 2m, a width of 1m, and a height of 0.5m. The tank is formed by splicing non-magnetic glass composite material with a thickness of 3cm, and the joints are sealed with modified epoxy resin. The material supply system includes a slurry pump with a volume of 200L and a wear-resistant hose. The slurry pump is connected to the upstream of the non-magnetic deposition simulation tank through the wear-resistant hose.

9. The three-dimensional fine characterization system for deep-water turbidite physical simulation of sedimentary bottom type based on nuclear magnetic resonance as described in claim 7, characterized in that, The static magnetic field generating device includes two planar superconducting magnet plates, which are respectively arranged in parallel on the upper and lower sides of the core observation section of the non-magnetic deposition simulation tank. The static magnetic field generating device forms a static magnetic field environment with a central magnetic field strength of 0.5T and a magnetic field uniformity better than 50ppm in the core observation section. The maximum linear gradient field strength of the triaxial gradient magnetic field coil is 20mT per meter.

10. The three-dimensional fine characterization system for deep-water turbidite physical simulation of sedimentary bottom types based on nuclear magnetic resonance as described in claim 7, characterized in that, The computer includes a three-dimensional voxel reconstruction unit, a relaxation spectrum inversion unit, a physical property parameter inversion unit, a sequence analysis unit, and a result output unit. The three-dimensional voxel reconstruction unit is used to take the original nuclear magnetic resonance signal dataset as input, and output a three-dimensional voxel matrix after three-dimensional frequency domain to spatial domain conversion. The relaxation spectrum inversion unit is used to output the lateral relaxation time distribution under the non-negative constraint of relaxation spectrum amplitude, taking the echo train of each voxel in the three-dimensional voxel matrix as input. The physical property parameter inversion unit is used to take the transverse relaxation time distribution, pure water signal reference, overall porosity scale, free fluid index and bound fluid index as inputs, and output porosity voxel field, particle size parameter voxel field and permeability voxel field according to porosity scale relationship, particle size empirical mapping relationship and permeability empirical relationship. The sequence analysis unit is used to take the three-dimensional voxel matrix of adjacent phases as input, identify the single-phase depositional sequence through voxel differences, and spatially superimpose the single-phase depositional sequences in chronological order to output a four-dimensional configuration model. The output unit is used to take a four-dimensional configuration model, a porosity voxel field, and a permeability voxel field as inputs, and output three-dimensional visualization results, three-dimensional geomorphological results of sedimentary bottom type, and quantitative evaluation results of microphysical properties.

Images