Seismic inversion velocity field reconstruction method and system based on drilling geochemical prior constraints

Abstract

The present disclosure provides a seismic inversion velocity field reconstruction method and system based on drilling geochemical prior constraints, relating to the field of seismic full waveform inversion technology, comprising: obtaining drilling noise passive source seismic data, drilling data and geochemical data; converting the drilling data and geochemical data into borehole velocity prior using a conversion model; based on the borehole velocity prior, constructing distance attenuation, anisotropy correlation, stratigraphic trend consistency and entropy value modulation spatial extrapolation weight function to generate a spatial velocity prior field and a constraint weight; constructing a borehole passive source seismic full waveform inversion joint loss function based on the spatial extrapolation weight function, performing gradient solving and optimization iteration on the joint loss function, and outputting an iterative optimization result; and obtaining a velocity field structure, interface position, anomaly body distribution, risk section and recommended supplementary detection area based on the iterative optimization result. The present disclosure improves the stability of the velocity field structure, hard and soft rock interface and fault fracture zone positioning.

Description

Technical Field

[0001] This disclosure relates to the field of seismic full-waveform inversion technology, specifically to a seismic inversion velocity field reconstruction method and system based on geochemical prior constraints during drilling. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] Seismic inversion is one of the core technologies for tunnel advanced geological prediction. By analyzing the propagation patterns of seismic waves within the rock mass, it infers the velocity structure, interface locations, and physical property distribution of the medium ahead of the tunnel face, thereby identifying adverse geological formations such as fault fracture zones, abrupt transitions between soft and hard rocks, water-rich weakened zones, karst, alteration zones, and areas with dense tectonic fractures. Failure to accurately obtain the velocity structure and stratigraphic interface locations ahead during advanced drilling and construction can easily lead to risks such as collapses, water and mud inrushes, over- or under-excavation, support failure, and construction delays.

[0004] Active-source seismic analysis is insufficient to meet the demands for continuous advance detection under rapid construction conditions. Borehole passive-source seismic analysis offers an important solution. This method utilizes the continuous vibrations generated during advance drilling, including drill bit rock breaking, drill rod vibration, and borehole wall coupling, as natural sources. However, the source timing and source function of passive sources are uncertain, and relying solely on full-waveform inversion of passive-source seismic analysis easily leads to multiple solutions and local minima, resulting in insufficient stability in velocity-structure inversion.

[0005] Drilling parameters and geochemical parameters are engineering data that can be continuously or quasi-continuously acquired during advanced drilling. They have the potential to constrain the inversion of passive source seismic velocity and can provide prior information for velocity field estimation from two dimensions: rock mechanical response and material composition.

[0006] However, drilling and geochemical parameters are mainly obtained along the borehole trajectory and are considered local observations, which cannot be simply used as hard constraints for the entire area. If the velocity or lithology results identified near the borehole are directly extrapolated to the entire exploration area, it may cause incorrect constraints in areas with significant lateral stratigraphic variations or complex structures, thereby reducing the reliability of the inversion results. Summary of the Invention

[0007] To address the aforementioned issues, this disclosure proposes a seismic inversion velocity field reconstruction method and system based on drilling geochemical prior constraints. It proposes a spatial extrapolation mechanism that can represent "strong constraints near the borehole and weak constraints at a distance" and "strong constraints in low-entropy regions and weak constraints in high-entropy regions," and incorporates this mechanism into the full-waveform inversion loss function in mathematical form. This deeply couples drilling parameters, geochemical parameters, and passive source seismic full-waveform inversion, avoiding the inversion ambiguity caused by relying solely on passive source seismic data, and also avoiding erroneous extrapolation caused by using drilling parameters or geochemical parameters as global hard constraints. This provides a basis for the advanced identification of fault fracture zones, abrupt changes in soft and hard rocks, water-rich weakening zones, and lithological change zones.

[0008] According to some embodiments, the present disclosure adopts the following technical solutions: Seismic inversion velocity field reconstruction methods based on drilling-while-drilling geochemical prior constraints include: Acquire seismic data, drilling data, and geochemical data of passive sources of drilling noise during tunnel drilling, and preprocess them; The preprocessed drilling and geochemical data are converted into borehole velocity priors using a transformation model; Based on the borehole velocity prior, a spatial extrapolation weight function is constructed based on distance attenuation, anisotropic correlation, formation trend consistency, and entropy modulation to generate a spatial velocity prior field and constraint weights. A joint loss function for full waveform inversion of borehole passive source seismic data is constructed based on a spatial extrapolation weight function. The joint loss function is then subjected to gradient solving and optimization iteration to output the iterative optimization results. Based on the iterative optimization results, the velocity field structure, interface location, anomaly distribution, risk sections, and suggested supplementary detection areas are obtained.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: The seismic inversion velocity field reconstruction system based on drilling-while-drilling prior constraints includes: The data acquisition module is used to acquire seismic data, drilling data, and geochemical data of passive sources of drilling noise during the tunnel drilling process, and to preprocess them. The velocity prior conversion module is used to convert preprocessed drilling data and geochemical data into borehole velocity priors using a conversion model; The spatial extrapolation function construction module is used to construct a spatial extrapolation weight function based on the borehole velocity prior, which is influenced by distance decay, anisotropic correlation, formation trend consistency, and entropy modulation, and generates a spatial velocity prior field and constraint weights. The full waveform inversion module is used to construct a joint loss function for full waveform inversion of borehole passive source seismic based on a spatial extrapolation weight function, perform gradient solving and optimization iteration on the joint loss function, and output the iterative optimization result. The uncertainty assessment and result output module is used to obtain the velocity field structure, interface location, anomaly distribution, risk sections, and suggested supplementary detection areas based on the iterative optimization results.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: A computer program product includes a computer program that, when executed by a processor, implements the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints.

[0011] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints.

[0012] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints.

[0013] Compared with the prior art, the beneficial effects of this disclosure are as follows: This disclosed seismic inversion velocity field reconstruction method based on drilling-while-drilling geochemical prior constraints constructs a spatial extrapolation function that simultaneously conforms to geostatistical spatial correlation, stratigraphic spatial trend, and interpretation uncertainty modulation. This results in strong constraints near the borehole and weak constraints further away; strong constraints in low-entropy, high-confidence regions and weak constraints in high-entropy, low-confidence regions; and the ability to set anisotropic correlation lengths based on bedding, structural strike, or tunnel axis direction. The prior velocity mean, prior variance, and constraint weights generated by this spatial extrapolation function are introduced into the borehole passive source seismic full-waveform inversion loss function to form a joint objective function consisting of a data fitting term, a drilling-geochemical velocity prior constraint term, a geostatistical covariance constraint term, and a spatial trend regularization term. By iteratively optimizing the joint loss function, the velocity field structure, velocity anomaly location, stratigraphic interface location, and spatial uncertainty distribution in front of the tunnel face or borehole are finally obtained. This allows for the simultaneous use of seismic waveform information and direct borehole disturbance information during the passive source seismic full waveform inversion process, avoiding the inversion ambiguity caused by relying solely on passive source seismic data, and also avoiding erroneous extrapolation caused by using drilling parameters or geochemical parameters as global hard constraints. This provides a basis for the advanced identification of fault fracture zones, abrupt changes in soft and hard rocks, water-rich weakening zones, and lithological change zones.

[0014] The seismic inversion velocity field reconstruction method based on drilling geochemical prior constraints disclosed herein unifies drilling parameters, geochemical parameters, and passive source seismic full waveform inversion into a joint loss function. This allows drilling and geochemical information to be directly incorporated into the velocity field inversion, rather than relying on manual interpretation. By establishing an interpretable mapping from drilling-geochemical data to velocity priors through the rock physics relationship between rock strength, lithology, and wave velocity, the physical rationality of the constraint terms is improved.

[0015] The seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints disclosed herein constructs a spatial extrapolation function that makes the constraints strong near the borehole and weak far from the borehole, and makes the constraints strong in low-entropy regions and weak in high-entropy regions, thus avoiding the hard extrapolation of local borehole information to the entire domain.

[0016] The seismic inversion velocity field reconstruction method based on drilling geochemical prior constraints disclosed herein reduces the ambiguity and local minima risk of passive source seismic FWI by using a joint loss function, thereby improving the stability of velocity field structure, soft and hard rock interfaces and fault fracture zone localization. Attached Figure Description

[0017] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.

[0018] Figure 1 This is a flowchart illustrating the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints, according to an embodiment of this disclosure. Detailed Implementation

[0019] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.

[0020] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.

[0021] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this disclosure. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0022] Example 1 One embodiment of this disclosure provides a seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints, the method steps of which include: Step 1: Acquire seismic data, drilling data, and geochemical data of passive sources of drilling noise during the tunnel drilling process, and preprocess them; Step 2: Convert the preprocessed drilling data and geochemical data into borehole velocity priors using a transformation model; Step 3: Based on the borehole velocity prior, construct a spatial extrapolation weight function for distance attenuation, anisotropic correlation, formation trend consistency, and entropy modulation to generate a spatial velocity prior field and constraint weights; Step 4: Construct a joint loss function for full waveform inversion of borehole passive source seismic waves based on the spatial extrapolation weight function, perform gradient solving and optimization iteration on the joint loss function, and output the iterative optimization result; Step 5: Based on the iterative optimization results, obtain the velocity field structure, interface location, anomaly distribution, risk sections, and suggested supplementary detection areas.

[0023] As one embodiment, this disclosure presents a seismic inversion velocity field reconstruction method based on drilling-while-drilling geochemical prior constraints. First, it utilizes the correlation between drilling parameters and geochemical parameters and rock strength and lithology. Then, it further constructs a velocity prior field through the rock physical relationship between rock strength, lithology, and wave velocity. Unlike existing practices that directly extrapolate drilling or geochemical information from the borehole to the entire exploration area, this disclosure constructs a spatial extrapolation function that simultaneously conforms to geostatistical spatial correlation, stratigraphic spatial trends, and interpretation uncertainty modulation. This function provides strong constraints near the borehole and weak constraints further away; it provides strong constraints in low-entropy, high-confidence regions and weak constraints in high-entropy, low-confidence regions; and it allows for setting anisotropic correlation lengths based on bedding, structural strike, or tunnel axis direction. The prior mean, prior variance, and constraint weights of the velocity generated by the spatial extrapolation function are introduced into the borehole passive source seismic full waveform inversion loss function to form a joint objective function composed of data fitting term, drilling-geochemical velocity prior constraint term, geostatistical covariance constraint term, and spatial trend regularization term. The joint loss function is iteratively optimized to finally obtain the velocity field structure, velocity anomaly location, formation interface location, and spatial uncertainty distribution at the tunnel face or in front of the borehole.

[0024] In this embodiment, the velocity field reconstruction ahead of the tunnel face during drill-and-blast tunnel construction is used as the application scenario. The tunnel is excavated using the drill-and-blast method, and there are potential adverse geological risks ahead of the tunnel face, such as lithological changes, weak interlayers, fault fracture zones, water-rich structures, or abrupt changes in rock mass integrity. To obtain lithological information ahead of the tunnel face without additional construction time, the drilling disturbance during advanced geological drilling is used as a passive source. Combined with drilling parameters and borehole cuttings geochemical parameters, fault fracture zones, abrupt changes in soft and hard rock, water-rich weakening zones, and lithological change zones within a certain range ahead of the tunnel face are identified in advance. The specific implementation process of this method is as follows: Step 1: Acquire seismic data, drilling data, and geochemical data of passive sources of drilling noise during the tunnel drilling process, and preprocess them.

[0025] Specifically, during the advanced geological drilling process of drill-and-blast tunnels, a borehole passive source seismic observation system, a drilling parameter acquisition system, and a geochemical parameter acquisition system are deployed. A seismic sensor array is arranged behind or on the sidewall of the drill-and-blast tunnel face. Sensors can be velocity detectors, accelerometers, or fiber optic vibration sensors. The sensor deployment should cover the borehole axis and the tunnel's lateral position to record drill bit rock-breaking vibration, drill rod vibration, and surrounding rock response during the advanced drilling process. The drilling parameter acquisition system is synchronously connected to the drilling rig control system. Geochemical parameters can be obtained through drill cuttings, flushing fluid, return water, or portable elemental analysis equipment. Acquired data includes passive source seismic data of drilling noise, drilling speed, torque, thrust, rotational speed, impact pressure, flushing pressure, return water volume, borehole depth, elemental content, ion concentration, mineral composition, or other geochemical parameters. All data must be synchronized with timestamps and borehole depth coordinates.

[0026] For continuous passive source seismic data, the following processes are performed: mean removal, trend removal, outlier removal, bandpass filtering, spectral whitening, 1-bit normalization, and short-window segmentation. Reference sensors or reference time windows are selected for cross-correlation, and the data are superimposed to form a virtual shot gather. Travel time, separation, envelope, phase, or full waveform residual data domains are extracted as needed. For drilling parameters, depth registration, sliding window statistics, differential feature extraction, and drilling specific energy calculation are performed. For geochemical parameters, element normalization, outlier processing, and principal component extraction are performed.

[0027] During advanced drilling, the drill bit breaking rock and the drill pipe vibration generate continuous passive source signals. Let the first... i Each detector record is as follows Select a reference detector Virtual gun sets can be constructed through cross-correlation:

[0028] in, This indicates that the first detector or reference detector is used as the virtual source, and the second detector is used as the virtual source. i Each detector receives the cross-correlation response or virtual shot collection channel from the receiving channel; Indicates the first i Each detector lags behind the reference channel. τ The subsequent passive source vibration signal; t Indicates the continuous recording time; τ This represents the cross-correlation delay or virtual gun set time axis.

[0029] Superimposing the cross-correlation results of multiple short time windows can improve the signal-to-noise ratio of the stable propagation response.

[0030] in, Indicates the first i The stable cross-correlation response between the Dao and the reference Dao after multiple short time windows are superimposed; Indicates the number of short time windows involved in the overlay; n Indicates the short time window number; Indicates the first n The cross-correlation response calculated within a short time window; τ This indicates the cross-correlation delay.

[0031] The virtual shot collection or the waveform recorded after source estimation is denoted as... Given a velocity model m The main parameters are V s ( x , z ), through forward calculus F ( m Generate simulated passive source seismic records:

[0032] in, s This represents the equivalent source function for a passive source. If the source function is unknown, source estimation can be performed in each iteration, or a weakly source-dependent data domain such as a cross-correlation virtual shot gather or envelope residual can be used.

[0033] Let the first k The feature vector is composed of drilling parameters and geochemical parameters at each drilling depth or borehole location:

[0034] in, For drilling speed, M For torque, For propulsion, n For rotational speed, To impact pressure, For flushing pressure, For return water volume, To drill down to the specific energy, Geographic indicators such as element content, ion concentration, or mineral composition. This represents the change or local difference in drilling speed within adjacent depth windows. This represents the change or local difference in torque within adjacent depth windows.

[0035] Furthermore, the drilling specific energy can be expressed as:

[0036] in,A This indicates the effective rock-breaking area of ​​the drill bit, cutterhead, or bottom of the hole; if the borehole is approximately circular, A can be calculated based on the drill bit diameter. D b calculate:

[0037] in, D b This indicates the drill bit diameter. If a non-circular cutter or a combination of multiple cutters is used, A can be taken as the equivalent contact area or equivalent bottom area of ​​the actual rock breaking.

[0038] Step 2: Convert the preprocessed drilling data and geochemical data into borehole velocity priors using a transformation model.

[0039] Specifically, a rock strength estimation model and a lithology identification model are constructed:

[0040]

[0041] in, For the first k Rock strength estimates at each location; This represents a rock strength estimation model; Indicates the first k Drilling-geochemical joint feature vectors at each location; This indicates the error in the intensity estimation model; For the first k The position belongs to the first j The probability of lithological similarity; Indicates the first k A random variable representing the lithological category at each location; j Indicates the lithological category number; This indicates the total number of lithology categories.

[0042] The uncertainty in lithology identification is represented by normalized information entropy:

[0043] The lower the value, the more certain the lithology identification result; The higher the value, the greater the uncertainty in lithological classification.

[0044] Using exposed strata, borehole logging, or laboratory rock strength test data to train rock strength models and lithology identification models, outputting... , and .

[0045] Furthermore, based on the rock physics relationship between rock strength, lithology, and wave velocity, a priori velocity at the borehole can be established, which can be achieved using empirical relationships, lithology-specific regression models, or machine learning regression models.

[0046] Drilling parameters and geochemical parameters are successively converted into velocity priors through rock strength model, lithology classification model and rock physical velocity conversion model. The rock strength model can be an empirical model, a machine learning model or a combination of both. The lithology classification model can be a random forest, support vector machine, extreme gradient boosting, neural network or ensemble learning model.

[0047] In one implementation, the logarithmic velocity can be expressed as:

[0048] in, For the velocity prior at the drilling point, , , For parameters related to lithology, For the localization parameter vector, This represents the model error.

[0049] Considering the uncertainty of lithological probability, the velocities under different lithological conditions can be probabilistically weighted to obtain the prior mean velocity at the borehole:

[0050] The prior variance of borehole velocity can be composed of lithological uncertainty, strength estimation error, geochemical measurement error, and velocity conversion model error.

[0051] This formula indicates that if the velocity varies significantly among different lithologies, or if the uncertainty in lithology identification is high, the prior variance of the borehole velocity will increase.

[0052] Step 3: Based on the borehole velocity prior, construct a spatial extrapolation weight function for distance attenuation, anisotropic correlation, formation trend consistency, and entropy modulation, and generate a spatial velocity prior field and constraint weights.

[0053] To avoid rigidly extrapolating local borehole priors to the entire domain, this invention constructs a spatial extrapolation weighting function. The spatial extrapolation function includes a distance attenuation term, an anisotropic spatial correlation term, a formation trend consistency term, a drilling / geochemical identification confidence term, and an information entropy modulation term, which makes the constraints strong near the borehole and weak far from the borehole, and makes the constraints strong in low-entropy regions and weak in high-entropy regions.

[0054] Let the target cell location be u =( x , z ), No. kThe borehole constraint positions are: Local coordinate components are defined based on bedding, structural strike, or tunnel axis direction. and And establish anisotropic space-related kernels:

[0055] in, and These are the correlation lengths along the geological trend direction and the vertical trend direction, respectively. Generally, the correlation length is larger along the direction of bedding or structural extension, and smaller vertically.

[0056] To ensure that the extrapolation conforms to the spatial trend of the stratigraphy, a trend consistency term is introduced. Let... ζ ( u Let be the stratigraphic trend coordinates estimated from geological logging, face sketches, existing profiles, or initial velocity models, then:

[0057] By integrating drilling / geochemical identification confidence, spatial correlation, trend consistency, and information entropy modulation, an unnormalized extrapolation weight is constructed:

[0058] in, For sample quality factor, This is the lithology identification entropy value. This represents the data quality factor for drilling and geochemical parameters. A higher weight is assigned to samples that are reliable, have low entropy, are close together, and exhibit consistent trends; conversely, a lower weight is assigned to samples that are uncertain, far apart, or have inconsistent trends.

[0059] Furthermore, the normalized extrapolation weights are:

[0060] This yields the prior mean field of spatial velocity:

[0061] The spatial velocity prior variance field is defined as:

[0062] The first term represents the uncertainty of the drilling speed prior itself and the discreteness between different drilling priors; the second term represents the compensation for the extrapolation uncertainty when far from the constraint region.

[0063] The spatial extrapolation weights are further transformed into the prior constraint strengths in the full waveform inversion:

[0064] in, For spatial lithology probability entropy, This represents the local data quality factor for passive source earthquakes. The larger the value, the stronger the constraint of the drilling-geometry prior on the velocity inversion at that location.

[0065] Step 4: Construct a joint loss function for full waveform inversion of borehole passive source seismic data based on the spatial extrapolation weight function, perform gradient solving and optimization iteration on the joint loss function, and output the iterative optimization result.

[0066] From a Bayesian perspective, the target velocity model m Given passive source seismic data Drilling parameters Geochemical parameters and geological trend information The posterior distribution after that is:

[0067] If the observation error follows a Gaussian distribution, then the likelihood of passive source seismic data can be written as:

[0068] The probabilistic form of the prior constraint on drilling-geochemical velocity is:

[0069] The prior spatial covariance in geostatistics can be expressed as:

[0070] in This is the spatial covariance matrix constructed from anisotropic variation functions or covariance functions. Its typical form is:

[0071] Using μV(u) as the initial velocity field or a part of the initial velocity field, a joint loss function is constructed. Taking the negative logarithm of the posterior probability and ignoring the constant term, the joint loss function for full waveform inversion can be obtained:

[0072]

[0073]

[0074]

[0075]

[0076]

[0077] in, Constrained model fitting of passive source seismic data; To make the velocity field approximate the drilling-geometry velocity prior in the reliable region; Ensure that the velocity field conforms to geostatistical spatial correlation; Ensure that velocity changes conform to the geological trend; Suppress unreasonable high-frequency oscillations.

[0078] Furthermore, the loss function is solved for gradients and iterated for optimization. In the... n In this iteration, the velocity model is denoted as... m n The gradient of the joint loss function can be written as:

[0079] The velocity gradient of the prior constraint term for drilling-geochemical velocity is:

[0080] The gradient of the smoothing term can be written as:

[0081] Model updates can be performed using the conjugate gradient method or L-BFGS.

[0082]

[0083]

[0084] Iteration stops when the following formula is satisfied for a certain number of consecutive iterations:

[0085] Step 5: Based on the iterative optimization results, obtain the velocity field structure, interface location, anomaly distribution, risk sections, and suggested supplementary detection areas.

[0086] Based on the velocity field structure, residual convergence index, prior constraint weights, and posterior velocity uncertainty, the final velocity field is output. Velocity gradient map, spatial distribution of anomalies, location of formation interfaces, and drilling-geochemical constraint weight map. Speed ​​prior standard deviation Inversion residual plot and credibility evaluation plot. For Regions with low inversion residuals and high inversion residuals should be marked as high uncertainty regions, and it is recommended to supplement drilling, densify sensors, or add active source verification.

[0087] As one example, the rock strength model and lithology identification model are not independent interpretations of the results, but rather intermediate transformation modules in the velocity prior construction process. Their specific application process includes: The first step is to establish training samples. Using exposed borehole sections, face logging sections, core samples, or laboratory test samples as calibration samples, drilling parameters and geochemical parameters at the same depth or in the same spatial unit are registered with measured rock strength and lithology labels to form a training set.

[0088] in, Represents the training sample set; X k For the first k The joint feature vector of drilling and geochemical analysis for each sample; R c,k obs For the first k Rock strength obtained from measured or recorded inversion of individual samples; L k obs For the first k Lithological labels for each sample; N This represents the number of training samples.

[0089] The second step is to train the rock strength estimation model. X k Input rock strength estimation model f R The rock strength estimate was obtained. . Empirical regression models, random forests, gradient boosting trees, neural networks, or combined models can be used.

[0090] The third step is to train the lithology identification model. X k Input lithology identification model P L Output the probability of each lithology category. And supervised training using known lithology labels: The fourth step is field application. Drilling and geochemical parameters are collected from new boreholes or locations at new depths, and then processed using the same preprocessing method to form... X k ,Will X k Input the trained data respectively f R and P L ,get , and lithological entropy H kBy combining this with a rock physics velocity conversion model, velocity estimates under different lithological conditions are obtained, and the prior mean velocity at the borehole location is obtained by weighting the values ​​according to lithological probability. and variance : Fifth step, determine the drilling speed in advance. Prior variance And identifying uncertainty Input the spatial extrapolation function to generate the global velocity prior field, prior variance field and constraint weights, and then input the joint loss function for passive source earthquake full waveform inversion.

[0091] Example 2 One embodiment of this disclosure provides a seismic inversion velocity field reconstruction system based on drilling-while-drilling prior constraints, comprising: The data acquisition module is used to acquire seismic data, drilling data, and geochemical data of passive sources of drilling noise during the tunnel drilling process, and to preprocess them. The velocity prior conversion module is used to convert preprocessed drilling data and geochemical data into borehole velocity priors using a conversion model; The spatial extrapolation function construction module is used to construct a spatial extrapolation weight function based on the borehole velocity prior, which is influenced by distance decay, anisotropic correlation, formation trend consistency, and entropy modulation, and generates a spatial velocity prior field and constraint weights. The full waveform inversion module is used to construct a joint loss function for full waveform inversion of borehole passive source seismic based on a spatial extrapolation weight function, perform gradient solving and optimization iteration on the joint loss function, and output the iterative optimization result. The uncertainty assessment and result output module is used to obtain the velocity field structure, interface location, anomaly distribution, risk sections, and suggested supplementary detection areas based on the iterative optimization results.

[0092] Example 3 One embodiment of this disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints.

[0093] Example 4 One embodiment of this disclosure provides a non-transitory computer-readable storage medium for storing computer instructions. When executed by a processor, the computer instructions implement the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints.

[0094] Example 5 One embodiment of this disclosure provides an electronic device, including a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints.

[0095] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0096] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0097] While the specific embodiments of this disclosure have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of this disclosure. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this disclosure are still within the scope of protection of this disclosure.

Claims

1. A seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints, characterized in that, include: Acquire seismic data, drilling data, and geochemical data of passive sources of drilling noise during tunnel drilling, and preprocess them; The preprocessed drilling and geochemical data are converted into borehole velocity priors using a transformation model; Based on the borehole velocity prior, a spatial extrapolation weight function is constructed based on distance attenuation, anisotropic correlation, formation trend consistency, and entropy modulation to generate a spatial velocity prior field and constraint weights. A joint loss function for full waveform inversion of borehole passive source seismic data is constructed based on a spatial extrapolation weight function. The joint loss function is then subjected to gradient solving and optimization iteration to output the iterative optimization results. Based on the iterative optimization results, the velocity field structure, interface location, anomaly distribution, risk sections, and suggested supplementary detection areas are obtained.

2. The seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints as described in claim 1, characterized in that, The acquisition and preprocessing of passive source seismic data, drilling data, and geochemical data of drilling noise during tunnel drilling includes: Seismic sensor arrays are arranged behind or on the sidewall of the tunnel face using the drill-and-blast method. The sensor layout should cover the borehole axis and the transverse position of the tunnel to record the rock-breaking vibration of the drill bit, the vibration of the drill rod, and the response of the surrounding rock during the advance drilling process. The continuous passive source seismic data is subjected to mean removal, trend removal, outlier removal, bandpass filtering, spectral whitening, 1-bit normalization, and short window segmentation. Reference sensors or reference time windows are selected for cross-correlation and superimposed to form virtual shot sets. Drilling parameters are subjected to depth registration, sliding window statistics, differential feature extraction, and drilling specific energy calculation. Geochemical parameters are subjected to element normalization, outlier processing, and principal component extraction preprocessing.

3. The seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints as described in claim 1, characterized in that, The process of converting preprocessed drilling data and geochemical data into borehole velocity priors using a transformation model includes: The drilling parameters and geochemical parameters are combined to form a feature vector; A rock strength estimation model and a lithology identification model are constructed. The feature vectors are input into the rock strength estimation model and the lithology identification model to obtain the rock physical relationship between rock strength, lithology and wave velocity. Based on the rock physics relationship between rock strength, lithology and wave velocity, and considering the uncertainty of lithology probability, the velocities under different lithology conditions are probabilistically weighted to obtain the prior mean of the borehole velocity. The prior variance of borehole velocity is composed of lithological uncertainty, strength estimation error, geochemical measurement error, and velocity conversion model error.

4. The seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints as described in claim 1, characterized in that, The method, based on borehole velocity priors, constructs a spatial extrapolation weighting function considering distance attenuation, anisotropic correlation, formation trend consistency, and entropy modulation, generating a spatial velocity prior field and constraint weights, including: Local coordinate components are defined based on bedding, structural strike, or tunnel axis direction, and an anisotropic spatial correlation kernel is established. By integrating drilling / geochemical identification confidence, spatial correlation, trend consistency and information entropy modulation, an unnormalized extrapolation weight is constructed, thereby obtaining the spatial velocity prior mean field and the spatial velocity prior variance field. The spatial extrapolation weights are further transformed into the prior constraint strengths in the full waveform inversion.

5. The seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints as described in claim 1, characterized in that, The joint loss function for borehole passive source seismic full waveform inversion is constructed based on the spatial extrapolation weight function. The joint loss function is then subjected to gradient solving and iterative optimization to output the iterative optimization results, including: From a Bayesian perspective, obtain the posterior distribution of the target velocity model given passive source seismic data, drilling parameters, geochemical parameters, and geological trend information; Taking the negative logarithm of the posterior probability and ignoring the constant term yields the joint loss function for full waveform inversion; The model is updated using the conjugate gradient method or L-BFGS, and the iteration stops when the condition is met for a certain number of consecutive iterations.

6. The seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints as described in claim 1, characterized in that, The velocity field structure, interface location, anomaly distribution, risk zones, and suggested supplementary detection areas obtained based on the iterative optimization results include: Based on the velocity field structure, residual convergence index, prior constraint weights and velocity posterior uncertainty, the final velocity field, velocity gradient map, anomaly spatial distribution, formation interface location, drilling-geochemical constraint weight map, velocity prior standard deviation, inversion residual map and credibility evaluation map are output. Regions with low drilling-geochemical constraint weights and high inversion residuals are marked as high uncertainty regions.

7. A seismic inversion velocity field reconstruction system based on drilling-while-drilling prior constraints, characterized in that, include: The data acquisition module is used to acquire seismic data, drilling data, and geochemical data of passive sources of drilling noise during the tunnel drilling process, and to preprocess them. The velocity prior conversion module is used to convert preprocessed drilling data and geochemical data into borehole velocity priors using a conversion model; The spatial extrapolation function construction module is used to construct a spatial extrapolation weight function based on the borehole velocity prior, which is influenced by distance decay, anisotropic correlation, formation trend consistency, and entropy modulation, and generates a spatial velocity prior field and constraint weights. The full waveform inversion module is used to construct a joint loss function for full waveform inversion of borehole passive source seismic based on a spatial extrapolation weight function, perform gradient solving and optimization iteration on the joint loss function, and output the iterative optimization result. The uncertainty assessment and result output module is used to obtain the velocity field structure, interface location, anomaly distribution, risk sections, and suggested supplementary detection areas based on the iterative optimization results.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints as described in any one of claims 1-6.

9. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the seismic inversion velocity field reconstruction method based on drilling-while-drilling prior constraints as described in any one of claims 1-6.

10. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the seismic inversion velocity field reconstruction method based on drilling geochemical prior constraints as described in any one of claims 1-6.

Images