A drill bit steering control system and method based on ground sound monitoring

Abstract

The application provides a drill bit guiding control system and method based on ground sound monitoring, which comprises a ground ground sound monitoring subsystem, a ground drilling machine console, a while-drilling information transmission mechanism and a downhole guiding execution mechanism; a ground sound sensor array picks up continuous active ground sound signals generated by rock breaking of a drill bit rotation and passive ground sound signals generated by surrounding rock breaking in real time, and transmits the signals to a data processing center through a signal collection station; the data processing center classifies and identifies the ground sound signals, solves a while-drilling three-dimensional space position and inverts a geological environment, and generates a while-drilling deviation correction control instruction; the ground drilling machine console triggers a drill bit sensor according to the instruction, modulates a steering action command into a fluid pressure wave and transmits the fluid pressure wave to a downhole receiving probe; the downhole receiving probe transmits the command to the downhole guiding execution mechanism after demodulation, and drives the drill bit to adjust a drilling posture, so that the drill bit track while-drilling monitoring, deviation correction control and geological environment inversion are cooperatively closed, and the guiding response timeliness is improved.

Description

Technical Field

[0001] This invention relates to the field of geological exploration and drilling engineering technology, and in particular to a drill bit guidance control system and method based on ground acoustic monitoring. Background Technology

[0002] In mineral resource exploration and geological drilling engineering, directional drilling technology is a key means to accurately hit the target strata and improve resource extraction efficiency. In order to ensure that the borehole trajectory strictly follows the design route, it is necessary to conduct precise monitoring and dynamic guidance control of the three-dimensional spatial position of the downhole drill bit.

[0003] Currently, mainstream measurement-while-drilling (MWD) technologies primarily rely on MWD instruments installed within the drill string assembly. However, due to the severe mechanical vibrations and impacts generated by the drill bit during rock cutting, traditional measurement sensors are typically placed inside the drill collar, several meters or even tens of meters behind the cutting edge of the drill bit, to protect delicate electronic components. This structural design results in an inherent geometric blind zone between the measurement point and the actual drill bit contact point, making it impossible to directly reflect the real-time trajectory of the drill bit tip. Furthermore, existing downhole data transmission methods mostly utilize physical media such as fluid pressure waves (mud pulses), which have extremely low signal transmission rates, leading to severe intermittent lag in trajectory data acquired by the surface control console. When facing thin reservoirs, faults, and complex geological interfaces, the combination of this "physical blind zone" and "transmission lag" often prevents timely adjustments to guidance commands from the surface, causing the drill bit to deviate from the target layer, severely impacting drilling quality and operational safety.

[0004] On the other hand, while existing ground acoustic monitoring technologies have some applications in the acquisition of peripheral acoustic signals, they are mainly used in passive safety protection areas such as monitoring surrounding rock pressure or early warning of rockbursts during deep mining, and are rarely directly applied to real-time closed-loop guidance control during drilling. How to effectively utilize the acoustic characteristics during drilling and integrate large-scale surface acoustic positioning with downhole guidance tool assemblies in a closed loop remains a bottleneck that is difficult to overcome in this field.

[0005] Therefore, there is a need to provide a drill bit guidance control system and method based on ground acoustic monitoring, which can realize drill bit trajectory monitoring and active closed-loop guidance control with no blind spots and high real-time performance, and can simultaneously complete the advanced inversion prediction of the physical properties of the underground rock mass in front of the drill bit. Summary of the Invention

[0006] In view of the above-mentioned defects in the prior art, the purpose of this invention is to provide a drill bit guidance control system and method based on ground acoustic monitoring, which aims to solve the technical problems of drill bit geometric blind spots and intermittent lag in trajectory data transmission in traditional measurement while drilling, which makes closed-loop guidance control impossible during drilling.

[0007] To achieve the above objectives, in a first aspect, the present invention provides a drill bit guidance control system based on ground acoustic monitoring, comprising: The ground acoustic monitoring subsystem includes multiple acoustic sensors arranged in an array, a signal acquisition station electrically connected to each of the acoustic sensors, and a data processing center connected to the signal acquisition station. Each acoustic sensor is used to pick up continuous active acoustic signals generated by the rotation of the drill bit breaking the rock, as well as passive acoustic signals generated by the fracturing of the surrounding rock in real time. The data processing center is used to perform signal classification and identification, drilling spatial position calculation, and geological environment inversion based on the active and passive acoustic signals, and output drilling deviation correction control commands. The ground drilling rig control console is bidirectionally connected to the data processing center and is used to issue directional adjustment commands according to the drilling correction control instructions. The drilling information transmission mechanism includes a drill bit sensor connected to the control console of the surface drilling rig and a downhole receiving probe located at the bottom of the well. The drill bit sensor is used to modulate the directional action command into a fluid pressure wave and transmit it downhole. The downhole receiving probe is used to receive and demodulate the fluid pressure wave to restore the directional action command.

[0008] The downhole guiding actuator is connected to the downhole receiving probe and the drill bit, respectively, and is used to drive the drill bit to adjust the drilling attitude according to the demodulated directional action command.

[0009] As a further improvement to the above technical solution, each of the ground acoustic sensors is arranged in a star-shaped or mesh topology around the drilling site, and the straight-line distance between any two adjacent ground acoustic sensors is no more than 200m.

[0010] As a further improvement to the above technical solution, the ground acoustic sensor adopts a multi-band sensing and acquisition chip to perform continuous ground acoustic signal acquisition around the clock, and multiple ground acoustic sensors are networked and coordinated to achieve simultaneous drilling monitoring of multiple wells.

[0011] As a further improvement to the above technical solution, the data processing center is equipped with a speed model correction module and an adaptive noise filter, which is used to filter out the mechanical background noise of the ground drilling rig.

[0012] As a further improvement to the above technical solution, the downhole guiding actuator is a rotary guiding system or a screw motor with a bend.

[0013] Secondly, the present invention also provides a drill bit guidance control method based on ground acoustic monitoring, which applies the system described in the first aspect and includes the following dynamic collaborative steps during drilling: S1. During the drilling process, the ground acoustic monitoring subsystem is used to continuously pick up the underground acoustic signals in real time, and the earth acoustic signals are transmitted to the data processing center through the signal acquisition station and optical fiber communication cable. S2. The data processing center performs time-frequency transformation preprocessing on the received ground acoustic signals and uses a preset neural network classification model to automatically identify and distinguish between continuous active source signals excited by the rotation of the drill bit breaking the rock and passive source signals generated by the fracturing of the surrounding rock. S3. Based on the identified active seismic source signal, the data processing center uses the time difference positioning algorithm to invert and calculate the real-time three-dimensional spatial coordinates of the drill bit, and compares them with the preset design trajectory to calculate the deviation vector. S4. When the deviation vector exceeds the preset threshold, the data processing center automatically generates a drilling correction control command and sends it to the ground drilling rig control console. The ground drilling rig control console triggers the drill bit sensor to convert the drilling correction control command into a fluid pressure wave and transmit it to the downhole receiving probe to control the downhole guide actuator to adjust the drilling attitude of the drill bit until the drill bit returns to the designed trajectory. S5. The data processing center simultaneously performs geological environment inversion, specifically using the active source signal and the passive source signal to perform travel time residual inversion, and imaging to obtain the wave velocity distribution parameters of the underground medium in the monitoring area, so as to invert the physical property occurrence state of the underground rock mass in real time and complete the closed-loop collaborative control of drilling correction and geological advance prediction.

[0014] As a further improvement to the technical solution, the specific method for identifying and distinguishing between continuous active source signals and passive source signals in step S2 is as follows: The acquired raw ground acoustic signals are preprocessed by time-frequency transformation to extract a two-dimensional feature map that characterizes the evolution of ground acoustic wave energy with time and frequency, and the energy distribution numerical matrix of the two-dimensional feature map is normalized. The normalized energy distribution numerical matrix is ​​used as the feature input. Multi-layer convolution and pooling downsampling operations are performed using convolution kernels to reduce the feature dimension and extract the vibration acoustic topology features of deep surrounding rock, resulting in a one-dimensional feature vector. The one-dimensional feature vector is input into a pre-trained neural network classification model for identification, and the corresponding ground sound signal category identifier is output.

[0015] As a further improvement to the technical solution, in step S3, the specific method for inverting and calculating the current real-time three-dimensional spatial coordinates of the drill bit is as follows: Obtain the arrival time observation value of the same ground acoustic signal picked up by each triggered ground acoustic sensor, and calculate the arrival time difference observation value between any two triggered ground acoustic sensors. ; Based on the spatial coordinates of the target seismic source Based on the equivalent wave velocity V, a theoretical regression equation for the arrival time difference between each of the aforementioned ground acoustic sensors is constructed. The objective function is constructed with the optimization constraint of minimizing the sum of squared deviations between the observed time difference values ​​and the theoretical regression equation output values ​​among all triggered ground acoustic sensors. : ; In the formula, and The first The and the first The known calibrated three-dimensional coordinates of the triggered ground acoustic sensor. The total number of ground acoustic sensors that were triggered. This is the time difference observation value. This is a theoretical prediction value. The time difference deviation between sensor i and sensor j represents the difference between the "observed value" and the "theoretical predicted value". Solve for the objective function The solution obtained at the minimum value determines the real-time three-dimensional spatial coordinates of the drill bit during drilling. .

[0016] As a further improvement to the technical solution, the specific method for geological environment inversion in step S5 is as follows: Based on the obtained locations of each seismic source, arrival time of the seismic source, and known calibration coordinates of each of the aforementioned ground acoustic sensors, the theoretical travel time of the seismic wave from each seismic source to each corresponding ground acoustic sensor is calculated using the initial velocity structure model. Calculate the time residual between the actual observed time and the theoretical time. ; The ray propagation path in the monitoring area is divided into grids, and the travel time residual matrix equation is constructed, which includes the source coordinate positioning error, the source release time error, the propagation path velocity structure difference error, and the observation error: ; In the formula, Let be the travel time residual vector. The vector to be determined includes the positioning correction error term of each seismic source on the three-dimensional coordinate axis, the release time correction error term, and the slow disturbance parameter term in each subdivided grid. This is a coefficient matrix determined by the ray path length; This is the observation noise error vector; The least squares method is used to iteratively solve the travel time residual matrix equation to obtain the underground medium wave velocity distribution parameters in each segmented grid. Tomographic imaging is then performed based on the wave velocity distribution parameters to invert and identify the vein boundary, fault, or non-uniform medium property occurrence state in front of the drill bit.

[0017] As a further improvement to the technical solution, the neural network classification model is pre-trained using a constructed active source waveform image sample database and a passive source waveform image sample database; wherein, respectively, selected Group of active source signal waveforms and The passive source signal waveforms are used as samples in the active source waveform image sample database and the passive source waveform image sample database, and .

[0018] As a further improvement to the technical solution, before continuously picking up the underground acoustic signals transmitted in real time during the drilling process as described in step S1, the method further includes: establishing an initial velocity model of the work area, conducting a blasting experiment in the work area, and using the known excitation signal generated by the blasting experiment to correct the initial velocity model of the work area, so as to ensure the inversion positioning accuracy of the three-dimensional spatial coordinates while drilling.

[0019] Because the present invention adopts the above technical solutions, the beneficial effects of this application are as follows: This invention provides a drill bit guidance control system based on ground acoustic monitoring. Through the closed-loop dynamic coordination of a large-scale ground acoustic joint sensing system and the downhole guiding actuator, it effectively solves the problems of spatial geometric blind spots and data transmission lag in traditional measurement while drilling. The specific beneficial effects are mainly reflected in the following aspects:

[0020] Firstly, it eliminates the spatial measurement blind zone of the drilling trajectory: This system utilizes an array of ground-based or shallow-hole acoustic sensors to directly and in real-time acquire continuous active ground-based acoustic signals generated by the drill bit's rotation and cutting of rock underground. The location of the vibration source is equivalent to the front contact point of the drill bit currently cutting the rock. By directly inverting and solving this drilling source using a time-difference positioning algorithm, the geometric blind zone caused by the traditional requirement of placing drilling measurement sensors behind the drill collar is avoided from a physical spatial layout perspective. This improves the directness of drilling position monitoring and the accuracy of leading-edge positioning.

[0021] Secondly, it enables highly real-time active closed-loop control during drilling: This invention constructs a bidirectional, rapid transmission loop between the surface and the wellbore through a "surface acoustic monitoring subsystem - surface drilling rig control console - drilling information transmission mechanism - downhole guidance actuator". The speed at which acoustic signals travel through the formation medium to the surface is much faster than the discrete upload speed of traditional downhole mud pulses, enabling the surface data processing center to continuously acquire high-density coordinate data and perform trajectory comparison in real time. When the deviation exceeds the limit, the correction command is instantly converted into a fluid pressure wave and transmitted down through the drilling rig control console. The downhole receiving probe demodulates the wave and directly drives the actuator to adjust the attitude. This system collaboration of "high-speed surface acquisition and calculation - real-time surface decision-making - real-time downhole execution" overcomes the technical drawbacks of the intermittent lag of traditional mud pulse signals, achieving active closed-loop feedback control during the drilling process.

[0022] Third, it enables deep synergy between drilling-while-drilling correction and geological advance prediction: the system's ground acoustic sensor array can simultaneously capture active seismic source signals from the drill bit and passive seismic source signals generated by surrounding rock fractures. The data processing center uses the precise location results of the active seismic source as known boundary constraints, combined with the spatiotemporal distribution of the passive seismic source, and uses the travel-time residual matrix equation to perform gridded tomographic imaging of the ray propagation path in the monitoring area. Thus, without interrupting normal drilling or introducing additional downhole advance detection equipment, it can inversely deduce the medium wave velocity distribution parameters and rock mass physical properties of the unexplored area in front of the drill bit, providing intuitive data support for avoiding fault geological anomalies and accurately tracking ore veins.

[0023] 2. This invention provides a drill bit guidance control method based on ground acoustic monitoring. Through dynamic coordination during drilling in steps S1 to S5, it deeply integrates a large-scale surface acoustic joint sensing system with downhole positioning and actuators, forming a closed-loop control process of "sensing-identification-positioning-correction-inversion". Based on rigorous physical and algorithmic logic, this method achieves the following technical effects:

[0024] Firstly, it eliminates the spatial measurement blind zone of the drilling trajectory from a physical mechanism perspective: According to steps S1 and S2, the system continuously picks up underground acoustic signals during drilling and uses a neural network classification model to accurately extract the continuous active seismic source signals excited by the rotation of the drill bit breaking the rock. Since the active seismic source is directly generated at the cutting contact surface between the drill bit tip and the rock mass, the three-dimensional spatial coordinates calculated by the time-difference positioning algorithm in step S3 directly represent the current true position of the drill bit's front end. This combination of features fundamentally avoids the geometric blind zone caused by the fact that sensors must be placed behind the drill collar in traditional drilling measurement technology, thus improving the directness of drilling trajectory monitoring and the accuracy of front-end positioning.

[0025] Secondly, it significantly improves the real-time performance of drilling directional control, achieving true active closed-loop control. In traditional technologies, downhole data transmission mainly relies on intermittent and extremely slow mud pulses. In this method, the ground acoustic signal picked up in step S1 travels through the formation to the surface at the speed of sound waves. Step S3 can continuously and frequently calculate the drill bit spatial coordinates and the deviation vector during drilling. Once the deviation exceeds the limit, step S4 generates and transmits a drilling correction control command in real time through the surface drilling rig control console, directly driving the downhole directional actuator to adjust the drill attitude. This dynamic coordination of "high-speed surface acoustic wave reception and high-frequency calculation" and "real-time surface decision-making and command transmission" overcomes the time lag bottleneck of downhole data transmission in traditional technologies, upgrading the system from traditional "post-event passive adjustment" to "active closed-loop control during drilling."

[0026] Third, it enables deep synergy between drilling-while-drilling correction and geological advance prediction, improving construction safety: Based on steps S2 and S5, the data processing center not only uses active seismic source signals for precise drilling-while-drilling positioning of the drill bit itself, but also simultaneously identifies and utilizes passive seismic source signals generated by surrounding rock fractures. During the inversion process, the active seismic source with known coordinates is used as a precise boundary and velocity field correction constraint, and the travel-time residual inversion is performed in conjunction with the spatiotemporal distribution of passive seismic sources, thereby obtaining the subsurface medium wave velocity distribution parameters within the monitoring area through tomographic imaging. This feature allows the system to simultaneously invert the rock mass physical properties of the unexplored area ahead of the drill bit while drilling, without interrupting normal drilling or adding additional downhole electrical or radioactive advance detection instruments, providing crucial geological data support for the early optimization of the active guidance route. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0028] Figure 1 This is a block diagram of a drill bit guidance control system based on ground acoustic monitoring disclosed in this invention; Figure 2 This is a flowchart illustrating a drill bit guidance control method based on ground acoustic monitoring disclosed in this invention.

[0029] The realization of the objective, functional characteristics and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] It should be noted that the technical solutions of the various embodiments of the present invention can be combined with each other, but only if they are based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0032] Example 1

[0033] See Figure 1 This invention provides a drill bit guidance control system based on ground acoustic monitoring. The system includes a ground acoustic monitoring subsystem, a ground drilling rig control console, a drilling-while-drilling information transmission mechanism, and a downhole guidance actuator.

[0034] Specifically, the ground acoustic monitoring subsystem includes multiple acoustic sensors arranged in an array, a signal acquisition station electrically connected to each acoustic sensor, and a data processing center connected to the signal acquisition station. To ensure the breadth of acquisition and spatial positioning accuracy of weak underground acoustic signals, each acoustic sensor is arranged in a star or mesh topology around the drilling site, and the straight-line distance between any two adjacent acoustic sensors is controlled to be no more than 200m. The acoustic sensors employ multi-band sensing and acquisition chips, capable of adapting to complex underground vibration responses at different frequencies, thus enabling continuous all-weather acoustic signal acquisition. Simultaneously, through the networked collaboration of multiple acoustic sensors, a single ground sensing network can be used to simultaneously monitor multiple wells within the work area during drilling, significantly reducing the overall equipment investment and maintenance costs during concurrent multi-well operations.

[0035] During system operation, once drilling is initiated via the ground drilling rig control console, the drill bit, rotating and cutting underground, acts as a natural passive vibration source, spontaneously and continuously emitting active ground acoustic signals. Simultaneously, drilling causes localized fracturing of the surrounding rock around the working face, generating passive ground acoustic signals. The ground acoustic sensor array picks up these two types of underground ground acoustic signals in real time and transmits them remotely to the data processing center via the signal acquisition station and fiber optic communication cable.

[0036] In this embodiment, the data processing center is equipped with a velocity model correction module and an adaptive noise filter. Since the ground drilling rig generates strong mechanical vibrations during operation, the adaptive noise filter can specifically filter out the mechanical background noise of the ground drilling rig, thereby extracting underground acoustic waveform data with a high signal-to-noise ratio. The velocity model correction module is used to import the work area velocity field model, which has been pre-corrected through blasting experiments and other methods, to provide accurate velocity medium parameters for subsequent high-precision drilling spatial positioning.

[0037] The data processing center first preprocesses the received ground acoustic signals using a pre-set neural network machine learning model, and automatically identifies and distinguishes between continuous active source signals excited by the drill bit's rotation breaking the rock and passive source signals generated by the fracturing of the surrounding rock. Since the continuous active source signals are directly generated on the physical contact surface where the drill bit tip cuts the rock, the data processing center can use a time-difference positioning algorithm to invert and calculate the drill bit's current real-time three-dimensional spatial coordinates based on the identified active source signals. Through this acoustic inversion positioning method, the calculated coordinates are the actual contact point position of the drill bit tip, thus eliminating the geometric blind spots in the drilling trajectory caused by the rear-mounted sensors in traditional measurement-while-drilling (MWD) systems.

[0038] During the trajectory decision-making and closed-loop control phase, the data processing center compares the calculated real-time three-dimensional spatial coordinates of the drill bit with the preset design trajectory in real time, and automatically calculates the deviation vector between the two. When the deviation vector exceeds a preset threshold, it indicates that the drill bit has deviated from the design target point, and the data processing center automatically generates a drilling correction control command and sends it to the ground drilling rig control console.

[0039] After receiving a control command, the surface drilling rig control console sends a directional adjustment command to the drilling information transmission mechanism. The drilling information transmission mechanism includes a drill bit sensor connected to the surface drilling rig control console and a downhole receiving probe located at the bottom of the well. The drill bit sensor modulates the received directional adjustment command into a fluid pressure wave and transmits it downhole. Because the ground acoustic signal is transmitted to the surface at the formation sound speed, and the surface calculation and processing and command issuance are in a high-frequency continuous state, this fundamentally overcomes the intermittent data lag problem caused by the extremely low transmission rate when traditional drilling data is transmitted from downhole using mud pulses, thus improving the real-time performance of drilling control.

[0040] The downhole receiving probe receives and demodulates the fluid pressure wave, reconstructs the initial directional control command, and transmits it to the downhole directional actuator. The downhole directional actuator is connected to both the downhole receiving probe and the drill bit, and can specifically employ a rotary steering system (RSS) or a screw motor with an angle. Based on the demodulated directional control command, the downhole directional actuator drives the drill bit to adjust its drilling attitude (e.g., adjusting the directional tool face angle), guiding the drill bit towards the designed trajectory until it completely returns to the preset design trajectory, thereby achieving efficient and real-time active closed-loop directional control between the surface and downhole.

[0041] Meanwhile, to maximize the utilization of the acquired acoustic data, the data processing center simultaneously performs geological environment inversion while controlling the trajectory. Specifically, the data processing center collaboratively utilizes the active and passive seismic source signals to perform travel-time residual inversion. Using the precise location of the active seismic source as a known boundary constraint, it solves the travel-time residual matrix equation after ray path gridding, thereby imaging and obtaining the subsurface medium wave velocity distribution parameters within the monitoring area. Through tomographic imaging using these wave velocity distribution parameters, the physical properties of the subsurface rock mass in unexplored areas ahead of the drill bit can be inverted in real time without interrupting normal drilling and without the need for additional expensive electrical or radioactive advanced detection instruments installed downhole. This allows for the identification of vein boundaries, faults, or non-homogeneous medium anomalies. This enables a deep closed-loop synergy between drilling-while-drilling deviation control and geological advanced prediction, avoiding hidden geological hazards and improving the overall construction safety and material economy of mineral exploration or oil drilling.

[0042] Example 2

[0043] See Figure 2 This invention also provides a drill bit guidance control method based on ground acoustic monitoring. Applying a drill bit guidance control system based on ground acoustic monitoring as described in Example 1, this method addresses the technical problems of drill bit geometric blind spots and intermittent lag in trajectory data transmission that prevent real-time closed-loop guidance in traditional measurement-while-drilling (MWD) technology. This method achieves dynamic coordination between a large-scale surface acoustic joint sensing system and the downhole MWD guidance actuator. The specific execution steps are as follows:

[0044] During drilling operations, step S1 is executed: the ground acoustic monitoring subsystem continuously acquires ground acoustic signals transmitted from underground in real time, and remotely transmits these signals to the data processing center via a signal acquisition station and fiber optic communication cable. In this process, the ground acoustic signals travel directly through the formation medium to the surface at the speed of sound waves. This transmission speed is much faster than the traditional uplink speed of downhole mud pulses, providing a fundamental physical transmission path for the surface control terminal to acquire high-density, high-frequency downhole data. This effectively overcomes the discreteness and lag in the upward transmission of downhole signals in traditional technologies.

[0045] Subsequently, step S2 is executed: After receiving the ground acoustic signal, the data processing center first performs time-frequency transformation preprocessing to extract a two-dimensional feature map characterizing the evolution of ground acoustic wave energy over time and frequency. Then, using a preset neural network machine learning model, it automatically identifies and distinguishes between continuous active source signals excited by the rotation of the drill bit breaking the rock and passive source signals generated by the fracturing of the surrounding rock. Through the combination of time-frequency transformation and feature extraction from the neural network model in this step, a pure continuous active source signal specifically characterizing the cutting state of the drill bit tip can be accurately extracted from the complex underground integrated vibration response, providing reliable data support for subsequent high-precision trajectory positioning.

[0046] Next, step S3 is executed: Based on the continuous active seismic source signal identified and separated in step S2, the data processing center uses a time-difference positioning algorithm to invert and calculate the real-time three-dimensional spatial coordinates of the drill bit while drilling, and compares them with the preset design trajectory inside the system to calculate the deviation vector. Since the continuous active seismic source signal is directly generated on the physical contact surface where the drill bit tip cuts the rock, the coordinates calculated directly by the time-difference positioning algorithm directly represent the true spatial position of the drill bit's front end. This combination of features eliminates the geometric blind zone of the drilling trajectory caused by the sensor having to be placed behind the drill collar in traditional measurement techniques, thus improving the directness of drilling position monitoring and the accuracy of front-end positioning.

[0047] When the deviation vector calculated in step S3 exceeds a preset threshold, it indicates that the drill bit has deviated from the design target point. At this time, step S4 is executed: the data processing center automatically generates a drilling correction control command and sends it to the surface drilling rig control console. The surface drilling rig control console triggers the drill bit sensor, converting the drilling correction control command into a fluid pressure wave and transmitting it to the downhole receiving probe to control the downhole directional actuator to adjust the drilling attitude of the drill bit until the drill bit returns to the design trajectory. In this step, "high-speed acoustic wave reception and high-frequency continuous calculation on the surface" and "real-time decision-making and transmission of pressure wave commands on the surface" construct a rapid bidirectional transmission closed loop between the surface and the well bottom. This dynamic collaborative architecture for drilling can overcome the technical drawbacks of intermittent data lag in traditional drilling measurement, enabling the system to make a second-level response based on trajectory deviation, realizing true active drilling closed-loop feedback control.

[0048] While executing trajectory control, step S5 is performed simultaneously: the data processing center synchronously performs geological environment inversion, specifically using the active source signal and the passive source signal to perform travel-time residual inversion, and imaging to obtain the subsurface medium wave velocity distribution parameters within the monitoring area, so as to invert the physical properties of the subsurface rock mass in real time. In this step, the data processing center uses the accurate positioning results of the active source obtained in step S3 as known boundary constraints and velocity field correction parameters, combined with the spatiotemporal distribution of the passive source, and uses the travel-time residual matrix equation to perform gridded segmentation and analysis imaging of the ray propagation path within the monitoring area. This combination of features enables the system to synchronously invert the medium wave velocity distribution parameters and rock mass physical properties of the unexplored area ahead of the drill bit without interrupting normal drilling and without the need to install expensive electrical or radioactive advanced detection instruments downhole. This achieves a deep closed-loop synergy between drilling-while-drilling deviation control and geological advance prediction, which can provide key geological data support for the early optimization design of active deviation-while-drilling routes, and effectively avoid hidden geological hazards such as faults and non-uniform media ahead, greatly improving the overall construction safety factor and exploration economy of directional drilling.

[0049] In a preferred embodiment, the specific implementation process and system logic for identifying and distinguishing between continuous active source signals and passive source signals in step S2 are as follows: Before monitoring while drilling, a corresponding waveform database is first established using manually identified active and passive seismic source signals. Specifically, waveforms are selected respectively. Group of active source signal waveforms and The passive source signal waveforms are used as a waveform image sample database, and the size of the two satisfies the following relationship: ; In the formula, This represents the number of active source signal waveforms selected from the sample database. This represents the number of passive source signal waveform groups selected from the sample database. By ensuring that the cardinality of the two types of source waveform image samples is completely equal, the sample asymmetry bias during machine learning training can be effectively eliminated, ensuring the robustness of the subsequent identification mechanism in classification accuracy.

[0050] During actual drilling operations, the ground data processing center performs time-frequency transformation preprocessing on the acquired raw ground acoustic signals. Because the underground drilling environment contains complex background mechanical noise and stratum scattering interference, one-dimensional time-domain waveforms are difficult to intuitively distinguish the source material properties. Through time-frequency transformation processing, these signals can be converted into a two-dimensional feature map characterizing the evolution of ground acoustic wave energy over time and frequency, jointly mapping the frequency domain evolution trend of the acoustic signal with the time domain energy distribution.

[0051] To eliminate differences in acoustic energy attenuation or sensor coupling caused by varying sensor spacing, the sizes of the two types of acquired waveform images need to be adjusted to be the same. The grayscale value of each pixel in each waveform image is then extracted to construct a corresponding two-dimensional matrix; the grayscale value range of each pixel in the waveform image is defined as 0~255. Subsequently, the resulting two-dimensional matrix is ​​normalized. This restricts the values ​​of each element within the matrix to a uniform standard range, eliminating the interference of absolute amplitude values ​​on feature weights, thereby significantly enhancing the stability of the input features and improving the convergence speed of subsequent algorithm models.

[0052] Subsequently, the normalized energy distribution numerical matrix (i.e., the two-dimensional feature map) is used as the bottom-level feature input, and multi-layer convolution and pooling downsampling operations are performed using a preset convolution kernel. During this process, the two-dimensional feature map undergoes sliding window convolution operations to obtain a new feature matrix, which can adaptively extract the acoustic topological features of deep surrounding rock vibrations. These acoustic topological features include specific frequency band energy ripples generated by continuous drill cutting or transient pulse boundary features generated by surrounding rock fracturing. Simultaneously, pooling downsampling is performed, and through several alternating convolution-pooling operations, the dimensionality of the feature map is significantly reduced while preserving the core acoustic topological features. This effectively prevents network overfitting due to feature redundancy and reduces the computational overhead of the data processing center during drilling. Finally, after several dimensionality reduction processes, the two-dimensional feature matrix is ​​transformed into a one-dimensional row vector.

[0053] Finally, the one-dimensional feature vector is input into a pre-established and trained neural network machine learning model for identification. The neural network machine learning model then outputs the corresponding ground sound signal category identifier, thereby accurately determining whether the ground sound signal belongs to the active source signal of the drill bit breaking rock or the passive source signal of the surrounding rock fracturing.

[0054] By introducing the aforementioned combined identification mechanism based on time-frequency image transformation and neural network machine learning model, the traditional lagging identification method that relies on human experience is replaced. This enables the system to achieve rapid, fully automatic, and accurate classification of drill bit rock breaking signals and surrounding rock fracture signals in complex drilling environments with severe vibrations and high background noise. This provides a highly reliable data classification foundation and closed-loop logic support for subsequent steps S3, which uses a clean, continuous active seismic source for precise positioning of the drill bit tip without blind spots, and for step S5, which uses a passive seismic source to perform travel-time residual inversion tomography of the underground medium.

[0055] As a preferred embodiment, the specific implementation process, mathematical model, and internal control logic of step S3, which involves inverting and calculating the current real-time three-dimensional spatial coordinates of the drill bit, are as follows: 1. Establishment of the Nonlinear Space Physics Equations While Drilling: During drilling operations, the drill bit acts as a continuously excited active seismic source as it rotates and cuts through rock underground. Assume the spatial coordinates of the location of the seismic source while drilling are... The equivalent propagation speed of seismic waves in the void and rock mass is Both are unknowns in their initial state. When the seismic wave signal generated by the rock-breaking vibration propagates outward through the formation medium, let the initial moment of the seismic wave release from the source be... After propagating along the path, the signal reaches the first [unit / entry point] positioned on the ground or in a shallow hole. One ground acoustic sensor (of which) The absolute arrival times of (=1, 2...n) are The corresponding straight-line distance of its spatial propagation path is The geometric and temporal mapping relationships of the signals received by a single-node sensor satisfy the following two equations: ; ; Similarly, to achieve global positioning, the ground acoustic monitoring subsystem physically combines the spatial propagation paths, equivalent wave velocities, initial excitation times, and absolute arrival times of all sensor nodes that receive the signal across the entire network, constructing the following set of nonlinear spatial physical equations during drilling: ; In the formula, , , , They are respectively the first, second, and so on up to the first. The known pre-calibrated high-precision spatial coordinates of the successfully triggered ground acoustic sensor; , , , These are the actual measured values ​​of the clock time when the acoustic waveform is captured by the corresponding sensor node; This represents the total number of triggering sensors participating in collaborative localization within the current sensing cycle.

[0056] 2. Construction of the Time Difference Theory Regression Equation System: In actual complex work areas, although the aforementioned nonlinear equation system is mathematically feasible for solving spatial coordinates, the non-uniformity of underground rock wave velocity during propagation and the unavoidable small observation errors in long-distance transmission and time synchronization at each monitoring station lead to biases in the directly solved results. To eliminate the unmeasurable initial excitation time at the well bottom... To mitigate interference with the positioning accuracy of the control terminal, this embodiment further performs pairwise difference transformations on the equation system, converting the absolute time relationship into a relative time difference, and constructs the following theoretical regression equation system for the time difference between any two trigger sensors:

[0057] ; In the formula, Let be the theoretical regression value of the arrival time difference between the i-th and j-th ground acoustic sensors regarding the same seismic wave signal, where and Through this mathematical transformation, the unknown term... The absolute clock hard synchronization constraint between the downhole excitation source and the ground station is completely decoupled from the physical mechanism during the subtraction process.

[0058] 3. Construction of the Deviation Equation Set and the Global Drilling Control Objective Function: The time difference calculated by the above formula is the theoretical regression value, which is compared with the actual observed value measured by the field hardware. There are certain statistical residuals between them. In order to quantitatively characterize and incorporate these measurement residuals, this embodiment constructs the following set of drilling residual deviation equations for the time difference between each pair of sensors: ; In the formula, The measured time difference between any two ground acoustic sensors during drilling is obtained by collecting and identifying data from the ground data processing center. Let $\frac{i}{j}$ be the arrival time difference deviation term between the $i$-th sensor and the $j$-th sensor, used to accurately measure the statistical difference between the theoretically predicted curve and the discrete data points measured in the industrial field. Based on the optimization mathematical principle that a smaller deviation results in a higher degree of fit, in order to obtain the most noise-resistant solution in the noisy environment of drilling with severe mechanical vibration, this embodiment transforms the trajectory positioning problem into an optimization statically indeterminate problem that minimizes the sum of squares of arrival time difference deviations between all trigger sensors in the network. Therefore, a global deviation objective function is defined. for: ; In the formula, and The first The and the first The known calibrated three-dimensional coordinates of the triggered ground acoustic sensor. The total number of ground acoustic sensors that were triggered. This is the time difference observation value. This is a theoretical prediction value. Let be the time difference deviation between sensor i and sensor j, representing the difference between the "observed value" and the "theoretical predicted value"; the ground data processing center uses the objective function constructed above... The solution is performed to find the optimal parameter vector that minimizes the sum of squared residuals of the global fit. The optimal convergent solution obtained at this point is ( This is determined as the real-time three-dimensional spatial coordinates of the drill bit on the working surface.

[0059] Specifically, in solving the global bias objective function At that time, there exists a set Make To reach a minimum, when the deviation reaches a minimum, the following condition is met: ; ; ; .

[0060] Since the active vibration source for the drill bit's rotational rock breaking occurs at the working surface thousands of meters underground, it is usually impossible to directly install a hard-synchronized trigger clock on the ground to measure its accurate initial excitation time. This embodiment constructs a regression equation by introducing a difference mechanism based on the absolute arrival time of dual sensors, thus making the unknown variables... The signals are directly eliminated during the subtraction process. This means that the ground acoustic monitoring subsystem does not need to perform high-precision absolute time hard synchronization with the drilling tool face thousands of meters underground during operation; it only relies on the relative time difference of the signals received by each sensor node on the ground. This allows for the convergence calculation of the coordinates across the entire drilling space. At the underlying algorithm level, this mechanism decouples the time clock synchronization constraints between the surface sensing end and the downhole execution end. This not only significantly reduces the stringent requirements of the drilling monitoring system for long-distance hardware clock synchronization but also avoids clock drift errors caused by the harsh industrial environment at the bottom of the well, effectively ensuring the real-time performance of the drilling trajectory calculation and the engineering feasibility of the system.

[0061] In a preferred embodiment, step S5 includes the following process for geological environment inversion.

[0062] During the drilling process, the data processing center acquires the active seismic source signals of rock breaking and the passive seismic source signals of surrounding rock fracturing, which were identified through the aforementioned steps. It also records the initial position of each source, the arrival time of the source, and the calibration coordinates of the ground acoustic sensors. The positions of the aforementioned sources can be initially determined by the aforementioned source positioning steps, and the coordinates of the ground acoustic sensors are pre-calibrated known quantities.

[0063] Subsequently, the data processing center invoked the initial velocity-structure model of the monitoring area to calculate the theoretical travel time from the j-th seismic source to the i-th ground acoustic sensor. Let L be the propagation path from the j-th seismic source to the i-th ground acoustic sensor. ij Then its walking time Represented as: ; in, For the seismic wave along path L ij The velocity components at spatial coordinates (x, y, z) during propagation. This represents the slowness component at the corresponding location, where slowness is the reciprocal of wave speed. Let the length of the infinitesimal element on this path be . This corresponds to the theoretical timeframe.

[0064] After obtaining the theoretical travel time, calculate the travel time residual between the observed travel time and the theoretical travel time: ; in, Let be the travel time residual from the j-th seismic source to the i-th ground acoustic sensor. This represents the observation travel time corresponding to the actual signal received by the ground acoustic sensor from the seismic source. This is the theoretical travel time calculated based on the initial velocity structure model.

[0065] Considering that the source coordinates obtained from the aforementioned positioning steps may still have a small error, and that there may also be differences between the velocity structure of the underground medium and the initial model, the travel time residual is decomposed into source coordinate positioning error, source release time error, propagation velocity structure difference error, and observation error, resulting in the following residual expression: ; ; in, , , These represent the coordinate corrections for the j-th earthquake source in the X, Y, and Z axes, respectively. The error caused by the coordinate positioning of earthquake source j; This is a correction amount for the moment of earthquake source release; For path L ij The slow perturbation at the upper spatial location (x,y,z); This is the observation error term for the source-sensor pair.

[0066] Furthermore, the monitoring area is divided into K grids along the ray propagation path, and the propagation path length of the ray in the k-th grid is set to R. k The slowness within this grid is Then the theoretical travel time from the j-th seismic source to the i-th ground acoustic sensor can be written as: ; in, Let be the path length of the ray within the k-th grid. Let K be the slowness within the k-th grid, and K be the number of grids participating in the inversion within the monitoring area.

[0067] Based on the above gridding process, the travel time residual can be further written as: ; in, The wave velocity obtained by inversion within the k-th grid; This represents the reference wave velocity in the initial velocity structure model. The above summation term characterizes the effect of the difference between the actual grid wave velocity and the initial reference wave velocity on the travel time residual.

[0068] Furthermore, the above formula can be equivalently expressed as: ; in, Let be the theoretical travel time component of the ray from the j-th seismic source to the i-th ground acoustic sensor within the k-th grid.

[0069] Furthermore, let ,in, Let be the velocity perturbation parameter of the k-th grid relative to the initial velocity structure model, used to characterize the deviation of the actual wave velocity within this grid from the reference wave velocity. Then the travel time residual equation can be expressed as: .

[0070] When calculating the source coordinate correction, the partial derivative relationship in the source coordinate error term can be expressed as: ; in, , , Let be the three-dimensional coordinates of the i-th ground acoustic sensor; , , Let be the three-dimensional coordinates of the j-th earthquake source; Let j be the release time of the j-th seismic source; The theory is outdated.

[0071] When there are M ground acoustic sensors, N1 active rock-breaking sources from drill bits, and N2 passive rock-fracture sources within the monitoring area, the total number of sources is N, and N1 + N2 = N. The travel time residuals of all source-sensor pairs are organized into a residual vector: ; in, is the transpose of the travel time residual vector; M is the number of ground acoustic sensors; N is the total number of seismic sources.

[0072] The coordinate corrections for each seismic source, the release time corrections, and the velocity perturbation parameters for each grid are organized into a vector of parameters to be determined: ; Where ω is the parameter vector to be determined; to , to , to These represent the coordinate corrections for each seismic source on the three-dimensional coordinate axes; to These are the correction amounts for the release time of each seismic source; to These are the velocity perturbation parameters within each subdivided grid.

[0073] Therefore, the overall travel time residual equation within the monitoring area can be written in matrix form: ; in, ω is the travel time residual vector; A is the coefficient matrix, whose elements are determined by the ray path length, the partial derivative of theoretical travel time with respect to source coordinates, and the theoretical travel time components in each grid; ω is the parameter vector to be determined; ε is the observation noise error vector.

[0074] The data processing center uses the least squares method to iteratively solve the above residual matrix equations to obtain the correction values ​​for each source coordinate, the correction value for the release time, and the velocity disturbance parameters within each segmented grid. Based on this, the subsurface medium wave velocity distribution within the monitoring area is updated. Then, tomographic imaging is performed based on the updated wave velocity distribution to form velocity anomaly images of the medium in front of and around the drill bit, which are used to identify vein boundaries, fault locations, and the physical properties of non-homogeneous media.

[0075] Through the above processing, the guidance judgment is no longer based solely on the geometric deviation between the current position of the drill bit and the preset trajectory. Instead, it can combine the drill bit's rock-breaking signal and the surrounding rock fracturing signal to perform inverse analysis of the geological environment ahead of the drill bit. This method helps reduce positioning deviations caused by initial velocity model errors and can, to some extent, provide data for the drill bit to avoid faults, track vein boundaries, or adjust the drilling direction, thereby reducing time and material consumption caused by ineffective drilling and repeated drilling.

[0076] In a preferred embodiment, the neural network classification model is pre-trained using an active source waveform image sample database and a passive source waveform image sample database.

[0077] Specifically, select A database of active source waveform image samples was established using active source signals from the drill bit rock-breaking system. A database of passive source waveform image samples was established based on the passive source signal waveforms of the surrounding rock fracture. .in, This represents the number of active source signal sample groups. The number of passive source signal sample groups is set to the same value. Having the same number of both types helps reduce the impact of sample class imbalance on model training. During training, both types of waveforms are uniformly converted into waveform images of the same size. The grayscale values ​​of each pixel are extracted to form a two-dimensional matrix, which is then normalized. After convolution, pooling, and feature expansion, the matrix is ​​input into the neural network classification model, allowing the model to learn the differences in waveform morphology and local features between active and passive source signals. After training, the data processing center processes real-time acquired ground acoustic waveforms in the same way and inputs them into the model to obtain the classification results for active or passive sources. This provides a relatively reliable signal source for drill bit positioning, guidance judgment, and drilling environment inversion.

[0078] In a preferred embodiment, before step S1, an initial velocity model for the work area is first established and calibrated. Specifically, after the ground acoustic sensors are deployed and their coordinates are calibrated, the data processing center establishes an initial velocity model based on the geological data, borehole data, and lithological distribution data of the work area. Subsequently, blasting experiments are conducted at known test excitation points within the work area. The known excitation signals generated by the blasting are used as calibration signals, which are received and recorded by the ground acoustic sensors at each location. The data processing center compares the actual observed arrival times with the theoretical arrival times calculated based on the initial velocity model and corrects the propagation velocity parameters in the velocity model according to the arrival time difference. The calibrated velocity model is used for subsequent drill bit source positioning, travel time calculation, and geological environment inversion to reduce the impact of rock mass heterogeneity, void distribution, and lithological changes on the positioning results and improve the stability of the three-dimensional spatial coordinate inversion while drilling.

[0079] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's specification and drawings under the inventive concept of the present invention, or direct or indirect applications in other related technical fields, are within the patent protection scope of the present invention.

Claims

1. A drill bit guidance control method based on ground acoustic monitoring, characterized in that, A drill bit guidance control system based on ground acoustic monitoring is provided, comprising a surface ground acoustic monitoring subsystem, a surface drilling rig control console, and a downhole guidance actuator. The surface ground acoustic monitoring subsystem includes multiple ground acoustic sensors arranged in an array, a signal acquisition station electrically connected to each ground acoustic sensor, and a data processing center connected to the signal acquisition station. The drill bit guidance control method includes the following dynamic collaborative steps during drilling: S1. During the drilling process, the ground acoustic monitoring subsystem is used to continuously pick up the underground acoustic signals in real time, and the acoustic signals are transmitted to the data processing center through the signal acquisition station and optical fiber communication cable. S2. The data processing center performs time-frequency transformation preprocessing on the received ground acoustic signals and uses a preset neural network classification model to automatically identify and distinguish between continuous active source signals excited by the rotation of the drill bit breaking the rock and passive source signals generated by the fracturing of the surrounding rock. S3. Based on the identified active seismic source signal, the data processing center uses the time difference positioning algorithm to invert and calculate the real-time three-dimensional spatial coordinates of the drill bit, and compares them with the preset design trajectory to calculate the deviation vector. S4. When the deviation vector exceeds the preset threshold, the data processing center automatically generates a drilling correction control command and sends it to the ground drilling rig control console. The ground drilling rig control console triggers the drill bit sensor to convert the drilling correction control command into a fluid pressure wave and transmit it to the downhole receiving probe to control the downhole guide actuator to adjust the drilling attitude of the drill bit until the drill bit returns to the designed trajectory. S5. The data processing center simultaneously performs geological environment inversion. Specifically, it uses the active source signal and the passive source signal to perform travel time residual inversion, and images to obtain the wave velocity distribution parameters of the underground medium in the monitoring area. This allows for real-time inversion of the physical properties of the underground rock mass, providing key geological data support for the early optimization design of the active correction route while drilling. It can also effectively avoid hidden geological hazards ahead and complete the closed-loop collaborative control of correction while drilling and geological advance prediction.

2. The drill bit guidance control method based on ground acoustic monitoring according to claim 1, characterized in that, In step S2, the specific method for identifying and distinguishing between continuous active source signals and passive source signals is as follows: The acquired raw ground acoustic signal is preprocessed by time-frequency transformation to extract a two-dimensional feature map that characterizes the evolution of ground acoustic wave energy with time and frequency, and the energy distribution numerical matrix of the two-dimensional feature map is normalized. The normalized energy distribution numerical matrix is ​​used as the feature input. Multi-layer convolution and pooling downsampling operations are performed using convolution kernels to reduce the feature dimension and extract the vibration acoustic topology features of deep surrounding rock to obtain a one-dimensional feature vector. The one-dimensional feature vector is input into a pre-trained neural network classification model for identification, and the corresponding ground sound signal category identifier is output.

3. The drill bit guidance control method based on ground acoustic monitoring according to claim 1, characterized in that, In step S3, the specific method for inverting and calculating the current real-time three-dimensional spatial coordinates of the drill bit is as follows: Obtain the arrival time observation value of the same ground acoustic signal picked up by each triggered ground acoustic sensor, and calculate the arrival time difference observation value between any two triggered ground acoustic sensors. ; Based on the spatial coordinates of the target seismic source Based on the equivalent wave velocity V, a theoretical regression equation for the arrival time difference between each of the aforementioned ground acoustic sensors is constructed. The objective function is constructed with the optimization constraint of minimizing the sum of squared deviations between the observed time difference values ​​and the theoretical regression equation output values ​​among all triggered ground acoustic sensors. : ； In the formula, and The first The and the first The known calibrated three-dimensional coordinates of the triggered ground acoustic sensor. The total number of ground acoustic sensors that were triggered. This is the time difference observation value. This is a theoretical prediction value. The time difference deviation between sensor i and sensor j; Solve for the objective function The solution obtained at the minimum value determines the real-time three-dimensional spatial coordinates of the drill bit during drilling. .

4. The drill bit guidance control method based on ground acoustic monitoring according to claim 1, characterized in that, The specific method for geological environment inversion in step S5 is as follows: based on the obtained locations of each seismic source, the arrival time of the seismic source, and the known calibration coordinates of each of the aforementioned ground acoustic sensors, the theoretical travel time of the seismic wave from each seismic source to each corresponding ground acoustic sensor is calculated in conjunction with the initial velocity structure model. Calculate the time residual between the actual observed time and the theoretical time. ; The ray propagation path in the monitoring area is divided into grids, and the travel time residual matrix equation is constructed, which includes the source coordinate positioning error, the source release time error, the propagation path velocity structure difference error, and the observation error: ； In the formula, Let be the travel time residual vector. The vector to be determined includes the positioning correction error term of each seismic source on the three-dimensional coordinate axis, the release time correction error term, and the slow disturbance parameter term in each subdivided grid. This is a coefficient matrix determined by the ray path length; This is the observation noise error vector; The least squares method is used to iteratively solve the travel time residual matrix equation to obtain the underground medium wave velocity distribution parameters in each segmented grid. Tomographic imaging is then performed based on the wave velocity distribution parameters to invert and identify the vein boundary, fault, or non-uniform medium property occurrence state in front of the drill bit.

5. The drill bit guidance control method based on ground acoustic monitoring according to claim 2, characterized in that, The neural network classification model was pre-trained using a constructed active source waveform image sample database and a passive source waveform image sample database; wherein, respectively, active source waveform image sample databases were selected. Group of active source signal waveforms and The passive source signal waveforms are used as samples in the active source waveform image sample database and the passive source waveform image sample database, and .

6. The drill bit guidance control method based on ground acoustic monitoring according to claim 1, characterized in that, Before continuously picking up the underground acoustic signals in real time during the drilling process as described in step S1, the method further includes: establishing an initial velocity model of the work area, conducting a blasting experiment in the work area, and using the known excitation signal generated by the blasting experiment to correct the initial velocity model of the work area, so as to ensure the inversion positioning accuracy of the three-dimensional spatial coordinates while drilling.

7. The drill bit guidance control method based on ground acoustic monitoring according to claim 1, characterized in that, The ground acoustic sensors are used to pick up the ground acoustic signals generated by the rotation of the drill bit breaking the rock, as well as the ground acoustic signals generated by the fracturing of the surrounding rock. The data processing center is used to preprocess the received ground acoustic signals to automatically identify and distinguish the continuous active source signals generated by the rotation of the drill bit breaking the rock and the passive source signals generated by the fracturing of the surrounding rock. Based on the identified active source signals, the center performs drilling spatial position calculation and geological environment inversion, and outputs drilling deviation correction control commands. The ground drilling rig control console is bidirectionally connected to the data processing center and is used to issue directional adjustment commands according to the drilling deviation correction control commands. The drill bit guidance control system also includes a drilling information transmission mechanism, which includes a drill bit sensor connected to the surface drilling rig control console and a downhole receiving probe located at the bottom of the well. The drill bit sensor is used to modulate the directional action command into a fluid pressure wave and transmit it downhole. The downhole receiving probe is used to receive and demodulate the fluid pressure wave to reconstruct the directional action command. The downhole guidance actuator is connected to the downhole receiving probe and the drill bit, respectively, and is used to drive the drill bit to adjust the drilling attitude according to the demodulated directional action command.

8. The drill bit guidance control method based on ground acoustic monitoring according to claim 1, characterized in that, Each of the aforementioned ground acoustic sensors is arranged in a star-shaped or mesh-like topology around the drilling site, and the straight-line distance between any two adjacent ground acoustic sensors is no greater than 200m.

9. The drill bit guidance control method based on ground acoustic monitoring according to claim 1, characterized in that, The data processing center is equipped with a velocity model correction module and an adaptive noise filter, which is used to filter out the mechanical background noise of the ground drilling rig.

10. The drill bit guidance control method based on ground acoustic monitoring according to claim 1, characterized in that, The downhole guiding actuator is a rotary guiding system or a screw motor with a bend.

Images