A mobile rare earth ore sampling drilling machine and a sampling method

Abstract

The application discloses a mobile rare earth ore sampling drilling machine and a sampling method, and belongs to the field of rare earth ore sampling. The mobile rare earth ore sampling drilling machine comprises a caterpillar walking mechanism, an XY moving platform arranged on the caterpillar walking mechanism, a drilling tool arranged on the output end of the XY moving platform, a multi-source sensing assembly for collecting environment and operation parameters, and an intelligent control module. The multi-source sensing assembly is electrically connected with the input end of the intelligent control module. The output end of the intelligent control module is electrically connected with the XY moving platform and the drilling tool respectively, so that the operation parameters can be adjusted according to the collected environment and operation parameters. The mobile rare earth ore sampling drilling machine and the sampling method are suitable for the automatic sampling operation of the shallow rare earth ore under the complex terrain without pollution, with high precision and high efficiency.

Description

Technical Field

[0001] This invention relates to the field of rare earth ore sampling technology, and in particular to a mobile rare earth ore sampling drill and sampling method. Background Technology

[0002] Rare earth minerals are a national mineral resource. The collection of shallow surface (0~30cm) mineral samples is the core link in rare earth mineral resource exploration, reserve estimation and mining area environmental monitoring. The drilling and sampling process must strictly control the drilling depth and drilling position to prevent the core from being squeezed and deformed, so as to ensure that the collected samples accurately reflect the occurrence characteristics and distribution patterns of underground rare earth minerals.

[0003] Currently, shallow rare earth ore sampling operations mainly use two types of equipment, both of which have significant technical defects: one type is handheld drilling rigs. These devices have a simple structure but cannot adjust the drill bit torque in real time according to the soil hardness. The matching of the drill bit operating parameters with the formation conditions is extremely poor, which not only causes a large amount of ineffective energy consumption but also accelerates the wear of the drill bit. At the same time, they rely entirely on manual hand operation, which consumes a lot of physical strength for the operators and poses safety hazards such as mechanical collisions and soil collapses. The other type is traditional mechanical drilling rigs. Although they can provide a certain amount of drilling power, they are generally bulky, time-consuming and laborious to disassemble and transport, and have poor mobility in the field. Moreover, the drilling position can only be roughly aligned by visual inspection, and precise positioning and control cannot be achieved.

[0004] At a deeper level, existing shallow rare earth ore sampling technologies generally suffer from two fundamental defects: First, the positioning accuracy is severely insufficient. Traditional sampling equipment relies on manual visual alignment, which cannot achieve accurate positioning and exploration grid matching in complex field environments without GPS signals, such as slopes and gravel areas. The planar deviation of the sampling point can reach more than centimeters, making it difficult to meet the location requirements of high-precision exploration sampling. Second, soil hardness detection is lagging and relies on manual input. Existing drilling rigs cannot achieve online real-time soil hardness detection. Operators need to manually predict the Protodyakonov hardness of the soil and manually set drilling parameters. The hardness judgment error exceeds 20%, which directly leads to a very low match between key parameters such as drill bit speed, output torque, and descent speed and actual formation conditions. This further aggravates abnormal wear of the drill bit and energy waste, and at the same time, it is easy to cause core compression deformation and sampling distortion, which cannot provide real and reliable sample support for rare earth ore exploration. Summary of the Invention

[0005] The purpose of this invention is to provide a mobile rare earth ore sampling drill and sampling method to solve the above-mentioned technical problems.

[0006] To achieve the above objectives, the present invention provides a mobile rare earth ore sampling drill, including a tracked walking mechanism, an XY moving platform disposed on the tracked walking mechanism, a drill bit disposed on the output end of the XY moving platform, a multi-source sensing component for collecting environmental and operating parameters, and an intelligent control module. The multi-source sensing component is electrically connected to the input end of the intelligent control module, and the output end of the intelligent control module is electrically connected to the XY moving platform and the drill bit, respectively, so as to realize the adjustment of operating parameters according to the collected environmental and operating parameters.

[0007] A sampling method for a mobile rare earth ore sampling drill includes the following steps: S1. Perform power-on initialization, sensor calibration, and fault self-check on the sampling drilling rig, and output system ready signal and basic calibration parameters; S2. Based on the system ready signal and basic calibration parameters output by S1, laser SLAM-IMU-laser ranging fusion positioning is used to achieve field sampling point alignment, and the sampling point alignment completion signal and real-time positioning coordinates are output. S3. Based on the alignment completion signal output by S2, the soil Protodyakonov hardness is detected in real time and the hardness coefficient is output through LIBS-infrared spectral fusion sensing and multi-feature optimization noise reduction processing. S4. Based on the hardness coefficient output by S3 and the preset sampling depth, the BO-LSTM-FS-CBT hybrid model coupled with mechanical formulas is used for adaptive decision-making, outputting the optimal drill bit speed, optimal torque and optimal down-diving speed. S5. Based on the optimal rotational speed, optimal torque, and optimal descent speed output by S4, an anti-integral saturation PID algorithm is used to perform closed-loop control of the drill string parameters and output a stable operating signal for the drill string. S6, based on the stable operating signal output by S5, monitors the drilling status in all dimensions through multiple sensors and performs robust anomaly classification processing, outputting normal drilling commands; S7, based on the normal drilling command output by S6, completes fixed-depth pressure sampling and automatic drill string retraction according to the preset sampling depth, and outputs a sampling completion signal.

[0008] Therefore, the present invention employs the above-mentioned mobile rare earth ore sampling drill and sampling method, which has the following beneficial effects: 1. SLAM-IMU-Laser Fusion Positioning Innovation: The fusion positioning algorithm adopts LiDAR SLAM environment modeling + IMU attitude calculation + laser ranging closed-loop correction. In the absence of GPS, the positioning error is ≤ ±1cm (cm-level accurate positioning) and the XY alignment accuracy is ≤ ±0.5cm, which meets the requirements of high-precision exploration grid and solves the problem of positioning deviation in complex terrain. 2. Innovation in online multi-source spectral hardness detection: Integrating LIBS laser-induced breakdown spectroscopy and infrared spectral sensing, and through Savitzky-Golay (SG) filtering + Pearson correlation + mutual information + variance filtering multi-feature optimization, real-time online detection of soil Protodyakonov hardness is achieved with a detection error of less than 5%, replacing manual input and improving parameter adaptability; 3. Innovative Hybrid Intelligent Model Parameter Decision-Making: A hybrid prediction model of BO-LSTM-FS-CBT is constructed, integrating Bayesian optimization, feature selection, time series prediction, and physical and mechanical formulas. Rotation speed, torque, and descent speed are matched with soil characteristics in real time, reducing drill bit wear by 60% and achieving a sampling success rate of ≥98%. 4. Robust closed-loop control innovation: The full-process control mechanism of SG filtering noise reduction + anti-integral saturation PID + abnormal robust handling is adopted to suppress field noise and actuator saturation, identify and automatically handle abnormalities such as drill blockage, drill breakage, and overload in real time, and reduce equipment failure rate by 90%.

[0009] 5. Automated and efficient: The entire process is carried out autonomously, and the sampling time for a single sample is reduced to less than 3 minutes, increasing exploration efficiency by more than 3 times.

[0010] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of the overall structure of a mobile rare earth ore sampling drill according to the present invention. Figure 2 This is a schematic diagram of the drilling tool structure of a mobile rare earth ore sampling drilling rig according to the present invention. Figure 3 This is a schematic diagram of the XY mobile platform of a mobile rare earth ore sampling drilling rig according to the present invention.

[0012] Figure Labels 1. Drill bit; 11. Lifting assembly; 111. Lead screw; 112. Top fixing plate; 113. Lifting feed drive motor; 114. Chain; 115. Fixing frame; 116. Horizontally arranged bearing; 117. Bottom fixing plate; 118. Slider; 12. Sampling assembly; 121. Sampling drive motor; 122. Bracket; 123. Drill bit; 2. XY moving platform; 21. Moving drive motor; 22. Belt; 23. Mounting plate; 24. Common plate; 25. Moving plate; 26. Vertically arranged bearing; 3. Tracked walking mechanism; 4. Intelligent control module. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the embodiments of the present invention and are not intended to limit the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of this application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.

[0014] It should be noted that the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as a process, method, system, product, or server that includes a series of steps or units, not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or device.

[0015] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0016] like Figures 1-3 As shown, a mobile rare earth ore sampling drilling rig includes a tracked walking mechanism 3, an XY mobile platform 2 mounted on the tracked walking mechanism 3, a drill string 1 mounted on the output end of the XY mobile platform 2, a multi-source sensing component for collecting environmental and operating parameters, and an intelligent control module 4. The multi-source sensing component is electrically connected to the input end of the intelligent control module 4, and the output end of the intelligent control module 4 is electrically connected to the XY mobile platform 2 and the drill string 1 respectively, so as to adjust the operating parameters according to the collected environmental and operating parameters.

[0017] The XY mobile platform 2 includes an X-direction moving component fixed to the chassis of the tracked walking mechanism and a Y-direction moving component connected to the output end of the X-direction moving component. Both the X-direction and Y-direction moving components include a mounting plate 23, a moving drive motor 21 fixed to one end of the mounting plate 23, a belt 22, and a moving plate 25. One end of the belt 22 is connected to the output end of the moving drive motor 21, and the other end of the belt 22 is rotatably mounted on the other end of the mounting plate 23 via a vertically arranged bearing 26. The top end of the belt 22 is fixedly connected to the moving plate 25. In this embodiment, the moving plate of the X-direction moving component and the mounting plate of the Y-direction moving component share a single plate, which is designated as the common plate 24.

[0018] The drill string 1 includes a lifting assembly 11 and a sampling assembly 12 connected to the output end of the lifting assembly 11. The lifting assembly 11 includes a bottom fixing plate 117 fixed to the top of the moving plate 25 of the Y-direction moving assembly, a top fixing plate, a lead screw 111, and a lifting feed motor. The four corners of the top of the bottom fixing plate are fixedly connected to the four corners of the top fixing plate 112 via guide rods. The center of the top of the bottom fixing plate and the center of the bottom of the top fixing plate 112 are also horizontally rotatably connected to the lead screw 111 via a horizontally arranged bearing 116. The bottom end of the lead screw 111 is also connected to a sprocket and a chain. 114 is connected to the lifting feed motor fixed on the bottom fixing plate 117. In this embodiment, the lifting feed drive motor 113 is fixed on the bottom fixing plate 117 via the fixing bracket 115. The sampling assembly 12 includes a bracket 122 that is slidably connected to the guide rod, a sampling drive motor 121 fixed on the bracket 122, and a drill bit 123 fixed to the bottom output end of the sampling drive motor 121. The bracket 122 is also fixedly connected to the nut on the lead screw 111. In this embodiment, the bracket 122 is slidably connected to the two guide rods located on the same side via the slider 118.

[0019] The multi-source sensing components include a lidar sensing unit, an inertial measurement unit, a spectral detection sensing unit, a laser ranging sensing unit, a torque sensing unit, a speed sensing unit, a current sensing unit, a temperature sensing unit, a tilt sensing unit, and a limit sensing unit. Specifically, the lidar sensing unit is used to acquire 3D point cloud data of the field environment, obstacle location information, and terrain feature data; the inertial measurement unit is used to acquire overall acceleration, angular velocity, and heading angle attitude data of the drilling rig; the spectral detection sensing unit is used to acquire laser-induced breakdown spectral data and infrared radiation spectral data of the soil surface; the laser ranging sensing unit is used to acquire drilling bit depth data and XY moving platform displacement deviation data; and the torque sensing unit is used for… The system includes: a real-time output torque data of drill bit 123; a speed sensing unit for collecting rotational speed data of drill bit 123; a current sensing unit for collecting operating current data of lifting feed motor and sampling drive motor 121; a temperature sensing unit for collecting operating temperature data of drill bit 123 and housing temperature data of sampling drive motor 121; an inclination sensing unit for collecting horizontal inclination angle data of XY moving platform 2 and drilling rig attitude tilt data; a limit sensing unit for collecting trigger signal data of upper and lower limit positions of drill bit 123; and a spectral detection sensing unit including a LIBS laser-induced breakdown spectral sensor and an infrared spectral sensor, both coaxially mounted on the front end of bracket 122 of sampling assembly 12 and located above and to the side of drill bit 123.

[0020] It should be noted that the above electronic components are all mature products on the market. This embodiment only requires purchasing them and connecting them according to the instruction manual. No modifications have been made to them. Therefore, their circuit connection structure and principle will not be described in detail here.

[0021] A sampling method for a mobile rare earth ore sampling drill includes the following steps: S1. Perform power-on initialization, sensor calibration, and fault self-check on the sampling drilling rig, and output system ready signal and basic calibration parameters.

[0022] S2. Based on the system readiness signal and basic calibration parameters output by S1, laser SLAM-IMU-laser ranging fusion positioning is used to achieve field sampling point alignment, and the sampling point alignment completion signal and real-time positioning coordinates are output.

[0023] S3, based on the alignment completion signal output by S2, uses LIBS-infrared spectral fusion sensing and multi-feature optimization noise reduction processing to detect the Protodyakonov hardness of the soil in real time and output the hardness coefficient.

[0024] S4. Based on the hardness coefficient output by S3 and the preset sampling depth, the BO-LSTM-FS-CBT hybrid model coupled with mechanical formulas is used for adaptive decision-making, outputting the optimal drill bit speed, optimal torque, and optimal down-diving speed.

[0025] S5, based on the optimal rotational speed, optimal torque and optimal descent speed output by S4, uses an anti-integral saturation PID algorithm to perform closed-loop control of drill string parameters and outputs a stable operating signal for the drill string.

[0026] S6, based on the stable operating signal output by S5, monitors the drilling status in all dimensions through multiple sensors and performs robust anomaly classification processing, outputting normal drilling commands.

[0027] S7, based on the normal drilling command output by S6, completes fixed-depth pressure sampling and automatic drill string retraction according to the preset sampling depth, and outputs a sampling completion signal.

[0028] Step S2 specifically includes the following steps: S21. Synchronously collect three-dimensional point cloud data, inertial motion data, and relative distance data of the field environment. The inertial motion data includes the overall acceleration, angular velocity, and heading angle attitude data of the drilling rig. The relative distance data includes the drilling depth data and the displacement deviation data of the XY moving platform. In this embodiment, the lidar sampling frequency is set to 10Hz, the inertial motion sampling frequency is set to 100Hz, and the laser ranging sampling frequency is set to 50Hz.

[0029] S22. Based on the 3D point cloud data acquired by the LiDAR in S21, the GMAPping algorithm is used to complete point cloud matching and rasterization processing, generating a 5cm×5cm raster map Mgrid of the field operation environment. The initial positioning coordinates of the drilling rig in the raster map are extracted simultaneously. .

[0030] Meanwhile, based on the inertial motion data collected by S21, attitude calculation and dead reckoning are completed using a strapdown inertial navigation algorithm to obtain the real-time calculated coordinates of the drilling rig. and heading angle : ; In the formula, Indicates the initial positioning coordinates of the inertial measurement sensor unit; , These represent the real-time movement speeds of the drilling rig in the x and y directions, respectively. This indicates the z-axis angular velocity of the drilling rig; Indicates the initial heading angle; This indicates the dead reckoning time.

[0031] S23. Set the initial positioning coordinates of the drilling rig in the grid map. Real-time coordinate calculation of drilling rig and laser ranging correction coordinates Input the Kalman filter fusion model and solve for the final fused positioning coordinates of the drilling rig. : ; In the formula, , , Both represent weighting coefficients, and .

[0032] S24. Merge the calculated drilling rig positioning coordinates. This serves as the real-time positioning coordinate of the drilling rig and is compared with the coordinates of the target sampling points in the preset exploration grid. Perform the difference calculation to obtain the real-time alignment error of the x-axis. Real-time alignment error of the y-axis : .

[0033] S25, Based on real-time alignment error along the x-axis Real-time alignment error of the y-axis An incremental PID algorithm is used to drive the motors of the X-direction and Y-direction moving components, which corrects the planar position of the drill bit center in real time until the alignment error converges to the threshold range.

[0034] S26. Output the sampling point alignment completion signal and real-time positioning coordinates.

[0035] Step S3 specifically includes the following steps: S31. Based on the alignment completion signal output by S2, the LIBS laser-induced breakdown spectroscopy sensor and infrared spectroscopy sensor mounted on the drilling rig are triggered to work synchronously, simultaneously acquiring the emission spectrum data of elements in the soil surface. and soil infrared radiation spectral data .

[0036] S32. Obtain the original surface soil elemental emission spectrum data. and soil infrared radiation spectral data Inputting a 3rd-order polynomial and an SG filter module with a window length of 39 respectively, high-frequency noise caused by field dust and equipment vibration is removed, resulting in noise-reduced elemental emission spectrum data of the soil surface. and soil infrared radiation spectral data .

[0037] S33. For the denoised spectral data obtained in S32, Pearson correlation analysis, variance filtering, and mutual information method are used sequentially to complete the triple feature threshold screening (in this embodiment, the triple feature threshold screening strategy is: , The linear correlation coefficient between spectral characteristics and soil hardness is represented. Represents the discrete variance of spectral features. (Representing the mutual information value between spectral features and soil hardness), redundant spectral features with weak correlation to soil hardness are removed, and the high-contribution preferred feature set is retained. .

[0038] S34. Select the optimal feature set Input a pre-trained spectral-hardness regression model, and calculate the predicted value of the Protodyakonov hardness coefficient of the soil through linear weighted fitting. : ; In the formula, This represents the spectral preference feature weight matrix; This represents the bias term in the spectral-hardness regression model.

[0039] S35. Predicted values ​​based on Protodyakonov soil hardness coefficient Compared with the model calibration true value Calculate the confidence level of hardness test : .

[0040] S36. Determine if the condition is met. If the conditions are met, the hardness test data is deemed valid, and the predicted Protodyakonov hardness coefficient of the soil is used. As the final Protodyakonov hardness coefficient of soil If the hardness test is not completed, a hardness test completion signal is generated simultaneously; otherwise, return to step S31.

[0041] Step S4 specifically includes the following steps: S41. The final Protodyakonov hardness coefficient of the soil output in step S3. Preset sampling depth Ambient operating temperature Perform min-max normalization to obtain the normalized decision feature set. .

[0042] S42. Normalize the decision feature set PCA dimensionality reduction was performed to obtain 6-dimensional principal component features. .

[0043] S43. Construct a BO-LSTM-FS-CBT hybrid model. The BO-LSTM-FS-CBT hybrid model includes, in sequence, an input layer, an FS feature selection layer, a CBT deep feature extraction layer, an LSTM temporal prediction layer, a feature fusion layer, a BO optimization layer, and an output layer. The input layer is used to process the principal component features... The data is converted into a 2D temporal tensor adapted to the 2D-CNN input format. The FS feature filtering layer is used to remove redundant features using the triple feature threshold filtering rules described in step S33. The CBT deep feature extraction layer is used to extract nonlinear correlation features of soil-drill string coupling through 2D-CNN multi-dimensional feature extraction, capture the temporal dependencies of the drilling process through BiLSTM temporal dependency, and enhance the weight of key operations through TPA temporal pattern attention weighting. The LSTM temporal prediction layer is used to fit the drilling time series data patterns using the LSTM network to obtain the initial predicted values ​​of rotational speed, torque, and down-drilling speed. The feature fusion layer is used to align and fuse the key temporal features after attention-weighted TPA temporal patterns with the drill string-soil coupled mechanical features, constructing a multi-dimensional joint feature vector that simultaneously includes data-driven features and physical constraint features; the BO optimization layer is used to correct the predicted values; the output layer is used to couple the mechanical formulas and output the optimal parameters that conform to engineering practice. ; These represent the predicted values ​​for drill bit rotation speed, drill bit torque, and descent speed, respectively. These represent the optimal speed, optimal torque, and optimal descent speed, respectively.

[0044] The mechanical formula is expressed as follows: ; In the formula, This represents the theoretically optimal drill bit torque; Indicates the torque coefficient; This represents the coefficient of friction between the drill bit and the soil sidewall; Indicates soil weight.

[0045] S44. Principal component features Input a pre-trained BO-LSTM-FS-CBT hybrid model and output the optimal parameters. .

[0046] S45, will and Perform error verification: determine whether the requirements are met. If so, retain the model parameters; otherwise, use... replace The corresponding rotational speed and downward speed are adjusted synchronously.

[0047] At the same time, the physical matching between rotational speed and downward speed is verified using the correlation formula between rotational speed and downward speed: ; In the formula, This indicates the feed rate per revolution of the drill bit; S46, Output after verification by S45 .

[0048] Step S5 specifically includes the following steps: S51, based on Generate and execute the initial drive preparation instructions for the motor.

[0049] S52. Synchronously acquire real-time drill bit speed during execution. Real-time drill bit torque and real-time downward speed .

[0050] S53, receive from S51 As a set value, collected by S52 As feedback values, the speed error is calculated separately. Torque error , downward speed error : ; Obtain the three-dimensional real-time control error set .

[0051] S54. Set the three-dimensional real-time control error set Input an anti-integral saturation PID controller to calculate the basic control voltage. : ; In the formula, , , These represent the proportional, integral, and differential coefficients, respectively. Represents the integral time variable The three-dimensional real-time control error set at each moment.

[0052] S55, the basic control voltage calculated from S54 Perform actuator saturation detection, calculate saturation error, and complete inverse calculation compensation of integral term: ; ; In the formula, Indicates the control voltage after saturation limiting; Indicates the maximum allowable drive voltage of the motor; Represents a symbolic function; express Constantly control voltage saturation error; This indicates that the integral is used to calculate the compensation gain. This indicates the final motor drive voltage after compensation and correction.

[0053] S56. The final motor drive voltage after compensation and correction obtained in S55. Input sampling drive motor and lifting feed motor to adjust the drill string operation status in real time. Refresh the control cycle every 20ms and determine whether the three-dimensional control error meets the stability threshold condition. If the condition is met, the drill string is determined to be running stably, and a stable drill string operation signal is output synchronously; otherwise, return to step S52.

[0054] In step S6, the anomaly determination formula is as follows: ; in, ; In the formula, Indicates the motor's operating current; Indicates the rated current of the motor; Indicates the surface temperature of the drill bit; This indicates the drill string's temperature exceeding the threshold. Indicates the amplitude of drill string vibration; This indicates that the vibration exceeds the threshold. Indicates the torque overload threshold; Indicates the current overload threshold; Indicates the safe operating temperature; This indicates the threshold for stopping the drilling operation due to blockage.

[0055] When a minor abnormality occurs, increase the torque setting by 10% to maintain continuous drilling; when a moderate abnormality occurs, control the lifting feed motor to raise the drill bit by 5cm, reverse the drill bit for 3 seconds, and then lower it again to continue drilling; when a severe abnormality occurs, immediately cut off the motor drive power, lock the drill bit, and trigger an audible and visual alarm.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A mobile rare earth ore sampling drill, characterized in that: It includes a tracked walking mechanism, an XY mobile platform mounted on the tracked walking mechanism, a drill string mounted on the output end of the XY mobile platform, a multi-source sensing component for collecting environmental and operating parameters, and an intelligent control module. The multi-source sensing component is electrically connected to the input end of the intelligent control module, and the output end of the intelligent control module is electrically connected to the XY mobile platform and the drill string respectively, so as to adjust the operating parameters according to the collected environmental and operating parameters.

2. The mobile rare earth ore sampling drill according to claim 1, characterized in that: The XY mobile platform includes an X-direction moving component fixed to the chassis of the tracked walking mechanism and a Y-direction moving component connected to the output end of the X-direction moving component. Both the X-direction moving component and the Y-direction moving component include a mounting plate, a moving drive motor fixed to one end of the mounting plate, a belt, and a moving plate. One end of the belt is connected to the output end of the moving drive motor, and the other end of the belt is rotatably mounted on the other end of the mounting plate via a vertically arranged bearing. The top end of the belt is fixedly connected to the moving plate.

3. A mobile rare earth ore sampling drill according to claim 2, characterized in that: The drilling tool includes a lifting assembly and a sampling assembly connected to the output end of the lifting assembly. The lifting assembly includes a bottom fixed plate, a top fixed plate, a lead screw, and a lifting feed motor, all fixed to the top of the moving plate of the Y-direction moving assembly. The top four corners of the bottom fixed plate are fixedly connected to the top four corners of the top fixed plate via guide rods. A lead screw is also horizontally rotatably connected between the top center of the bottom fixed plate and the bottom center of the top fixed plate. The bottom end of the lead screw is also connected to the lifting feed motor fixed to the bottom fixed plate via a sprocket and a chain. The sampling assembly includes a bracket that is slidably connected to the guide rod, a sampling drive motor fixed on the bracket, and a drill bit fixed to the bottom output end of the sampling drive motor. The bracket is also fixedly connected to the nut on the lead screw.

4. A mobile rare earth ore sampling drill according to claim 3, characterized in that: The multi-source sensing components include a lidar sensing unit, an inertial measurement sensing unit, a spectral detection sensing unit, a laser ranging sensing unit, a torque sensing unit, a speed sensing unit, a current sensing unit, a temperature sensing unit, a tilt sensing unit, and a limit sensing unit. The system includes: a lidar sensing unit for collecting 3D point cloud data of the field environment, obstacle location information, and terrain feature data; an inertial measurement unit for collecting overall acceleration, angular velocity, and heading angle attitude data of the drilling rig; a spectral detection sensing unit for collecting laser-induced breakdown spectral data and infrared radiation spectral data of the soil surface; a laser ranging sensing unit for collecting drill bit depth data and XY moving platform displacement deviation data; a torque sensing unit for collecting real-time output torque data of the drill bit; a speed sensing unit for collecting drill bit rotation speed data; a current sensing unit for collecting operating current data of the lifting feed motor and sampling drive motor; a temperature sensing unit for collecting drill bit operating temperature and sampling drive motor housing temperature data; a tilt sensing unit for collecting horizontal tilt angle of the XY moving platform and drilling rig attitude tilt data; and a limit sensing unit for collecting trigger signal data of the upper and lower limit positions of the drill bit lifting. The spectral detection sensing unit includes a LIBS laser-induced breakdown spectral sensor and an infrared spectral sensor, both of which are coaxially mounted on the front end of the sampling assembly's bracket and located above the drill bit.

5. The sampling method of a mobile rare earth ore sampling drill as described in claim 4, characterized in that: Includes the following steps: S1. Perform power-on initialization, sensor calibration, and fault self-check on the sampling drilling rig, and output system ready signal and basic calibration parameters; S2. Based on the system ready signal and basic calibration parameters output by S1, laser SLAM-IMU-laser ranging fusion positioning is used to achieve field sampling point alignment, and the sampling point alignment completion signal and real-time positioning coordinates are output. S3. Based on the alignment completion signal output by S2, the soil Protodyakonov hardness is detected in real time and the hardness coefficient is output through LIBS-infrared spectral fusion sensing and multi-feature optimization noise reduction processing. S4. Based on the hardness coefficient output by S3 and the preset sampling depth, the BO-LSTM-FS-CBT hybrid model coupled with mechanical formulas is used for adaptive decision-making, outputting the optimal drill bit speed, optimal torque and optimal down-diving speed. S5. Based on the optimal rotational speed, optimal torque, and optimal descent speed output by S4, an anti-integral saturation PID algorithm is used to perform closed-loop control of the drill string parameters and output a stable operating signal for the drill string. S6, based on the stable operating signal output by S5, monitors the drilling status in all dimensions through multiple sensors and performs robust anomaly classification processing, outputting normal drilling commands; S7, based on the normal drilling command output by S6, completes fixed-depth pressure sampling and automatic drill string retraction according to the preset sampling depth, and outputs a sampling completion signal.

6. The sampling method of a mobile rare earth ore sampling drill according to claim 5, characterized in that: Step S2 specifically includes the following steps: S21. Synchronously collect three-dimensional point cloud data, inertial motion data, and relative distance data of the field environment. The inertial motion data includes the overall acceleration, angular velocity, and heading angle attitude data of the drilling rig. The relative distance data includes the drill bit depth data and the displacement deviation data of the XY moving platform. S22. Based on the 3D point cloud data acquired by the LiDAR in S21, the GMAPping algorithm is used to complete point cloud matching and rasterization processing, generating a 5cm×5cm raster map Mgrid of the field operation environment. The initial positioning coordinates of the drilling rig in the raster map are extracted simultaneously. ; Meanwhile, based on the inertial motion data collected by S21, attitude calculation and dead reckoning are completed using a strapdown inertial navigation algorithm to obtain the real-time calculated coordinates of the drilling rig. and heading angle : ； In the formula, Indicates the initial positioning coordinates of the inertial measurement sensor unit; , These represent the real-time movement speeds of the drilling rig in the x and y directions, respectively. This indicates the z-axis angular velocity of the drilling rig; Indicates the initial heading angle; Indicates dead reckoning time; S23. Set the initial positioning coordinates of the drilling rig in the grid map. Real-time coordinate calculation of drilling rig and laser ranging correction coordinates Input the Kalman filter fusion model and solve for the final fused positioning coordinates of the drilling rig. : ； In the formula, , , Both represent weighting coefficients, and ; S24. Merge the calculated drilling rig positioning coordinates. This serves as the real-time positioning coordinate of the drilling rig and is compared with the coordinates of the target sampling points in the preset exploration grid. Perform the difference calculation to obtain the real-time alignment error of the x-axis. Real-time alignment error of the y-axis : ； S25, Based on real-time alignment error along the x-axis Real-time alignment error of the y-axis An incremental PID algorithm is used to drive the motors of the X-direction and Y-direction moving components to correct the planar position of the drill bit center in real time until the alignment error converges to the threshold range. S26. Output the sampling point alignment completion signal and real-time positioning coordinates.

7. The sampling method of a mobile rare earth ore sampling drill according to claim 6, characterized in that: Step S3 specifically includes the following steps: S31. Based on the alignment completion signal output by S2, the LIBS laser-induced breakdown spectroscopy sensor and infrared spectroscopy sensor mounted on the drilling rig are triggered to work synchronously, simultaneously acquiring the emission spectrum data of elements in the soil surface. and soil infrared radiation spectral data ; S32. Obtain the original surface soil elemental emission spectrum data. and soil infrared radiation spectral data Inputting a 3rd-order polynomial and an SG filter module with a window length of 39 respectively, high-frequency noise caused by field dust and equipment vibration is removed, resulting in noise-reduced elemental emission spectrum data of the soil surface. and soil infrared radiation spectral data ; S33. For the denoised spectral data obtained in S32, Pearson correlation analysis, variance filtering, and mutual information method are used sequentially to complete triple feature threshold screening, eliminating redundant spectral features that are weakly correlated with soil hardness and retaining the high-contribution optimized feature set. ; S34. Select the optimal feature set Input a pre-trained spectral-hardness regression model, and calculate the predicted value of the Protodyakonov hardness coefficient of the soil through linear weighted fitting. : ； In the formula, This represents the spectral preference feature weight matrix; This represents the bias term in the spectral-hardness regression model; S35. Predicted values ​​based on Protodyakonov soil hardness coefficient Compared with the model calibration true value Calculate the confidence level of hardness test : ； S36. Determine if the condition is met. If the conditions are met, the hardness test data is deemed valid, and the predicted Protodyakonov hardness coefficient of the soil is used. As the final Protodyakonov hardness coefficient of soil If the hardness test is not completed, a hardness test completion signal is generated simultaneously; otherwise, return to step S31.

8. The sampling method of a mobile rare earth ore sampling drill according to claim 7, characterized in that: Step S4 specifically includes the following steps: S41. The final Protodyakonov hardness coefficient of the soil output in step S3. Preset sampling depth Ambient operating temperature Perform min-max normalization to obtain the normalized decision feature set. ; S42. Normalize the decision feature set PCA dimensionality reduction was performed to obtain 6-dimensional principal component features. ; S43. Construct a BO-LSTM-FS-CBT hybrid model. The BO-LSTM-FS-CBT hybrid model includes, in sequence, an input layer, an FS feature selection layer, a CBT deep feature extraction layer, an LSTM temporal prediction layer, a feature fusion layer, a BO optimization layer, and an output layer. The input layer is used to process the principal component features... The data is converted into a 2D temporal tensor adapted to the 2D-CNN input format. The FS feature filtering layer is used to remove redundant features using the triple feature threshold filtering rules described in step S33. The CBT deep feature extraction layer is used to extract nonlinear correlation features of soil-drill string coupling through 2D-CNN multi-dimensional feature extraction, capture the temporal dependencies of the drilling process through BiLSTM temporal dependency, and enhance the weight of key operations through TPA temporal pattern attention weighting. The LSTM temporal prediction layer is used to fit the drilling time series data patterns using the LSTM network to obtain the initial predicted values ​​of rotational speed, torque, and down-drilling speed. The feature fusion layer is used to align and fuse the key temporal features after attention-weighted TPA temporal patterns with the drill string-soil coupled mechanical features, constructing a multi-dimensional joint feature vector that simultaneously includes data-driven features and physical constraint features; the BO optimization layer is used to correct the predicted values; the output layer is used to couple the mechanical formulas and output the optimal parameters that conform to engineering practice. ; These represent the predicted values ​​for drill bit rotation speed, drill bit torque, and descent speed, respectively. These represent the optimal speed, optimal torque, and optimal descent speed, respectively. The mechanical formula is expressed as follows: ； In the formula, This represents the theoretically optimal drill bit torque; Indicates the torque coefficient; This represents the coefficient of friction between the drill bit and the soil sidewall; Indicates soil weight; S44. Principal component features Input a pre-trained BO-LSTM-FS-CBT hybrid model and output the optimal parameters. ; S45, will and Perform error verification: determine whether the requirements are met. If so, retain the model parameters; otherwise, use... replace Simultaneously correct the corresponding rotational speed and downward speed; At the same time, the physical matching between rotational speed and downward speed is verified using the correlation formula between rotational speed and downward speed: ； In the formula, This indicates the feed rate per revolution of the drill bit; S46, Output after verification by S45 .

9. The sampling method of a mobile rare earth ore sampling drill according to claim 8, characterized in that: Step S5 specifically includes the following steps: S51, based on Generate and execute the initial drive preparation instructions for the motor; S52. Synchronously acquire real-time drill bit speed during execution. Real-time drill bit torque and real-time downward speed ; S53, receive from S51 As a set value, collected by S52 As feedback values, the speed error is calculated separately. Torque error , downward speed error : ； Obtain the three-dimensional real-time control error set ; S54. Set the three-dimensional real-time control error set Input an anti-integral saturation PID controller to calculate the basic control voltage. : ； In the formula, , , These represent the proportional, integral, and differential coefficients, respectively. Represents the integral time variable The three-dimensional real-time control error set at each moment; S55, the basic control voltage calculated from S54 Perform actuator saturation detection, calculate saturation error, and complete inverse calculation compensation of integral term: ； ； In the formula, Indicates the control voltage after saturation limiting; Indicates the maximum allowable drive voltage of the motor; Represents a symbolic function; express Constantly control voltage saturation error; This indicates that the integral is used to calculate the compensation gain. This indicates the final motor drive voltage after compensation and correction. S56. The final motor drive voltage after compensation and correction obtained in S55. Input sampling drive motor and lifting feed motor to adjust the drill string operation status in real time. Refresh the control cycle every 20ms and determine whether the three-dimensional control error meets the stability threshold condition. If the condition is met, the drill string is determined to be running stably, and a stable drill string operation signal is output synchronously; otherwise, return to step S52.

10. The sampling method of a mobile rare earth ore sampling drill according to claim 9, characterized in that: In step S6, the anomaly determination formula is as follows: ； in, ； In the formula, Indicates the motor's operating current; Indicates the rated current of the motor; Indicates the surface temperature of the drill bit; This indicates the drill string's temperature exceeding the threshold. Indicates the amplitude of drill string vibration; This indicates that the vibration exceeds the threshold. Indicates the torque overload threshold; Indicates the current overload threshold; Indicates the safe operating temperature; Indicates the threshold for stopping the drilling operation due to blockage; When a minor anomaly occurs, increase the torque setting by 10% to maintain continuous drilling; When a moderate abnormality occurs, control the lifting feed motor to raise the drill bit by 5cm, and after the drill bit reverses for 3 seconds, resume drilling. In the event of a severe abnormality, immediately disconnect the motor drive power, lock the drill bit, and trigger an audible and visual alarm.

Images