A coal rock cutting dust production experimental system and method based on real-time correlation of cracks and dust

Abstract

The application discloses a coal rock cutting dust production experimental system and method based on real-time correlation of cracks and dust, and belongs to the technical field of coal rock cutting experiments. The system comprises a coal rock sample, a cutting simulation device, a crack evolution data acquisition unit, a dust and wind speed monitoring unit, a synchronous data acquisition module, a data processing and analysis module and a visual display module. The application realizes millisecond-level time alignment of crack and dust data through FPGA hardware synchronous triggering and dust transmission delay compensation, combines resistivity tomography of conductive similar materials, acoustic emission array and multi-modal crack monitoring of high-speed photography, and establishes a real-time correlation between crack evolution and dust production. The application first realizes the space-time correspondence between internal cracks and external dust in coal rock cutting, quantitatively reveals the influence of different crack types on dust particle size distribution and dust concentration, and provides an experimental basis for coal mine dust source control and cutting parameter optimization.

Description

Technical Field

[0001] This invention belongs to the field of coal and rock cutting experimental technology, specifically relating to a coal and rock cutting dust generation experimental system and method based on real-time correlation between cracks and dust. Background Technology

[0002] In fully mechanized mining and tunneling faces of coal mines, the cutting of coal and rock by mining machinery is the primary source of dust generation underground. High concentrations of dust not only seriously endanger the respiratory health of workers but also pose significant safety hazards such as coal dust explosions. A thorough understanding of the dust generation mechanism during cutting and the establishment of a causal relationship between internal damage evolution and external dust generation are crucial prerequisites for achieving dust source control.

[0003] Currently, various experimental systems exist for simulating dust generation during coal and rock mass cutting. For example, Chinese patent application CN110006771A discloses a simulation experimental system for dust generation during coal and rock mass cutting, capable of controlling cutting parameters, applying ground stress, switching cutting tooth types, and using mechanical monitoring devices and dust monitoring devices (such as high-speed cameras and phase Doppler interferometers) to collect mechanical parameters and data on dust concentration, particle size, and velocity during the cutting process. Chinese patent application CN114544440A proposes a cutting dust generation experimental system that uses a sliding mechanism, a dust collection module, and a dynamic collection module to collect and visualize dust during the cutting process. Chinese patent application CN116124692A discloses a mining cutting dust generation condition investigation system, which designs a triaxial confining pressure clamping module, a closed dust collection hood, and a multi-stage inertial drainage dust collection pipeline, and adds a coal and rock mass crack monitoring camera to observe surface cracks.

[0004] However, the existing experimental systems mentioned above all focus on controlling cutting parameters, dust collection and measurement, or single optical observation of surface cracks, and none of them have set up a real-time monitoring unit for the initiation and propagation process of internal cracks in coal and rock samples. Specifically: First, the existing systems cannot simultaneously obtain the initiation location, propagation path, and damage accumulation degree of internal microcracks in the coal and rock mass during the cutting process. This makes it difficult to monitor the direct source of dust generation—the evolution of internal damage in the coal and rock mass—and makes it impossible to establish a causal relationship between internal damage and external dust generation. Second, although resistivity tomography (ERT) and acoustic emission (AE) technologies have been applied to monitoring mining-induced fractures and identifying precursors of coal and rock instability, respectively, existing applications are limited to quasi-static loading (such as the groundwater migration simulation platform for coal mining disclosed in Chinese patent application CN117268892A) or specific dynamic loading scenarios (such as the blasting method for hard rock faults in fully mechanized tunneling faces disclosed in Chinese patent application CN114577078A), and have not yet been adapted to high-speed dynamic cutting loading processes. Moreover, conventional coal and rock materials have poor conductivity, and their resistivity response sensitivity to micro-cracks is insufficient to meet the real-time acquisition requirements of the cutting process. Third, and more critically, in existing technologies, crack monitoring (if any) and dust monitoring are completely disconnected: dust often experiences a transmission delay of hundreds of milliseconds to several seconds from the cutting area to the sensor via the airflow transmission pipe, and existing systems lack hardware synchronization triggering mechanisms and pneumatic transmission delay compensation algorithms, resulting in a misalignment between crack evolution data and dust concentration data on the time axis, making it impossible to establish a real-time correlation between the two. Furthermore, the existing solutions do not achieve a unified coordinate matching between the crack location and the dust source in the spatial dimension.

[0005] In summary, existing technologies are unable to reveal the core mechanism by which internal crack evolution drives external dust generation during coal and rock cutting under a unified spatiotemporal coordinate system. There is an urgent need for a system and method that can realize real-time correlation and inversion between crack evolution and dust generation. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an experimental system and method for dust generation in coal and rock cutting based on real-time correlation between cracks and dust. By using hardware synchronous triggering and dust transmission delay compensation, millisecond-level spatiotemporal alignment between crack evolution and dust generation is achieved. Combined with multimodal crack monitoring and quantitative correlation modeling, this invention reveals for the first time the real-time correspondence between internal cracks driving external dust generation in coal and rock cutting. This fills the technical gap in the study of the causal relationship between internal damage and external dust generation in the dust generation mechanism of coal and rock cutting, and has significant scientific value and engineering application prospects.

[0007] This invention is achieved through the following technical solution:

[0008] An experimental system for dust generation during coal and rock cutting based on real-time crack-dust correlation includes:

[0009] Coal and rock samples are used to simulate the object to be cut.

[0010] A cutting simulation device is used to cut the coal and rock sample.

[0011] The crack evolution data acquisition unit is used to acquire crack evolution data in real time, reflecting the crack initiation and propagation process inside coal and rock samples.

[0012] The dust and wind speed monitoring unit is used to collect real-time data on dust concentration and / or dust particle size distribution and wind speed generated during cutting.

[0013] A synchronous data acquisition module is connected to the crack evolution data acquisition unit and the dust and wind speed monitoring unit to realize the synchronous triggering and acquisition of all data; the synchronous data acquisition module uses a field-programmable gate array as the main controller to generate a unified sampling clock and trigger signal;

[0014] The data processing and analysis module is connected to the synchronous data acquisition module and is used to execute experimental method steps;

[0015] The visualization module, connected to the data processing and analysis module, is used to output the correspondence between the crack evolution process and the dust generation characteristics.

[0016] Preferably, the coal-rock sample is a conductive similar material, which is formed by uniformly mixing a coal-rock similar material matrix with a conductive filler; the conductive filler is one or more of carbon nanotubes, carbon fibers, graphene, carbon black, steel slag, or conductive polymer materials, with a dosage range of 0.5 wt% to 10 wt%.

[0017] The crack evolution data acquisition unit includes:

[0018] An electrode array arranged in a preset network inside or on the surface of the coal and rock sample, and a resistivity measuring device connected to the electrode array, are used to collect resistivity distribution data in real time to identify internal cracks.

[0019] An acoustic emission sensor array arranged on the surface of the coal and rock sample, and an acoustic emission data acquisition device connected to the acoustic emission sensor array, are used to acquire micro-fracture event data in real time.

[0020] A high-speed camera used to acquire image sequences of surface cracks in the coal and rock samples.

[0021] Preferably, it also includes a sealed experimental chamber for accommodating the coal and rock sample, the cutting simulation device, the crack evolution data acquisition unit, and the dust and wind speed monitoring unit; the sealed experimental chamber is provided with an air inlet and an air outlet, the air outlet is connected to an airflow transmission pipe, and the dust and wind speed monitoring unit is installed at the air outlet of the airflow transmission pipe;

[0022] The dust and wind speed monitoring unit includes a dust concentration meter, a dust particle counter, an online laser particle size analyzer, and a wind speed sensor.

[0023] Preferably, the data processing and analysis module has a built-in deep neural network model for establishing a nonlinear mapping relationship between crack evolution characteristics and dust generation characteristics; the crack evolution characteristics include one or more of crack propagation rate, crack length, acoustic emission energy rate, and resistivity anomaly zone volume; the dust generation characteristics include dust concentration and dust particle size distribution.

[0024] An experimental method for dust generation during coal and rock cutting based on real-time crack-dust correlation, the method being based on the aforementioned experimental system, includes the following steps:

[0025] Step 1) During the cutting of the coal and rock sample, multi-source monitoring data is collected synchronously through the synchronous data acquisition module. The multi-source monitoring data includes: crack evolution data reflecting the crack initiation and propagation process inside the coal and rock sample, and dust concentration data and / or dust particle size distribution data reflecting the characteristics of cutting dust generation.

[0026] Step 2) Synchronize and match the crack evolution data and dust data collected in Step 1) in time and space to establish a real-time correlation between crack evolution and dust generation;

[0027] Step 3) Based on the real-time correlation established in Step 2), output the correspondence between the crack evolution process and the dust generation characteristics.

[0028] Preferably, the crack evolution data in step 1) is acquired by an array of acoustic emission sensors arranged on the surface of the coal and rock sample. The acoustic emission waveform data is used for real-time localization of micro-fracture events and crack type identification. The crack type is identified by the rise angle RA and average frequency AF of the acoustic emission signal: if AF is lower than a preset threshold and RA is higher than a preset threshold, it is identified as a tension crack; if AF is higher than a preset threshold and RA is lower than a preset threshold, it is identified as a shear crack.

[0029] The crack evolution data also includes: a sequence of surface crack images of coal and rock samples acquired by a high-speed camera, used for digital image correlation analysis to extract surface crack propagation information, the surface crack propagation information including crack length and crack propagation speed, and compared and verified with the internal crack results obtained by resistivity inversion;

[0030] The multi-source monitoring data also includes: force data of the cutting tooth collected by a triaxial force sensor installed at the root of the cutting tooth, which is used to help determine the load changes during crack propagation.

[0031] Preferably, the specific method for time synchronization in step 2) is as follows: all sensors use the same hardware trigger signal for synchronous acquisition to ensure that the time alignment accuracy of multi-source data reaches the microsecond level; the dust generated during cutting is transported from the cutting area to the dust and wind speed monitoring unit installed at the air outlet of the airflow transmission pipe through the airflow transmission pipe; the airflow speed is measured in real time by the wind speed sensor set in the airflow transmission pipe; according to the time delay Δt of the dust being transported from the cutting area to the dust and wind speed monitoring unit, the timestamp of the dust data is time-aligned with the crack evolution data, and the time delay Δt is determined by the formula Δt=L / v(t), where L is the length of the airflow transmission pipe and v(t) is the airflow speed measured in real time by the wind speed sensor;

[0032] The specific method for spatial matching is as follows: a unified sample coordinate system is established, and the spatial coordinates of crack evolution data and the spatial positions of dust sampling points are uniformly transformed to the same coordinate system to achieve spatial correspondence between crack positions and dust sources; the spatial coordinates of crack evolution data include the three-dimensional coordinates of internal cracks obtained by resistivity tomography inversion, the spatial coordinates of micro-fracture events obtained by acoustic emission time difference positioning, and the surface crack propagation coordinates obtained by high-speed camera combined with digital image correlation analysis.

[0033] Preferably, it also includes a combined acoustic emission-resistivity monitoring step, as follows:

[0034] The acoustic emission localization results and resistivity imaging results are fused in a unified spatiotemporal coordinate system to identify the spatial correspondence between areas of dense acoustic emission events and areas of resistivity anomalies.

[0035] Preferably, step 3) specifically includes:

[0036] Identify the crack evolution stages corresponding to dust generation rate and intensity to determine the main timing of dust generation;

[0037] The correspondence between different crack types and dust particle size distribution was analyzed, and the crack types included tensile cracks and shear cracks.

[0038] Establish a quantitative relationship model between crack propagation rate, crack length, crack width and dust generation.

[0039] Preferably, it also includes a variable influence analysis step, as follows:

[0040] Repeat steps 1) to 3) with different cutting parameters to analyze the changing relationship between crack evolution and dust generation under different cutting parameters; the cutting parameters include one or more of the following: cutting speed, traction speed, cutting tooth angle, and cutting tooth arrangement.

[0041] Repeat steps 1) to 3) with different coal and rock samples to analyze the changing relationship between crack evolution and dust generation in different coal and rock samples; the differences between the various coal and rock samples lie in one or more of the mechanical properties, pore structure, degree of coalification or rock type; the mechanical properties include one or more of hardness, brittleness and strength.

[0042] The beneficial effects of this invention are as follows:

[0043] (1) By adopting FPGA hardware synchronous triggering technology and combining it with a dynamic compensation algorithm for dust transport delay based on the length of the airflow transmission pipe and the real-time wind speed (Δt = L / v(t)), the precise spatiotemporal alignment of internal coal and rock crack evolution data and external dust generation data was achieved for the first time on a millisecond time scale. This effectively solved the technical problem of the misalignment of crack monitoring and dust monitoring on the time axis in traditional experimental systems, and provided a reliable data foundation for revealing the micromechanical mechanism of dust generation from cutting.

[0044] (2) This invention comprehensively employs three monitoring methods—conductive similar materials, resistivity tomography, acoustic emission array, and high-speed imaging—to achieve full-scale capture from internal damage perspective images of coal and rock to surface crack propagation fields. Specifically, the conductive similar materials, by incorporating conductive fillers such as carbon nanotubes, significantly improve the resistivity response sensitivity to microcracks, meeting the real-time acquisition requirements of the cutting process; the acoustic emission array is used for real-time localization of micro-fracture events and crack type identification; and high-speed imaging combined with digital image correlation analysis is used to extract surface crack propagation information. The three monitoring methods cross-validate each other, significantly improving the accuracy and confidence of crack evolution characterization.

[0045] (3) This invention can identify tension cracks and shear cracks in real time by using the rise angle (RA) and average frequency (AF) parameters of the acoustic emission signal. Based on this, a quantitative correspondence between different crack types and dust particle size distribution (such as PM10 proportion) can be established, revealing the mechanism by which tension cracks are prone to generating fine dust and shear cracks are prone to generating coarse dust, providing a scientific basis for dust source control from the perspective of crack control.

[0046] (4) The system of this invention incorporates a deep neural network model, which takes crack evolution parameters such as crack propagation rate, crack length, acoustic emission energy rate, and resistivity anomaly zone volume as inputs, and dust concentration and particle size distribution as outputs, to establish a nonlinear mapping relationship between crack evolution and dust generation. Through comparative experiments on different cutting parameters and different coal and rock samples, the influence of factors such as cutting speed, traction speed, cutting tooth angle, and coal and rock mechanical properties and pore structure on the crack-dust correlation can be quantitatively analyzed, providing experimental basis for coal mine dust source control and cutting process parameter optimization.

[0047] (5) This invention establishes a unified sample coordinate system, transforming the three-dimensional coordinates of resistivity imaging, acoustic emission positioning coordinates, and two-dimensional coordinates of surface images into the same coordinate system. This achieves the spatial correspondence between crack locations and dust sampling points, enabling monitoring data from different sources and dimensions to be fused, analyzed, and visualized within a unified spatial framework. This facilitates an intuitive understanding of the correspondence between crack spatial distribution and dust generation areas. By fusing acoustic emission positioning results and resistivity imaging results in a unified spatiotemporal coordinate system, the spatial correspondence between areas of dense acoustic emission events and resistivity anomalies is identified, achieving mutual verification between internal micro-fracture events and resistivity anomalies. This further improves the reliability and accuracy of crack evolution monitoring.

[0048] (6) By changing different cutting parameters (cutting speed, traction speed, cutting tooth angle, cutting tooth arrangement) and different coal and rock samples (different hardness, brittleness, coal type), the present invention can systematically analyze the influence law of each factor on the crack-dust correlation, and provide a comprehensive experimental basis for coal mine dust source control and cutting process parameter optimization. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the module structure of the coal and rock cutting dust generation experimental system based on real-time crack-dust correlation in Example 1;

[0050] Figure 2 This is a data processing flowchart of the coal and rock cutting dust generation experimental method based on real-time correlation between cracks and dust in Example 2. Detailed Implementation

[0051] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0052] Example 1

[0053] This embodiment provides an experimental system for dust generation during coal and rock cutting based on real-time correlation between cracks and dust. For example... Figure 1As shown, the system includes: coal and rock samples, cutting simulation device, crack evolution data acquisition unit, dust and wind speed monitoring unit, synchronous data acquisition module, data processing and analysis module, and visualization display module.

[0054] 1. Coal and rock samples

[0055] Coal and rock samples were prepared using a conductive similar material to simulate cutting objects with different mechanical properties. The conductive similar material consisted of a coal and rock similar material matrix and a conductive filler, uniformly mixed. In this embodiment, taking medium-hardness bituminous coal as an example, the similar material matrix was prepared with a cement:gypsum:coal powder mass ratio of 1:1:2 and a water-cement ratio of 0.4 to simulate the mechanical properties of medium-hardness bituminous coal. Carbon nanotubes were used as the conductive filler, with a dosage of 2% of the total mass of cement and gypsum. The carbon nanotubes were pre-dispersed ultrasonically in a sodium dodecylbenzenesulfonate solution for 30 min, then mixed and stirred with the matrix material, and poured into a 300 mm × 300 mm × 300 mm mold. During pouring, a stainless steel electrode array was pre-embedded at 100 mm intervals in a 3 × 3 × 3 three-dimensional grid arrangement, with electrode leads extending from the side of the sample. The samples were cured at room temperature for 28 days before use. The conductive filler can also be one or more of carbon fiber, graphene, carbon black, steel slag, or conductive polymer materials, with a dosage range of 0.5% to 10% (mass fraction). The spacing between adjacent electrodes in the electrode array is 50 to 200 mm.

[0056] 2. Cutting simulation device

[0057] The cutting simulation device is used to cut coal and rock samples. In this embodiment, the cutting simulation device includes a cutting motor, cutting teeth, a traction mechanism, and a feeding mechanism. The cutting motor drives the cutting teeth to rotate; the traction mechanism drives the cutting teeth to move laterally along the sample surface; and the feeding mechanism adjusts the cutting depth of the cutting teeth. The device can perform cutting at a set cutting speed (e.g., 3 m / s), traction speed (e.g., 2 m / min), and cutting depth (e.g., 10 mm). The type of cutting teeth can be changed according to experimental needs, including pick-shaped cutting teeth and knife-shaped cutting teeth. Pick-shaped cutting teeth are mainly used to simulate point load coal breaking conditions, while knife-shaped cutting teeth are used to simulate line load coal cutting conditions. The cutting teeth can be arranged sequentially or staggered. Sequential arrangement means that the cutting teeth are arranged at equal intervals along the circumference of the drum, with one cutting tooth on each cutting line; staggered arrangement means that the cutting teeth on adjacent cutting lines are staggered at a certain angle in the circumferential direction, forming a staggered distribution. A triaxial force sensor is installed at the root of the cutting teeth to collect the force data of the cutting teeth in real time.

[0058] 3. Crack evolution data acquisition unit

[0059] The crack evolution data acquisition unit is used to acquire crack evolution data reflecting the crack initiation and propagation process inside coal and rock samples in real time. It consists of three parts:

[0060] (1) Resistivity Imaging Monitoring Subunit

[0061] The system includes an electrode array arranged in a preset network inside or on the surface of a coal or rock sample, and a resistivity measuring device connected to the electrode array. In this embodiment, the electrode array is a 3×3×3 three-dimensional grid with a spacing of 100 mm between adjacent electrodes. The resistivity measuring device uses a multi-channel resistivity acquisition instrument with a sampling frequency of 50 Hz. By injecting current into adjacent electrode pairs and measuring the potential difference, apparent resistivity data is obtained. Then, an inversion algorithm (such as the Gauss-Newton method) is used to reconstruct the resistivity distribution image inside the sample. When a crack initiates or propagates, the resistivity in the crack area increases significantly (e.g., a resistivity change exceeding 20% ​​is considered crack initiation), thereby identifying the crack initiation location, propagation path, and degree of damage accumulation.

[0062] (2) Acoustic emission monitoring subunit

[0063] The system includes an acoustic emission sensor array arranged on the surface of a coal and rock sample, and an acoustic emission data acquisition device connected to the array. In this embodiment, the acoustic emission sensor array has 8 channels, is attached to the sample surface in a triangular geometric layout, and has a sampling frequency of 2 MHz. The acoustic emission waveform data is used for real-time localization of micro-fracture events (using a time-difference localization algorithm) and crack type identification. The crack type is identified by the rise angle (RA) and average frequency (AF) of the acoustic emission signal: if AF is lower than a preset threshold (e.g., 100 kHz) and RA is higher than a preset threshold (e.g., 5 ms / V), it is identified as a tension crack; if AF is higher than the preset threshold and RA is lower than the preset threshold, it is identified as a shear crack. The acoustic emission sensor array can also have 6-12 channels, arranged in an inverted triangular geometric layout, with a sampling frequency of not less than 1 MHz.

[0064] (3) Optical monitoring subunit

[0065] A high-speed camera, positioned directly above the coal and rock sample surface, is used to acquire a sequence of surface crack images. In this embodiment, the high-speed camera operates at a frame rate of 5000 fps, and a speckle pattern projector is mounted in front of the lens. The acquired image sequence is analyzed using digital image correlation (DIC) to extract surface crack propagation information, including crack length, crack width, and crack propagation velocity. The DIC analysis results are compared and verified with the internal crack results obtained through resistivity inversion to improve the confidence level of crack evolution characterization.

[0066] 4. Dust and wind speed monitoring unit

[0067] The dust and wind speed monitoring unit is used to collect real-time data on dust concentration and / or dust particle size distribution and wind speed generated during cutting. In this embodiment, a sealed experimental chamber is also included to house the coal and rock sample, the cutting simulation device, the crack evolution data acquisition unit, and the dust and wind speed monitoring unit. The sealed experimental chamber has an air inlet and an air outlet, with the outlet connected to an airflow transmission duct. The dust and wind speed monitoring unit is installed at the air outlet of the airflow transmission duct and specifically includes: a dust concentration meter (for measuring total dust concentration), a dust particle counter or online laser particle size analyzer (for measuring dust particle size distribution, such as the proportion of PM10 and PM2.5), and a hot-wire anemometer (for real-time measurement of airflow velocity). The airflow transmission duct is L = 2.0 m long, and the anemometer sampling frequency is 10 Hz.

[0068] 5. Synchronous Data Acquisition Module

[0069] The synchronous data acquisition module is connected to the crack evolution data acquisition unit and the dust and wind speed monitoring unit to achieve synchronous triggering and acquisition of all data. In this embodiment, the synchronous data acquisition module uses a field-programmable gate array (FPGA) as the main controller to generate a unified sampling clock and trigger signal. The FPGA generates an external trigger signal and simultaneously connects to the resistivity acquisition instrument, acoustic emission acquisition card, high-speed camera, dust sensor, and wind speed sensor to ensure that the start timestamp error of each channel is less than 1 μs.

[0070] 6. Data Processing and Analysis Module

[0071] The data processing and analysis module is connected to the synchronous data acquisition module and is used to execute the experimental method steps. In this embodiment, the data processing and analysis module is a high-performance computer with built-in data preprocessing software, resistivity inversion software, acoustic emission localization and crack type identification software, and a deep neural network model.

[0072] A deep neural network model is used to establish a nonlinear mapping relationship between crack evolution characteristics (crack propagation speed, crack length, acoustic emission energy rate, resistivity anomaly zone volume, etc.) and dust generation characteristics (dust concentration, particle size distribution). In this embodiment, the model adopts a fully connected feedforward neural network architecture, specifically configured as follows: the input layer contains neurons corresponding to the crack evolution characteristics (crack propagation speed, crack length, acoustic emission energy rate, resistivity anomaly zone volume, etc.); it is then connected to three hidden layers, with the number of neurons in each hidden layer set to 64, 128, and 64; the activation function is ReLU (Modified Linear Unit) to accelerate convergence and alleviate gradient vanishing; the output layer contains neurons corresponding to dust concentration and particle size distribution. During training, the Adam optimizer and mean squared error (MSE) loss function can be used, with approximately 1000 iterations. This network architecture is only an exemplary configuration in the experiment. In actual applications, the network depth, number of neurons and hyperparameters can be adjusted according to the data scale and feature dimensions. For example, it can be replaced with a convolutional neural network (CNN) to process time-series image features, or a long short-term memory network (LSTM) can be used to mine time-series dependencies.

[0073] 7. Visualization Module

[0074] The visualization module is connected to the data processing and analysis module to output the correspondence between crack evolution and dust generation characteristics. In this embodiment, the visualization module is a high-resolution display that can simultaneously display the correlation results between crack evolution and dust generation in the form of time history curves, three-dimensional resistivity imaging maps, acoustic emission event cloud maps, and dust concentration curves.

[0075] Example 2

[0076] This embodiment provides an experimental method for dust generation during coal and rock cutting based on real-time crack-dust correlation. The method uses the system and cutting parameters described in Example 1 (cutting speed 3 m / s, traction speed 2 m / min, cutting depth 10 mm). Figure 2 As shown, the specific steps are as follows:

[0077] 1. Simultaneously collect multi-source monitoring data

[0078] During the cutting process of the coal and rock sample (same as in Example 1), all monitoring units were synchronously triggered via FPGA to collect multi-source monitoring data in real time. Specifically, this included:

[0079] (1) Crack evolution data

[0080] The injected current and measured voltage of the electrode array were collected by a resistivity measuring device. The temporal image of the resistivity distribution inside the sample was obtained by resistivity tomography inversion. The initiation location, propagation path and damage accumulation degree of the crack were identified based on the resistivity anomaly zone (the resistivity change threshold was set to 20%). Eight-channel waveform data were recorded by an acoustic emission data acquisition device. The spatial coordinates of the micro-fracture event were calculated by the time difference positioning algorithm. The crack type was identified based on the RA and AF values ​​(RA threshold 5ms / V, AF threshold 100 kHz). A surface speckle image sequence was acquired by a high-speed camera (frame rate 5000 fps). The surface crack propagation information, including crack length and crack propagation velocity, was extracted by the DIC algorithm and compared with the internal crack results obtained by resistivity inversion for verification.

[0081] (2) Dust data

[0082] Dust concentration and particle size distribution data are collected using a dust concentration meter and an online laser particle size analyzer; airflow velocity v(t) is measured in real time using an anemometer. The dust generated during cutting is transported from the cutting area to a dust and anemometer monitoring unit installed at the outlet of the anemometer via an airflow transmission duct.

[0083] (3) Auxiliary data

[0084] A triaxial force sensor installed at the root of the cutting tooth collects force data on the cutting tooth to help determine load changes during crack propagation.

[0085] 2. Time synchronization and spatial matching to establish real-time relationships.

[0086] (1) Time synchronization

[0087] All sensors use the same FPGA hardware trigger signal for synchronous acquisition, ensuring that the time alignment accuracy of multi-source data reaches the microsecond level. However, due to the time delay in the transmission of dust from the cutting area to the dust and wind speed monitoring unit, compensation is required. Based on the real-time airflow velocity v(t) measured by the wind speed sensor and the pipe length L, the time delay is calculated using the formula Δt = L / v(t). In this embodiment, L = 2.0 m, the measured average value of v(t) is 1.5 m / s, and Δt ≈ 1.33 s is calculated. The timestamp of the dust data is shifted forward by Δt to achieve physical time alignment with the crack data. After compensation, the peak dust concentration and the peak volume expansion rate of the resistivity anomaly zone coincide on the time axis.

[0088] (2) Spatial matching

[0089] A unified sample coordinate system (a right-handed rectangular coordinate system with the sample's geometric center as the origin) was established to uniformly transform the spatial coordinates of crack evolution data and the spatial positions of dust sampling points to the same coordinate system. The spatial coordinates of the crack evolution data include: the three-dimensional coordinates of internal cracks obtained through resistivity tomography inversion; the spatial coordinates of micro-fracture events obtained through acoustic emission time-of-flight localization; and the surface crack propagation coordinates obtained through high-speed camera combined with digital image correlation analysis (converted to a three-dimensional coordinate system through calibration). This achieves a spatial correspondence between crack locations and dust sources.

[0090] (3) Acoustic emission-resistivity joint monitoring

[0091] The acoustic emission localization results and resistivity imaging results are fused in a unified spatiotemporal coordinate system to identify the spatial correspondence between areas of dense acoustic emission events and areas of resistivity anomalies. In this embodiment, the areas of dense acoustic emission events and areas of resistivity anomalies highly overlap in space (overlap rate greater than 80%), verifying the accuracy of crack localization.

[0092] 3. Correspondence between the output crack evolution process and dust generation characteristics

[0093] The data processing and analysis module performs the following analysis and outputs the results:

[0094] (1) Identify the crack evolution stages corresponding to dust generation rate and intensity

[0095] Plot the time history curves of volume expansion rate, acoustic emission event rate and dust concentration in the resistivity anomaly region, and identify that the peak dust generation rate occurs at the moment of internal crack penetration.

[0096] (2) Analyze the correspondence between different crack types and dust particle size distribution.

[0097] Based on the acoustic emission identification results, the dust particle size distribution corresponding to the occurrence times of tensile crack and shear crack events was statistically analyzed. In this embodiment, during the stage of large-scale initiation of tensile cracks in the mid-stage of cutting, the proportion of respirable dust smaller than PM10 increased from 15% in the initial stage to 42%; while when shear cracks were dominant, the proportion of large-diameter dust increased.

[0098] (3) Establish a quantitative relationship model

[0099] Using crack propagation velocity, crack length, acoustic emission energy, and resistivity anomaly zone volume as inputs, and dust concentration and particle size distribution as outputs, a deep neural network model is trained to establish a quantitative mapping relationship between crack evolution parameters and dust generation characteristics. The analysis results are presented through a visualization module in the form of curves (crack evolution-dust generation curves), mapping diagrams (crack type-particle size distribution mapping diagrams), and three-dimensional images (three-dimensional dynamic images of crack propagation).

[0100] Example 3: Comparative Experiment with Different Traction Speeds

[0101] This embodiment, based on Example 2, keeps the coal and rock sample mix (same as Example 1) constant, but changes the traction speed of the cutting simulation device to analyze the effect of different traction speeds on the crack-dust correlation. Details are as follows:

[0102] The traction speed was set to 1 m / min, 2 m / min, and 3 m / min respectively, the cutting speed was kept at 3 m / s, and the cutting depth was kept at 10 mm. The experimental steps of Example 2 were repeated.

[0103] Experimental results show that when the traction speed is 1 m / min, crack propagation is relatively slow. Acoustic emission analysis results show that tension cracks account for approximately 68%, and the peak dust concentration is 85 mg / m³. 3 PM10 accounted for approximately 25%; when the traction speed increased to 3 m / min, the proportion of shear cracks rose to 41%, the proportion of tension cracks decreased to 59%, and the peak dust concentration rose to 142 mg / m³. 3 The proportion of PM10 decreased to 18%, while the proportion of large-diameter dust increased significantly.

[0104] The above results indicate that while increasing the traction speed increases the total dust generation, the proportion of respirable dust decreases. This is due to the increased shear damage ratio under high-speed cutting. The method of this invention can intuitively reveal the influence mechanism of different cutting parameters on dust particle size distribution, providing experimental basis for optimizing cutting process parameters.

[0105] Example 4: Comparative Experiment of Different Coal and Rock Samples

[0106] This embodiment uses coal and rock samples with different mechanical properties to analyze the influence of coal type differences on the crack-dust correlation. Details are as follows:

[0107] Three types of conductive smart similar material samples were prepared to simulate soft lignite (cement:gypsum:coal powder = 0.5:1:3, uniaxial compressive strength approximately 8 MPa), medium-hardness bituminous coal (cement:gypsum:coal powder = 1:1:2, uniaxial compressive strength approximately 15 MPa), and hard anthracite (cement:gypsum:coal powder = 2:1:2, uniaxial compressive strength approximately 28 MPa). The conductive filler content was 2% carbon nanotubes in all samples, and the electrode arrangement was the same as in Example 1. The experimental steps of Example 2 were repeated under the same cutting parameters (cutting speed 3 m / s, traction speed 2 m / min, cutting depth 10 mm).

[0108] Experimental results show that shear cracks account for as high as 55% during the cutting process of soft lignite samples, and the median diameter of dust particles D... 50The particle size was approximately 45 μm, with coarse particles mainly detaching. During the cutting process, tensile cracks accounted for 78% of the hard anthracite samples, and PM10 accounted for 48% of the dust, resulting in significantly higher levels of respirable dust compared to other coal types. The results for medium-hardness bituminous coal fell between these two categories. These differences are closely related to the brittleness, hardness, and strength characteristics of different coal types. Hard and brittle coal bodies are more prone to tensile failure, thus generating more fine-grained dust.

[0109] Example 5: Verification of the repeatability and reliability of the experimental system

[0110] To verify the stability and data reliability of this experimental method, the experimental steps of Example 2 were repeated five times under the same cutting parameters (cutting speed 3 m / s, traction speed 2 m / min, cutting depth 10 mm). Before each experiment, a brand new coal and rock sample from the same batch (same as in Example 1) was used. After the system initialization was completed and all sensors were confirmed to be synchronously triggered normally, the cutting simulation device was started, data was collected, and analyzed.

[0111] The results showed that the peak dust concentration in the five repeated experiments was 142.5 mg / m³. 3 138.2 mg / m 3 143.1 mg / m 3 139.8 mg / m 3 141.4 mg / m 3 The average value was 141.0 mg / m³. 3 The relative standard deviation was 1.5%; the proportions of PM10 were 41.8%, 40.5%, 42.5%, 41.2%, and 42.1%, respectively, with an average of 41.6% and a relative standard deviation of 1.9%. The relative standard deviations of the key data were all controlled within a small range (less than 5%), indicating that the experimental method has good repeatability and reliability.

[0112] Example 6: Validation of the Time Delay Compensation Algorithm

[0113] This embodiment is used to verify the effectiveness of the time delay compensation algorithm Δt = L / v(t). Specifically:

[0114] The same experimental conditions as in Example 1 were used, namely, medium-hardness bituminous coal samples, cutting speed of 3 m / s, traction speed of 2 m / min, and cutting depth of 10 mm. In this example, the length of the airflow transmission pipe L = 2.0 m, and the average value of the airflow velocity v(t) measured by the wind speed sensor during the experiment was approximately 1.5 m / s. According to the formula, Δt ≈ 1.33 s was calculated.

[0115] Two sets of data processing methods were compared. The experimental group used the delay compensation algorithm of this invention, calculating the delay time based on the measured airflow velocity v(t) using the formula Δt = L / v(t), and then aligning the timestamp of the dust data with the crack evolution data by shifting the timestamp forward by Δt. The control group did not perform delay compensation and directly aligned the timestamp of the dust data with the timestamp of the crack evolution data.

[0116] The average of the peak time of the volume expansion rate in the resistivity anomaly region and the peak time of the acoustic emission event rate in the crack evolution data is taken as the peak crack time t. crack The peak dust concentration time is taken as the peak dust time t. dust Calculate t after aligning the two sets of data. crack With t dust Time difference Δt peak =|t crack -t dust The smaller the time difference, the higher the time alignment accuracy between the crack event and the dust generation event.

[0117] Experimental results show that, in the control group without delay compensation, t crack It is 15.23 s, t dust The time difference is 16.58s, Δt peak The peak dust concentration lags significantly behind the peak crack concentration by 1.35 s. This lag time is basically consistent with the theoretically calculated Δt≈1.33 s, verifying that the delay originates from the pneumatic transport process of dust in the pipeline. After adopting the delay compensation algorithm, the experimental group achieved a delay of 1.35 s. dust After compensation, the time is 15.24 s, and the time difference Δt peak The time was reduced significantly to 0.01 s, a decrease of over 99%, and the peak dust concentration time basically coincided with the peak crack time.

[0118] The above comparison results show that the Δt=L / v(t) delay compensation algorithm proposed in this invention can effectively eliminate the time lag caused by dust pneumatic conveying, and achieve precise alignment of crack events and dust generation events on the time axis. It is a key technical means to ensure the accurate establishment of real-time correlation between cracks and dust.

[0119] Example 7: Construction, Training, and Validation of Deep Neural Network Models

[0120] This embodiment details the construction, training, and validation process of the deep neural network model in the data processing and analysis module. Specifically:

[0121] In this embodiment, the deep neural network model adopts a fully connected feedforward network architecture. The input layer has 5 nodes, corresponding to crack propagation velocity, crack length, acoustic emission energy rate, resistivity anomaly zone volume, and the resultant force of the three-dimensional forces of the cutting teeth, respectively; there are 3 hidden layers, with 64, 128, and 64 nodes in each layer, respectively; the output layer has 2 nodes, corresponding to total dust concentration and PM10 percentage, respectively. The hidden layer activation function is ReLU, and the output layer activation function is a linear function. The model training uses mean squared error as the loss function, Adam is selected as the optimizer, the learning rate is set to 0.001, the batch size is set to 32, and the training epochs are 1000.

[0122] The model training dataset was derived from all data collected in the experiments with different cutting parameters in Example 3 and the experiments with different coal types in Example 4. During the cutting process, samples were continuously collected in time series for each experiment. Crack evolution characteristic parameters and dust generation characteristic parameters were extracted from each sampling point, resulting in 500 valid samples. These 500 samples were randomly divided into a training set of 400 samples and a test set of 100 samples, with 80% and 20% respectively. The training set was used for model parameter optimization, and the test set was used to evaluate the model's predictive performance.

[0123] During training, the changes in the loss functions of the training and test sets were monitored simultaneously. After 1000 iterations, both the training and test set losses tended to stabilize and converge to a low level, with no significant difference between them, indicating that the model did not exhibit overfitting or underfitting.

[0124] After training, the model's prediction accuracy was quantitatively evaluated using 100 unseen samples in the test set. The root mean square error (RMSE) and coefficient of determination were selected as the evaluation metrics. The results showed that the RMSE for predicting total dust concentration was 4.2 mg / m³. 3 Compared to the peak dust concentration of 142 mg / m³ at a traction speed of 3 m / min in Example 3. 3 The relative error is approximately 3%; the coefficient of determination for predicting the proportion of PM10 is 0.89. These results demonstrate that the deep neural network model constructed in this invention can accurately establish the nonlinear mapping relationship between crack evolution characteristics and dust generation characteristics, with prediction accuracy meeting experimental requirements, providing a reliable technical means for the quantitative analysis of dust generation patterns during cutting.

[0125] It should be noted that the network structure and hyperparameter settings described above are merely exemplary configurations for this embodiment. In practical applications, the number of hidden layers and nodes can be adjusted according to the data scale and feature dimensions, or they can be replaced with other network architectures (such as CNN for processing temporal image features, or LSTM for mining time series dependencies).

[0126] The embodiments described above are only some, not all, of the embodiments of the present invention. The detailed description of the embodiments of the present invention is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments. The scope of protection of the present invention is determined by the scope claimed in the claims. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

Claims

1. An experimental system for dust generation during coal and rock cutting based on real-time correlation between cracks and dust, characterized in that, include: Coal and rock samples are used to simulate the object to be cut. A cutting simulation device is used to cut the coal and rock sample. The crack evolution data acquisition unit is used to acquire crack evolution data in real time, reflecting the crack initiation and propagation process inside coal and rock samples. The dust and wind speed monitoring unit is used to collect real-time data on dust concentration and / or dust particle size distribution and wind speed generated during cutting. A synchronous data acquisition module is connected to the crack evolution data acquisition unit and the dust and wind speed monitoring unit to realize the synchronous triggering and acquisition of all data. The synchronous data acquisition module uses a field-programmable gate array as the main controller to generate a unified sampling clock and trigger signal; The data processing and analysis module is connected to the synchronous data acquisition module and is used to execute experimental method steps; The visualization module, connected to the data processing and analysis module, is used to output the correspondence between the crack evolution process and the dust generation characteristics.

2. The experimental system for dust generation in coal and rock cutting based on real-time correlation of cracks and dust as described in claim 1, characterized in that, The coal and rock sample is a conductive similar material, which is formed by uniformly mixing a coal and rock similar material matrix with a conductive filler; the conductive filler is one or more of carbon nanotubes, carbon fibers, graphene, carbon black, steel slag, or conductive polymer materials, with a dosage range of 0.5 wt% to 10 wt%. The crack evolution data acquisition unit includes: An electrode array arranged in a preset network inside or on the surface of the coal and rock sample, and a resistivity measuring device connected to the electrode array, are used to collect resistivity distribution data in real time to identify internal cracks. An acoustic emission sensor array arranged on the surface of the coal and rock sample, and an acoustic emission data acquisition device connected to the acoustic emission sensor array, are used to acquire micro-fracture event data in real time. A high-speed camera used to acquire image sequences of surface cracks in the coal and rock samples.

3. The experimental system for dust generation in coal and rock cutting based on real-time correlation of cracks and dust as described in claim 1, characterized in that, It also includes a sealed experimental chamber for accommodating the coal and rock sample, the cutting simulation device, the crack evolution data acquisition unit, and the dust and wind speed monitoring unit; the sealed experimental chamber is provided with an air inlet and an air outlet, the air outlet is connected to an airflow transmission pipe, and the dust and wind speed monitoring unit is installed at the air outlet of the airflow transmission pipe. The dust and wind speed monitoring unit includes a dust concentration meter, a dust particle counter, an online laser particle size analyzer, and a wind speed sensor.

4. The experimental system for dust generation in coal and rock cutting based on real-time correlation of cracks and dust as described in claim 1, characterized in that, The data processing and analysis module has a built-in deep neural network model to establish a nonlinear mapping relationship between crack evolution characteristics and dust generation characteristics; the crack evolution characteristics include one or more of crack propagation rate, crack length, acoustic emission energy rate and resistivity anomaly zone volume; the dust generation characteristics include dust concentration and dust particle size distribution.

5. A method for dust generation during coal and rock cutting based on real-time correlation between cracks and dust, characterized in that, The experimental method is based on the experimental system as described in any one of claims 1-4, and includes the following steps: Step 1) During the cutting of the coal and rock sample, multi-source monitoring data is collected synchronously through the synchronous data acquisition module. The multi-source monitoring data includes: crack evolution data reflecting the crack initiation and propagation process inside the coal and rock sample, and dust concentration data and / or dust particle size distribution data reflecting the characteristics of cutting dust generation. Step 2) Synchronize and match the crack evolution data and dust data collected in Step 1) in time and space to establish a real-time correlation between crack evolution and dust generation; Step 3) Based on the real-time correlation established in Step 2), output the correspondence between the crack evolution process and the dust generation characteristics.

6. The experimental method for dust generation in coal and rock cutting based on real-time correlation of cracks and dust, as described in claim 5, is characterized in that... Step 1) The crack evolution data is collected by an array of acoustic emission sensors arranged on the surface of the coal and rock sample. The acoustic emission waveform data is used for real-time localization of micro-fracture events and crack type identification. The crack type is identified by the rise angle RA and average frequency AF of the acoustic emission signal: if AF is lower than a preset threshold and RA is higher than a preset threshold, it is identified as a tension crack. If AF is higher than a preset threshold and RA is lower than a preset threshold, it is identified as a shear crack; The crack evolution data also includes: a sequence of surface crack images of coal and rock samples acquired by a high-speed camera, used for digital image correlation analysis to extract surface crack propagation information, the surface crack propagation information including crack length and crack propagation speed, and compared and verified with the internal crack results obtained by resistivity inversion; The multi-source monitoring data also includes: force data of the cutting tooth collected by a triaxial force sensor installed at the root of the cutting tooth, which is used to help determine the load changes during crack propagation.

7. The experimental method for dust generation in coal and rock cutting based on real-time crack-dust correlation as described in claim 5, characterized in that, Step 2) The specific method for time synchronization is as follows: all sensors use the same hardware trigger signal for synchronous acquisition to ensure that the time alignment accuracy of multi-source data reaches the microsecond level; the dust generated during cutting is transported from the cutting area to the dust and wind speed monitoring unit installed at the air outlet of the airflow transmission pipe through the airflow transmission pipe; the airflow speed is measured in real time by the wind speed sensor set in the airflow transmission pipe; according to the time delay Δt of the dust being transported from the cutting area to the dust and wind speed monitoring unit, the timestamp of the dust data is time-aligned with the crack evolution data, and the time delay Δt is determined by the formula Δt=L / v(t), where L is the length of the airflow transmission pipe and v(t) is the airflow speed measured in real time by the wind speed sensor; The specific method for spatial matching is as follows: a unified sample coordinate system is established, and the spatial coordinates of crack evolution data and the spatial positions of dust sampling points are uniformly transformed to the same coordinate system to achieve spatial correspondence between crack positions and dust sources; the spatial coordinates of crack evolution data include the three-dimensional coordinates of internal cracks obtained by resistivity tomography inversion, the spatial coordinates of micro-fracture events obtained by acoustic emission time difference positioning, and the surface crack propagation coordinates obtained by high-speed camera combined with digital image correlation analysis.

8. The experimental method for dust generation in coal and rock cutting based on real-time correlation of cracks and dust, as described in claim 5, is characterized in that... It also includes a combined acoustic emission-resistivity monitoring step, as detailed below: The acoustic emission localization results and resistivity imaging results are fused in a unified spatiotemporal coordinate system to identify the spatial correspondence between areas of dense acoustic emission events and areas of resistivity anomalies.

9. The experimental method for dust generation in coal and rock cutting based on real-time correlation of cracks and dust, as described in claim 5, is characterized in that... Step 3) specifically includes: Identify the crack evolution stages corresponding to dust generation rate and intensity to determine the main timing of dust generation; The correspondence between different crack types and dust particle size distribution was analyzed, and the crack types included tensile cracks and shear cracks. Establish a quantitative relationship model between crack propagation rate, crack length, crack width and dust generation.

10. The experimental method for dust generation in coal and rock cutting based on real-time correlation of cracks and dust, as described in claim 5, is characterized in that... It also includes a variable impact analysis step, as detailed below: Repeat steps 1) to 3) with different cutting parameters to analyze the changing relationship between crack evolution and dust generation under different cutting parameters; the cutting parameters include one or more of the following: cutting speed, traction speed, cutting tooth angle, and cutting tooth arrangement. Repeat steps 1) to 3) with different coal and rock samples to analyze the changing relationship between crack evolution and dust generation in different coal and rock samples; the differences between the various coal and rock samples lie in one or more of the mechanical properties, pore structure, degree of coalification or rock type; the mechanical properties include one or more of hardness, brittleness and strength.

Images