Comprehensive mining method and system for coal mine shafts and tunnels based on acoustic, optical and electrical multi-source data fusion

Abstract

The present invention belongs to the technical field of coal mine shaft and tunnel mining, and specifically discloses a coal mine shaft and tunnel comprehensive mining method and system based on acoustic, optical and electrical multi-source data fusion, the method includes: judging whether there is a safety risk in the drilling process; obtaining the pick temperature of the cutting head during the coal mining process, and using the coal rock recognition model to obtain the preliminary recognition result of the coal rock; obtaining the reflection spectrum of the coal rock sample during the transportation process, and obtaining the secondary recognition result of the coal rock through machine learning; jointly analyzing the preliminary recognition result and the secondary recognition result of the coal rock to obtain the final coal rock recognition result; performing elemental testing on the coal sample, and using the prediction model of the coal quality index to obtain the quantitative detection results of ash, calorific value, volatile matter and sulfur, and determining the coal type based on the quantitative detection results, thereby realizing the sorting of different types of coal. The present invention can realize the integrated and continuous operation of safe coal mining, efficient identification of coal rock, and intelligent optimization of coal.

Description

Technical Field

[0001] The present invention relates to the technical field of coal mine shaft and tunnel mining, and in particular to a coal mine shaft and tunnel comprehensive mining method and system based on acoustic, optical and electrical multi-source data fusion. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In terms of disaster warning, traditional coal mining methods mainly monitor and warn of disasters based on the experience of construction personnel. There is a lack of comprehensive analysis of multi-source disaster monitoring data during shaft and tunnel excavation, resulting in low warning accuracy and poor targeted implementation of disaster prevention measures in coal mines. In terms of coal and rock identification, due to the complex underground environment, such as the electromagnetic fields generated by high-power equipment, strong vibrations and noise generated by various equipment, etc., the recognition results obtained by coal and rock identification technology based on a single technical principle are not accurate enough to meet the application requirements under actual working conditions. In terms of coal mine optimization, the main method is to transport coal gangue directly from underground to the ground coal preparation plant for centralized sorting. There is no preliminary screening of the characteristics of raw coal types in situ underground, and it is difficult to adopt the best matching sorting method for raw coal with different characteristics in the subsequent centralized sorting, thereby improving the actual coal preparation efficiency.

[0004] In summary, traditional coal mining methods cannot achieve the goal of safely avoiding potential risks and disasters during the mining process, accurately identifying coal rocks, and even performing real-time coal sorting. They cannot guarantee the continuity and effective connection of the entire process of drilling, digging, mining, and sorting in coal mines. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a comprehensive mining method and system for coal mine shafts and tunnels based on the fusion of multi-source data of sound, light and electricity, which can realize the integrated and continuous operation of safe coal mining, efficient identification of coal rocks and intelligent optimization of coal.

[0006] In some embodiments, the following technical solutions are adopted:

[0007] A comprehensive mining method for coal mine shafts and lanes based on acoustic, optical and electrical multi-source data fusion, comprising:

[0008] Real-time acquisition of drilling parameters during the drilling process of the drill pipe, microseismic data of the drill bit drilling feedback, transient electromagnetic data of the drilling receiving coil, and gas concentration monitoring data, and based on the acquired data, determine whether there are safety risks in the drilling process;

[0009] After safely drilling into the set distance, start coal mining, obtain the temperature of the cutting teeth of the cutting head during the coal mining process, and use the coal-rock identification model to obtain the preliminary identification results of the coal and rock;

[0010] Acquire the reflection spectrum of the coal-rock sample during the transportation process, and obtain the secondary recognition result of the coal-rock through machine learning; and judge whether the cutting head is at the coal-rock interface based on the secondary recognition result;

[0011] The preliminary and secondary identification results of coal and rock are jointly analyzed to obtain the final coal and rock identification results;

[0012] Based on the final coal-rock identification results, coal and rock samples are distributed and transmitted, and elemental tests are performed on coal samples. Based on the elemental test results and using the prediction model of coal quality indicators, quantitative test results of ash, calorific value, volatile matter and sulfur are obtained. Based on the quantitative test results, the coal type is determined, thereby achieving the sorting of different types of coal.

[0013] As a further solution, it is determined whether there is a safety risk in the drilling process based on the acquired data, specifically:

[0014] The drilling parameters of the drill pipe during drilling are obtained in real time, and the physical properties of the rock in the formation ahead are obtained using the pre-built drilling parameter-rock physical property inversion model;

[0015] Real-time acquisition of seismic wave signal data generated by the contact surface between the drill bit and the rock and transmitted to the rock fracture structure, extraction of abnormal characteristics of wave velocity, reflection coefficient and waveform characteristics, and analysis to obtain rock structure information of the front stratum;

[0016] A pulse current is applied in the borehole through a drilling transmitting coil to generate a primary magnetic field, and a secondary magnetic field is generated when the primary magnetic field encounters a conductive medium through a receiving coil. The abnormal resistance value data is analyzed to achieve directional detection of water sources in the surrounding rock mass and unknown areas.

[0017] Obtain the concentration of methane and carbon dioxide gases emitted during drilling in real time, and monitor the gas concentration in real time;

[0018] Based on the obtained rock physical properties, rock structure information, water source conditions and gas concentration, determine whether there are safety risks in the drilling process.

[0019] As a further solution, the reflection spectrum of the coal and rock samples during transportation is obtained, and the secondary recognition results of the coal and rock are obtained through machine learning, specifically:

[0020] The characteristic bands with significant differences in the visible-near-infrared reflectance spectrum curves of coal and rocks were selected respectively to establish a standard visible-near-infrared spectrum library for coal and rocks.

[0021] Obtain the visible-near infrared reflectance spectrum of the coal and rock samples during transportation, extract the corresponding characteristic bands, perform similarity analysis on the extracted characteristic bands and the bands in the standard visible-near infrared spectrum library, and identify whether the coal and rock samples are coal or rock.

[0022] As a further solution, judging whether the cutting head is at the coal-rock interface based on the secondary recognition result is as follows:

[0023] Assume that when the secondary identification result is coal, it is recorded as M, and when the secondary identification result is rock, it is recorded as R;

[0024] Each batch of coal and rock samples transported on the conveyor belt is tested separately, the test results are recorded, and the proportion of M and R results in the test results is analyzed. When the proportions of the two are equal, it is judged that the cutting head is at the coal-rock interface.

[0025] As a further solution, the preliminary identification results and the secondary identification results of coal and rock are jointly analyzed, specifically:

[0026] The F1 score was selected as the performance evaluation index to evaluate the coal-rock identification model used for preliminary identification and the machine learning model used for secondary identification.

[0027] Based on the F1 score evaluation values ​​of the two models, the weights of the initial recognition results and the secondary recognition results are determined;

[0028] Calculate the score for each recognition category based on the weights:

[0029] ;

[0030] Among them, k is the identification category, which is coal or rock; m is the number of models; x is the sample; is the prediction result of the i-th model for sample x; is an indicator function, which means that if If it is identified as category k, the function returns a value of 1, otherwise it returns a value of 0; is the weight of the i-th model;

[0031] The category with the highest score is selected as the final coal and rock identification result.

[0032] As a further solution, the coal and rock samples are distributed and transmitted based on the final coal-rock identification results, specifically:

[0033] Perform image recognition on the coal and rock samples on the conveyor belt and match the coal or rock category labels identified for the coal and rock sample images;

[0034] The specific location of the coal and rock samples is determined by sensors, and the corresponding coal and rock samples are transported to the corresponding transmission channels according to the category labels that match the coal and rock samples at each location.

[0035] As a further solution, the prediction model of coal quality indicators establishes the regression relationship between the main component information of the spectral lines of the elements and the coal quality indicators of the standard coal samples through a multivariate linear regression algorithm; the prediction model of coal quality indicators specifically includes an ash prediction model, a calorific value prediction model, a volatile matter prediction model and a sulfur content prediction model, which respectively realize the quantitative prediction of ash content, calorific value, volatile matter and sulfur content.

[0036] As a further solution, the coal type is determined based on the quantitative detection result, and then the different types of coal are sorted, specifically:

[0037] The quantitative test results of ash, calorific value, volatile matter and sulfur content are input into the trained coal type recognition model to identify the corresponding coal type;

[0038] Based on the coal type identification results of all coal samples within a set time interval, determine the proportion of different types of coal;

[0039] Set the speed of the sorter to screen the type of coal with the highest proportion; after the screening is completed, re-determine the quantitative test results of ash, calorific value, volatile matter and sulfur content for the remaining samples, and re-determine the proportion of different types of coal; re-screen the type of coal with the highest proportion; repeat this process until all types of coal samples have been screened.

[0040] In other embodiments, the following technical solutions are adopted:

[0041] A coal mine shaft and tunnel comprehensive mining system based on acoustic, optical and electrical multi-source data fusion, comprising:

[0042] The risk judgment module is used to obtain the drilling parameters of the drill rod during drilling, the microseismic data of the drill bit during drilling feedback, the transient electromagnetic data of the drilling receiving coil, and the gas concentration monitoring data in real time, and judge whether there is a safety risk in the drilling process based on the acquired data;

[0043] The coal-rock preliminary identification module is used to start coal mining after safely drilling a set distance, obtain the temperature of the cutting teeth of the cutting head during the coal mining process, and obtain the preliminary identification results of the coal and rock using the coal-rock identification model;

[0044] The coal-rock secondary identification module is used to obtain the reflection spectrum of the coal-rock sample during the transportation process, obtain the secondary identification result of the coal-rock through machine learning; and judge whether the cutting head is at the coal-rock interface based on the secondary identification result;

[0045] The coal-rock final identification module is used to jointly analyze the preliminary and secondary identification results of coal and rock to obtain the final coal-rock identification results;

[0046] The coal sorting module is used to distribute and transmit coal and rock samples based on the final coal-rock identification results, perform elemental testing on coal samples, and based on the elemental test results and using the prediction model of coal quality indicators, obtain quantitative test results of ash, calorific value, volatile matter and sulfur content, determine the coal type based on the quantitative test results, and then realize the sorting of different types of coal.

[0047] In other embodiments, the following technical solutions are adopted:

[0048] A terminal device includes a processor and a memory, wherein the processor is used to implement instructions; the memory is used to store multiple instructions, and the instructions are suitable for being loaded by the processor and executing the above-mentioned coal mine shaft and lane comprehensive mining method based on acoustic, optical and electrical multi-source data fusion.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] (1) In the drilling and excavation stage before the start of coal mining, the present invention obtains the drilling parameters of the drill rod during drilling, the microseismic data of the drill bit during drilling feedback, the transient electromagnetic data of the drilling receiving coil, and the gas concentration monitoring data in real time. Based on the acquired acoustic, optical and electrical ternary data, it is judged whether there is a safety risk in the drilling process, and comprehensive identification of disasters such as water damage, gas, and landslides is achieved, providing a new technical idea for the safety protection of the entire process of traditional coal mine tunneling.

[0051] (2) During the coal mining process, the present invention obtains the preliminary identification result of coal and rock through the cutting tooth temperature of the cutting head during the coal mining process, and obtains the secondary identification result of coal and rock through the reflection spectrum of the coal and rock sample; the preliminary identification result and the secondary identification result of coal and rock are integrated to distinguish the types of coal and rock; the coal and rock mining and identification are carried out simultaneously, and the accuracy of the coal and rock identification results is ensured at the same time.

[0052] (3) Based on the results of coal rock identification, the present invention locates and sorts the mined coal rock samples in real time, and obtains quantitative test results of ash content, calorific value, volatile matter and sulfur content by performing elemental testing on the coal samples, thereby determining the proportion of different types of coal, and then sorting different types of coal in situ; thus realizing the integrated and continuous process of safe coal mining, efficient coal rock identification and intelligent coal optimization.

[0053] Other features and advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flow chart of a comprehensive mining method for coal mine shafts and lanes based on the fusion of acoustic, optical and electrical multi-source data in an embodiment of the present invention;

[0055] Figure 2 Schematic diagram of characteristic bands with significant differences in the visible-near infrared reflectance spectrum curve of coal rock in an embodiment of the present invention;

[0056] Figure 3 The non-metal rod structure and working diagram of the micro electromagnetic transmission coil module in the embodiment of the present invention is shown. DETAILED DESCRIPTION

[0057] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0058] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0059] Embodiment 1

[0060] In one or more embodiments, a comprehensive mining method for coal mine shafts and lanes based on acoustic, optical and electrical multi-source data fusion is disclosed, comprising:

[0061] S101: Real-time acquisition of drilling parameters of the drill rod during drilling, microseismic data of the drill bit during drilling feedback, transient electromagnetic data of the drilling receiving coil, and gas concentration monitoring data, and based on the acquired data, determine whether there is a safety risk in the drilling process.

[0062] In this embodiment, the integrated drilling and excavation roadheader is equipped with two working devices, namely a drill rod and a cutting head. During the operation of the drill rod device, four data acquisition modules carried on the roadheader are used to obtain detailed technical parameter data of the drill rod drilling, microseismic data of the drill bit drilling feedback, transient electromagnetic data of the drilling receiving coil and real-time data of the while-drilling gas concentration monitoring device. By analyzing the above data one by one, a real-time judgment is made on the safety of coal mine tunneling. If there is a safety risk warning, the excavation is stopped; if there is no safety risk, the excavation is continued.

[0063] Specifically, during the construction and production of coal mines, shaft and tunnel construction goes through two main steps: drilling detection and tunneling excavation. In order to minimize the impact of drilling detection on construction, the present invention realizes comprehensive detection of hidden disasters ahead based on the comprehensive utilization of data such as sound, light and electricity.

[0064] In this embodiment, when the coal mine shaft and tunnel integrated drilling and tunneling machine performs geological directional drilling, various sensors carried on the integrated drilling and tunneling machine are used to obtain real-time drilling technical parameters such as torque, rotation speed, friction force, etc. during the drilling process of each equally spaced drill rod.

[0065] The pre-constructed coal mine shaft rock drilling parameter database is used to train the drilling parameter-rock physical property inversion model. The coal mine shaft rock drilling parameter database stores different drilling parameter combinations (combinations of torque, speed, friction and other data) and their corresponding rock physical property data. In the actual detection process, based on the acquired drilling technical parameter data, the inversion model is used to preliminarily infer the rock physical properties of the formation ahead, so as to achieve the effect of sensing the rock physical properties while drilling.

[0066] The compressive and shear strength of the rock can be determined by the physical properties of the formation rock. If the compressive and shear strength are less than a certain set threshold, it means that there is a safety risk in the drilling process.

[0067] In this embodiment, through the drill bit microseismic data receiving and analyzing module carried by the integrated drilling and tunneling machine, during the drilling process of the drill rod, the seismic wave signal data generated by the contact surface between the drill bit and the rock and transmitted to the rock fracture structure can be obtained in real time through the high-frequency seismic detector. Based on the time series signal processing method, the data is denoised. Since the wave velocity, reflection coefficient, and waveform characteristics in the seismic wave signal are related to the integrity and uniformity of the rock structure, generally speaking, the more complete and uniform the rock structure is, the greater the wave velocity, the higher the reflection coefficient, and the more complex the waveform characteristics are. By analyzing the abnormal manifestations of the wave velocity, reflection coefficient, and waveform characteristics, that is, the sudden change of the wave velocity, the abnormal distribution of the reflection coefficient, and the distorted scattering of the waveform, the rock structure information of the front stratum can be further analyzed.

[0068] Rock structure information can be used to determine whether the rock structure at different locations is intact and whether there are cracks, so that the rock crack situation can be known while drilling. If the seismic wave signal data fed back by the drill bit shows that the rock ahead is relatively broken (the specific degree of breakage can be set according to actual needs), it means that there is a safety risk during the drilling process.

[0069] In this embodiment, combined with Figure 3When the drilling is completed, the corresponding drill rod is withdrawn, and the non-metallic rod equipped with a micro electromagnetic transmitting and receiving coil module, a transmitting circuit, a receiving circuit and a master control circuit is inserted into the borehole. When it is inserted at a fixed rate, the annular component equipped with a micro transmitting coil rotates synchronously, and the receiving coil is fixed on the non-metallic rod; wherein the transmitting circuit includes an oscillator and a power amplifier that generate alternating current, and is used to provide a current of a specific frequency and power to the transmitting coil; the receiving circuit includes an amplifier, a filter, a detector and other components, and is used to process and extract the weak signal sensed by the receiving coil; the master control circuit is used to control the power response and data transmission of the entire device. The specific circuits are all existing structures and will not be described in detail.

[0070] When the device is working, the transmitting coil applies a pulse current to send a primary magnetic field to the surrounding rock mass and unknown areas. The receiving coil receives the secondary magnetic field generated by the primary magnetic field encountering the conductive medium. Due to the difference in resistivity of different media, when encountering a water-bearing area, the measured secondary magnetic field resistance value is lower. By analyzing the spatial position of these resistance anomalies and performing three-dimensional imaging, the specific location and scope of the water-bearing area can be determined, thereby realizing water source detection and achieving the effect of detecting hidden water hazards while drilling.

[0071] Since the tunnel surrounding rock mass is a high-resistance body and the water-containing area is a low-resistance body, a low resistance value appearing in a relatively high resistance value is an abnormal resistance value, indicating the existence of a water-containing area; if there is a hidden water source within the set range ahead (such as a 10-20m three-dimensional space range), it means that there is a safety risk during the drilling process.

[0072] In this embodiment, a small optical methane detection module is installed on the end of each drill rod of the integrated drilling and tunneling machine. The principle of optical interference is used to measure the concentration of gases such as methane and carbon dioxide that escape during the drilling process, and the gas concentration value is monitored in real time. When the gas concentration reaches the safety threshold, an alarm is issued in time, thereby achieving the effect of avoiding gas disasters while drilling.

[0073] If the gas concentration monitoring data shows that the gas concentration leaking out while drilling reaches the safety threshold, it means that there is a safety risk during the drilling process.

[0074] This embodiment utilizes the drilling parameters during the drilling process, the microseismic data fed back by the drill bit during drilling, the transient electromagnetic data of the drill receiving coil, and the gas concentration monitoring data to obtain the rock physical properties, rock structure information, water source conditions, and gas concentration conditions, respectively. If any of the above four data is abnormal, it is determined whether there is a safety risk in the drilling process.

[0075] This embodiment uses a combination of acoustic, optical and other data to detect risks ahead. It inverts the rock compressive and shear strength by acquiring the drilling parameters of the drill rod in real time during the drilling process, infers the structural integrity of the rock mass through the seismic wave signal data fed back by the drill bit, delineates the area of ​​low-resistance bodies such as water sources through the transient electromagnetic data of the drill receiving coil, and determines the risk of gas explosion through the gas concentration monitoring data. Based on the acquired acoustic, optical and other multi-dimensional data, it realizes the comprehensive identification of disasters such as rock collapse, hidden water hazards, and gas explosions, thereby overcoming the shortcoming of single data in identifying fewer disaster types and ensuring the safety and reliability of the entire process of coal mine tunneling.

[0076] S102: After safely drilling a set distance, coal mining begins, the temperature of the cutting teeth of the cutting head during the coal mining process is obtained, and a preliminary coal-rock identification result is obtained using a coal-rock identification model.

[0077] In this embodiment, step S101 is used to ensure the safety of the excavation stage in coal mine shaft and tunnel mining. After the coal mine shaft and tunnel are excavated for a certain distance, the entire mining workflow enters the coal mining stage. When the comprehensive coal mining machine is placed on the shaft and tunnel working face for excavation operations, in order to maximize the coal output rate and reduce equipment wear, infrared thermal imaging is used to identify coal rocks in real time.

[0078] Specifically, after a certain distance of safe mining in the coal mine, the coal mining machine at the fully mechanized mining face starts to work, and the cutting head equipped with an infrared thermal imaging recognition module performs cutting operations. The infrared thermal imaging recognition module is an existing technology that uses highly sensitive, passive infrared to measure the temperature changes when the coal mining machine contacts different geological materials such as coal seams, sandstone, shale, etc. At the same time, the temperature of the surface of different geological layers near the cutting teeth can measure a temperature difference as small as 0.1°C.

[0079] Based on the principle of infrared radiation, the infrared thermal imaging recognition module monitors in real time the temperature changes of the cutting teeth when the cutting head of the coal mining machine cuts the coal-rock mixed layer in the shaft and tunnel. In the initial stage of the cutting process, since the cutting head is not in complete contact with the coal-rock mixed layer, unstable jumping of the cutting teeth will occur, so the temperature rise of the cutting teeth is not obvious. At this time, it is necessary to conduct collaborative analysis with the visible-near infrared reflectance spectrum detection module of the rear section of the conveyor belt to determine whether the cutting head is currently at the coal-rock interface; the purpose of identifying whether it is at the coal-rock interface is to move the cutting head toward the coal layer so that it is in the coal layer as much as possible, thereby improving the coal mining efficiency and reducing the probability of the appearance of coal-rock mixed samples.

[0080] When the cutting process enters the stable stage, and the cutting teeth of the coal mining machine cutter head enter the stage of stable temperature change, according to the empirical rule that the higher the rock content in the coal rock, the higher the average temperature of the cutting teeth of the cutting head, and based on a large number of cutting tooth temperature data samples measured in advance in the laboratory at different coal-rock mixing ratios, a coal-rock recognition model is trained. The model can determine the coal-rock ratio contained in the sample according to the input cutting tooth temperature data, and thus identify the coal and rock type; therefore, in actual recognition, the infrared thermal imaging recognition module is used in real time to record the cutting tooth temperature data in the cutting operation, and it is input into the recognition model as input data to obtain preliminary coal-rock recognition results.

[0081] Specifically, the data set of the coal-rock identification model is the average value of a large number of pick temperature data measured at different coal-rock mixing ratios in the laboratory. The coal-rock samples are mixed in a changing pattern of 100:0, 99:1, 98:2, ..., 2:98, 1:99, and 0:100. The pick temperature is measured 1000 times at each mixing ratio to calculate the temperature average value at the ratio, and the coal-rock samples at different mixing ratios and the corresponding temperature average values ​​form a model data set; at the same time, the data set data is divided into a training set and a validation set at a ratio of 8:2, and a coal-rock identification model is established using common neural network regression and classification algorithms. The regression algorithm trains the model based on the established training set, and continuously optimizes the model through the validation set, so that the pick temperature data is input into the model to predict the coal-rock ratio in the sample. At the same time, the classification algorithm classifies the samples with a coal proportion of more than 90% as coal, and the samples with a rock proportion of more than 90% as rock in the prediction results, thereby realizing the identification of coal and rock types.

[0082] S103: Obtaining the reflection spectrum of the coal rock sample during the transportation process, and obtaining the secondary recognition result of the coal rock through machine learning;

[0083] In this embodiment, a visible-near-infrared reflection spectrum detection module is installed at the near end of the conveyor belt connected to the coal mining machine, which can perform spectral analysis on the coal and rock samples on the conveyor belt; the visible-near-infrared reflection spectrum detection module is implemented using existing technology; through the visible-near-infrared reflection spectrum module, the coal and rock samples on the conveyor belt that have been preliminarily identified by infrared thermal imaging can be identified by reflection spectrum; combined with Figure 2In certain bands, such as around 1400nm, 1900nm, and 2200nm, the rock spectrum shows obvious absorption valleys, while the coal spectrum curve is always flat. The two have significant differences in these bands, so these characteristic bands with significant differences are selected in advance to establish a standard visible-near infrared spectrum library for coal and rock; the data format stored in the standard visible-near infrared spectrum library for coal and rock is a value pair with wavelength-reflectivity as horizontal and vertical coordinates, and the horizontal axis range of the value pair is the wavelength range of the characteristic bands with significant differences in the visible-near infrared reflectance spectrum curves of coal and rock; plotting wavelength-reflectivity as a line graph is a common visible-near infrared reflectance spectrum curve form.

[0084] At the same time, the reflectance spectrum module performs data preprocessing on the measured reflectance spectrum of the coal and rock samples on the conveyor belt, extracts the characteristic bands therein, and uses the built-in machine learning algorithm in the module to perform similarity comparison and analysis on the extracted characteristic bands of the reflectance spectrum of the coal and rock samples with the bands in the standard spectrum library, thereby obtaining the secondary identification results of the coal and rock; among them, the machine learning algorithm uses existing known methods.

[0085] In this embodiment, in the initial stage of the cutting process, based on the secondary recognition result, it can be determined whether the cutting head is at the coal-rock interface. The specific process is as follows:

[0086] When the coal and rock samples cut in the initial stage of cutting are just transported to the visible-near infrared reflectance spectroscopy module for testing, the secondary recognition results of the batch of coal and rock are obtained. It is assumed that the secondary recognition result is coal and is recorded as M, and the recognition result is rock and is recorded as R. The recognition result of this batch of coal and rock is coal and is recorded as M1. At this time, the result is stored; as the conveyor belt rotates, the secondary recognition results of subsequent batches of coal and rock are obtained. It is assumed that the recognition result is rock and is recorded as R1; when the number of test batches gradually increases, the secondary recognition results of coal and rock form a data set, such as: C={M1, R 1, M2, M3, R2, … ,M p , R q}, by analyzing the proportion of M and R data in the set, when the proportions of the two are equal, or both change slightly within a range of about 50%, it is judged that the cutting head is at the coal-rock interface; the specific small change range can be set according to actual needs, for example, the change range does not exceed 5%. If it is at the coal-rock interface, the cutting head is moved to the coal layer to improve mining efficiency and coal quality.

[0087] It should be noted that a batch of coal rock refers to the coal samples on the conveyor belt within a set time interval.

[0088] S104: jointly analyzing the preliminary identification result and the secondary identification result of the coal and rock to obtain the final coal and rock identification result.

[0089] In this embodiment, the preliminary identification results and the secondary identification results of coal and rock are jointly analyzed, specifically:

[0090] S1041: Selecting the F1 score as a performance evaluation index, and performing F1 score index evaluation on the coal-rock recognition model used for preliminary recognition and the machine learning model used for secondary recognition;

[0091] S1042: Determine the weights of the preliminary recognition result and the secondary recognition result based on the F1 score evaluation value E1 of the coal-rock recognition model and the F1 score evaluation value E2 of the machine learning model;

[0092] In this embodiment, the performance of the two models is evaluated according to the index, and a large weight is given to a large performance index and a small weight is given to a small index. Specifically, the relative weights R1 and R2 of the two models on the performance index can be calculated. , , using R1 and R2 as weights and Thus, weights are given to the initial recognition results and the secondary recognition results.

[0093] S1043: Calculate the score of each recognition category based on the weight:

[0094] ;

[0095] Among them, k is the identification category, which is coal or rock; m is the number of models; x is the sample; is the prediction result of the i-th model for sample x; is an indicator function, which means that if If it is identified as category k, the function returns a value of 1, otherwise it returns a value of 0; is the weight of the i-th model;

[0096] S1044: Select the category with the highest score as the final coal-rock identification result; for example:

[0097] The recognition result of the coal-rock recognition model is coal (M), with a weight of 0.4, and the recognition result of the machine learning model is rock (R), with a weight of 0.6. The score Y value when k=M is ​​calculated as 0.4*1+0.6*0=0.4, and the score Y value when K=R is calculated as 0.4*0+0.6*1=0.6, so the final recognition result is rock (R).

[0098] This embodiment combines the advantages of fast, efficient and accurate analysis of infrared thermal imaging and visible-near infrared reflectance spectroscopy by jointly analyzing the preliminary recognition results and secondary recognition results of coal and rock, thereby realizing effective recognition of coal and rock types.

[0099] S105: Based on the final coal-rock identification results, the coal and rock samples are distributed and transmitted, and the coal samples are subjected to elemental testing; based on the elemental testing results, the prediction model of coal quality indicators is used to obtain quantitative test results of ash, calorific value, volatile matter and sulfur content, and the coal type is determined based on the quantitative test results, thereby realizing the sorting of different types of coal.

[0100] In this embodiment, according to the recognition results of coal and rock, the coal and rock are automatically distributed to two different conveyor belts through the automatic distribution device of coal and rock, and marked as coal conveyor belt and rock conveyor belt for subsequent coal sorting; the automatic distribution device is equipped with an image recognition module, a coal positioning sensor, a control module and a sorting module; specifically, the image of the coal and rock sample and the corresponding recognition result are transmitted to the image recognition module, the image recognition is performed on the sample entering the detection, the corresponding category labels "coal" and "rock" are matched, and the matching results are input into the control module; the conveyor belt samples are transported to the sorting module, at which time the control module determines the precise position of the sample through the sensor, and according to the matching result, the control module triggers the mechanical pusher in the sorting module to move the samples with the "coal" label and the "rock" label to the coal conveyor and rock conveyor belt tracks respectively.

[0101] Among them, the image recognition, position recognition and the sorting process of the sorting module according to the control instructions can all be achieved through existing technologies, and the specific process will not be described in detail.

[0102] In this embodiment, the LIBS&XRF combined testing module installed at the far end of the conveyor belt connected to the coal mining machine performs real-time analysis and testing of elements and other components on the identified coal samples, and realizes in-situ "α"-shaped circular sorting of the coal based on the spectral analysis results.

[0103] Specifically, a LIBS & XRF joint test module is mounted on the marked coal conveyor belt. The distributed coal passes through the conveyor belt and enters the joint test module, where LIBS and XRF are used to perform synchronous rapid tests on all elements of the coal and ash-forming elements.

[0104] Among them, the ash-forming elements of coal refer to a class of elements that do not volatilize during the combustion or gasification of coal and are eventually converted into ash (solid residue). These elements exist in coal in the form of minerals or organic matter, such as silicon (Si), aluminum (Al), iron (Fe), calcium (Ca), magnesium (Mg), potassium (K) and sodium (Na).

[0105] Synchronous rapid testing refers to the extraction of the main components of organic and inorganic element spectral lines related to ash, calorific value, volatile matter and sulfur content based on the standard LIBS and XRF data of coal samples through the principal component analysis method, which are then used as input variables after dimensionality reduction, and a prediction model for coal quality indicators is established using a regression algorithm; the measured spectral data is then preprocessed to extract the key spectral line components, which are input into the prediction model for coal quality indicators to obtain quantitative results of ash, calorific value, volatile matter and sulfur.

[0106] The prediction model of the coal quality index in this embodiment is to establish the regression relationship between the main component information of the spectral line of the element and the coal quality index of the standard coal sample through the multivariate linear regression algorithm, and establish the prediction model of each index respectively; existing studies have shown that the ash content in the index is related to the inorganic elements in the coal. In this embodiment, the spectral lines of elements such as Si, Ca, Na, Mg, Al, and Fe are selected as input variables of the ash content prediction model to obtain the quantitative results of the ash content. The calorific value in the index is related to both organic and inorganic elements in the coal. In this embodiment, the spectral lines of elements such as C, H, S, Si, Ca, Na, Mg, Al, and Fe are selected as input variables of the calorific value prediction model to obtain the quantitative results of the calorific value. The volatile matter in the index selects the spectral lines of elements such as C, H, and S as the input variables of the volatile matter prediction model to obtain the quantitative results of the volatile matter. For the sulfur content in the index, the spectral line of the S element and the spectral lines of elements overlapping with the spectral line of the S element are selected as the input variables of the sulfur content prediction model to obtain the quantitative results of the sulfur content.

[0107] The laboratory test values ​​of four types of coal quality indicators of standard coal samples and the corresponding key spectral line data are used as training and verification data to build a prediction model for coal quality indicators, and the model is optimized through the test set. Finally, the spectral line component data of the coal samples to be predicted are input into the four prediction models respectively, so that quantitative detection of ash content, calorific value, volatile matter and sulfur content can be achieved.

[0108] Based on the quantitative test results, the coal type is determined, and then the different types of coal are sorted. The specific process is as follows:

[0109] S1051: Input the quantitative test results of ash content, calorific value, volatile matter and sulfur content into the trained coal type recognition model to identify the corresponding coal type;

[0110] Specifically, according to the national standard for coal classification, the specific quantitative results and corresponding category labels of ash, calorific value, volatile matter and sulfur content of standard coal samples are sorted out to form a specimen sample data set. At the same time, the quantitative results of the four types of indicators of the measured coal samples and the sample numbers are formed into a measured sample data set. The data set is preprocessed and cleaned, and common classification algorithms such as support vector machines (SVM) are selected. The SVM model is trained based on the data of the standard sample data set, and the data of the measured sample data set are input into the model to predict the coal category.

[0111] S1052: Determine the proportion of different types of coal based on the coal type identification results of all coal samples within a set time interval;

[0112] A batch contains multiple coal samples. The proportion of each type of coal in the data set is predicted based on the coal samples in the set batch, that is, the proportion of different types of coal is determined.

[0113] S1053: Set the rotation speed of the sorter to first screen the coal type with the highest proportion; after removing the coal type with the highest proportion, use the same method to screen the coal type with the highest proportion in the remaining coal samples; repeat the above process until all types of coal samples are screened.

[0114] In this embodiment, different types of coal are sorted in-situ in a circular manner, and the specific process is as follows:

[0115] Coal samples are laid out in a single layer on the conveyor belt. When a batch of samples Sample1 is transported to the sorting detector, it has a built-in prediction model for coal quality indicators and coal categories. Based on the quantitative results of four types of coal quality indicators, the coal category is predicted, and the proportion of different types of coal in the batch of samples is determined.

[0116] The sample is then input into a small diameter heavy medium cyclone for sorting. At this time, the sorting controller optimizes the sorting by adjusting the speed of the heavy medium cyclone. The controller has a set of comparison relationships between coal types and heavy medium cyclone speeds obtained through a large number of tests. For example, coal type A at speed V a ±x% has the maximum sorting accuracy. When the detector calculates that the proportion of coal type A in the batch of samples is high, the controller adjusts the speed of the heavy medium cyclone to V a Within the range of ±x%, a sorting result with a higher sorting accuracy of coal type A is obtained; when the sorting of Sample 1 is completed, coal type A is retained, and the coal other than type A is added to the next batch of samples, Sample 2;

[0117] The next batch of samples, Sample 2, are transported back to the LIBS&XRF joint test module for quantitative detection of ash, calorific value, volatile matter and sulfur content. The coal is predicted and the proportion of different coal types is determined through the sorting detector. The coal is then transported to the sorter to sort the coal type with the largest proportion. After sorting, the remaining coal is added back to the next batch of samples, Sample 3.

[0118] And so on, until all types have been sorted, thus achieving cyclic intelligent optimization of coal and improving screening accuracy.

[0119] In this embodiment, a batch of samples is actually the coal samples on the conveyor belt within a certain time interval. For example, the coal samples on the conveyor belt from time A to time B are batch Sample 1, and the coal samples on the conveyor belt from time B to time C are batch Sample 2.

[0120] Based on the element test results, this embodiment uses the prediction model of coal quality indicators to obtain quantitative detection results of ash, calorific value, volatile matter and sulfur content, thereby determining the proportion of different types of coal, realizing rapid quantitative detection of in-situ coal quality indicators, and providing a sorting basis for subsequent cyclic sorting of coal.

[0121] This embodiment realizes the sorting of different types of coal through the method of cyclic intelligent optimization. Through the method of multiple batches of cyclic input, different types of coal can be identified relatively quickly and accurately, reducing the misjudgment rate, improving the sorting efficiency and quality, and at the same time realizing in-situ sorting underground, providing an optimization basis for further selection in subsequent coal preparation plants.

[0122] Embodiment 2

[0123] In one or more embodiments, a coal mine shaft and lane comprehensive mining system based on acoustic, optical and electrical multi-source data fusion is disclosed, comprising:

[0124] The risk judgment module is used to obtain the drilling parameters of the drill rod during drilling, the microseismic data of the drill bit during drilling feedback, the transient electromagnetic data of the drilling receiving coil, and the gas concentration monitoring data in real time, and judge whether there is a safety risk in the drilling process based on the acquired data;

[0125] The coal-rock preliminary identification module is used to start coal mining after safely drilling a set distance, obtain the temperature of the cutting teeth of the cutting head during the coal mining process, and obtain the preliminary identification results of the coal and rock using the coal-rock identification model;

[0126] The coal-rock secondary identification module is used to obtain the reflection spectrum of the coal-rock sample during the transportation process, obtain the secondary identification result of the coal-rock through machine learning; and judge whether the cutting head is at the coal-rock interface based on the secondary identification result;

[0127] The coal-rock final identification module is used to jointly analyze the preliminary and secondary identification results of coal and rock to obtain the final coal-rock identification results;

[0128] The coal sorting module is used to distribute and transmit coal and rock samples based on the final coal-rock identification results, perform elemental testing on coal samples, and based on the elemental test results and using the prediction model of coal quality indicators, obtain quantitative test results of ash, calorific value, volatile matter and sulfur content, determine the coal type based on the quantitative test results, and then realize the sorting of different types of coal.

[0129] The specific implementation of the above modules is the same as that in Example 1 and will not be described in detail.

[0130] Embodiment 3

[0131] In one or more embodiments, a terminal device is disclosed, which includes a processor and a memory, the processor is used to implement instructions; the memory is used to store multiple instructions, and the instructions are suitable for being loaded by the processor and executed by the comprehensive mining method of coal mine shafts and tunnels based on multi-source data fusion of sound, light and electricity described in Example 1.

[0132] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0133] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0134] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software.

[0135] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A comprehensive mining method for coal mine shafts and lanes based on acoustic, optical and electrical multi-source data fusion, characterized in that: include: Real-time acquisition of drilling parameters during the drilling process of the drill pipe, microseismic data of the drill bit drilling feedback, transient electromagnetic data of the drilling receiving coil, and gas concentration monitoring data, and based on the acquired data, determine whether there are safety risks in the drilling process; After safely drilling into the set distance, start coal mining, obtain the temperature of the cutting teeth of the cutting head during the coal mining process, and use the coal-rock identification model to obtain the preliminary identification results of the coal and rock; Acquire the reflection spectrum of the coal-rock sample during the transportation process, and obtain the secondary recognition result of the coal-rock through machine learning; and judge whether the cutting head is at the coal-rock interface based on the secondary recognition result; The preliminary and secondary identification results of coal and rock are jointly analyzed to obtain the final coal and rock identification results; Based on the final coal-rock identification results, the coal and rock samples are distributed and transmitted, and the coal samples are subjected to elemental testing. Based on the elemental testing results, the prediction model of coal quality indicators is used to obtain quantitative test results of ash, calorific value, volatile matter and sulfur content. The coal type is determined based on the quantitative test results, thereby realizing the sorting of different types of coal; Among them, judging whether there is a safety risk in the drilling process based on the acquired data is as follows: The drilling parameters of the drill pipe during drilling are obtained in real time, and the physical properties of the rock in the formation ahead are obtained using the pre-built drilling parameter-rock physical property inversion model; Real-time acquisition of seismic wave signal data generated by the contact surface between the drill bit and the rock and transmitted to the rock fracture structure, extraction of abnormal characteristics of wave velocity, reflection coefficient and waveform characteristics, and analysis to obtain rock structure information of the front stratum; A pulse current is applied in the borehole through a drilling transmitting coil to generate a primary magnetic field, and a secondary magnetic field is generated when the primary magnetic field encounters a conductive medium through a receiving coil. The abnormal resistance value data is analyzed to achieve directional detection of water sources in the surrounding rock mass and unknown areas. Obtain the concentration of methane and carbon dioxide gases emitted during drilling in real time, and monitor the gas concentration in real time; Based on the obtained rock physical properties, rock structure information, water source conditions and gas concentration, determine whether there are safety risks in the drilling process; The preliminary and secondary identification results of coal and rock are jointly analyzed, specifically: The F1 score was selected as the performance evaluation index to evaluate the coal-rock identification model used for preliminary identification and the machine learning model used for secondary identification. Based on the F1 score evaluation values ​​of the two models, the weights of the initial recognition results and the secondary recognition results are determined; Calculate the score for each recognition category based on the weights: ； Among them, k is the identification category, which is coal or rock; m is the number of models; x is the sample; is the prediction result of the i-th model for sample x; is an indicator function, which means that if If it is identified as category k, the function returns a value of 1, otherwise it returns a value of 0; is the weight of the i-th model; The category with the highest score is selected as the final coal and rock identification result.

2. A coal mine shaft and lane comprehensive mining method based on acoustic, optical and electrical multi-source data fusion as claimed in claim 1, characterized in that: Obtain the reflection spectrum of coal and rock samples during transportation, and obtain the secondary recognition results of coal and rock through machine learning, specifically: The characteristic bands with significant differences in the visible-near-infrared reflectance spectrum curves of coal and rocks were selected respectively to establish a standard visible-near-infrared spectrum library for coal and rocks. Obtain the visible-near infrared reflectance spectrum of the coal and rock samples during transportation, extract the corresponding characteristic bands, perform similarity analysis on the extracted characteristic bands and the bands in the standard visible-near infrared spectrum library, and identify whether the coal and rock samples are coal or rock.

3. A coal mine tunnel comprehensive mining method based on acoustic, optical and electrical multi-source data fusion as claimed in claim 1, characterized in that: Based on the secondary recognition result, it is judged whether the cutting head is at the coal-rock interface, specifically: Assume that when the secondary identification result is coal, it is recorded as M, and when the secondary identification result is rock, it is recorded as R; Each batch of coal and rock samples transported on the conveyor belt is tested separately, the test results are recorded, and the proportion of M and R results in the test results is analyzed. When the proportions of the two are equal, it is judged that the cutting head is at the coal-rock interface.

4. A coal mine tunnel comprehensive mining method based on acoustic, optical and electrical multi-source data fusion as claimed in claim 1, characterized in that: Based on the final coal-rock identification results, the coal and rock samples are distributed and transmitted, specifically: Perform image recognition on the coal and rock samples on the conveyor belt and match the coal or rock category labels identified for the coal and rock sample images; The specific location of the coal and rock samples is determined by sensors, and the corresponding coal and rock samples are transported to the corresponding transmission channels according to the category labels that match the coal and rock samples at each location.

5. A coal mine shaft and lane comprehensive mining method based on acoustic, optical and electrical multi-source data fusion as claimed in claim 1, characterized in that: The prediction model of coal quality indicators establishes the regression relationship between the main component information of the spectral lines of elements and the coal quality indicators of standard coal samples through a multivariate linear regression algorithm; the prediction model of coal quality indicators specifically includes an ash prediction model, a calorific value prediction model, a volatile matter prediction model and a sulfur content prediction model, which respectively realize the quantitative prediction of ash content, calorific value, volatile matter and sulfur content.

6. A coal mine shaft and lane comprehensive mining method based on acoustic, optical and electrical multi-source data fusion as claimed in claim 1, characterized in that: The coal type is determined based on the quantitative detection results, and then the different types of coal are sorted, specifically: The quantitative test results of ash, calorific value, volatile matter and sulfur content are input into the trained coal type recognition model to identify the corresponding coal type; Based on the coal type identification results of all coal samples within a set time interval, determine the proportion of different types of coal; Set the speed of the sorter to screen the type of coal with the highest proportion; after the screening is completed, re-determine the quantitative test results of ash, calorific value, volatile matter and sulfur content for the remaining samples, and re-determine the proportion of different types of coal; re-screen the type of coal with the highest proportion; repeat this process until all types of coal samples have been screened.

7. A coal mine shaft and tunnel comprehensive mining system based on acoustic, optical and electrical multi-source data fusion, characterized in that: include: The risk judgment module is used to obtain the drilling parameters of the drill rod during drilling, the microseismic data of the drill bit during drilling feedback, the transient electromagnetic data of the drilling receiving coil, and the gas concentration monitoring data in real time, and judge whether there is a safety risk in the drilling process based on the acquired data; The coal-rock preliminary identification module is used to start coal mining after safely drilling a set distance, obtain the temperature of the cutting teeth of the cutting head during the coal mining process, and obtain the preliminary identification results of the coal and rock using the coal-rock identification model; The coal-rock secondary identification module is used to obtain the reflection spectrum of the coal-rock sample during the transportation process, obtain the secondary identification result of the coal-rock through machine learning; and judge whether the cutting head is at the coal-rock interface based on the secondary identification result; The coal-rock final identification module is used to jointly analyze the preliminary and secondary identification results of coal and rock to obtain the final coal-rock identification results; The coal sorting module is used to distribute and transmit coal and rock samples based on the final coal-rock identification results, perform elemental testing on coal samples, obtain quantitative test results of ash, calorific value, volatile matter and sulfur content based on the elemental test results and using the prediction model of coal quality indicators, determine the coal type based on the quantitative test results, and then realize the sorting of different types of coal; Among them, judging whether there is a safety risk in the drilling process based on the acquired data is as follows: The drilling parameters of the drill pipe during drilling are obtained in real time, and the physical properties of the rock in the formation ahead are obtained using the pre-built drilling parameter-rock physical property inversion model; Real-time acquisition of seismic wave signal data generated by the contact surface between the drill bit and the rock and transmitted to the rock fracture structure, extraction of abnormal characteristics of wave velocity, reflection coefficient and waveform characteristics, and analysis to obtain rock structure information of the front stratum; A pulse current is applied in the borehole through a drilling transmitting coil to generate a primary magnetic field, and a secondary magnetic field is generated when the primary magnetic field encounters a conductive medium through a receiving coil. The abnormal resistance value data is analyzed to achieve directional detection of water sources in the surrounding rock mass and unknown areas. Obtain the concentration of methane and carbon dioxide gases emitted during drilling in real time, and monitor the gas concentration in real time; Based on the obtained rock physical properties, rock structure information, water source conditions and gas concentration, determine whether there are safety risks in the drilling process; The preliminary and secondary identification results of coal and rock are jointly analyzed, specifically: The F1 score was selected as the performance evaluation index to evaluate the coal-rock identification model used for preliminary identification and the machine learning model used for secondary identification. Based on the F1 score evaluation values ​​of the two models, the weights of the initial recognition results and the secondary recognition results are determined; Calculate the score for each recognition category based on the weights: ； Among them, k is the identification category, which is coal or rock; m is the number of models; x is the sample; is the prediction result of the i-th model for sample x; is an indicator function, which means that if If it is identified as category k, the function returns a value of 1, otherwise it returns a value of 0; is the weight of the i-th model; The category with the highest score is selected as the final coal and rock identification result.

8. A terminal device, comprising a processor and a memory, wherein the processor is used to implement instructions; and the memory is used to store multiple instructions, characterized in that: The instructions are suitable for being loaded by a processor and executing the comprehensive mining method for coal mine shafts and lanes based on acoustic, optical and electrical multi-source data fusion as described in any one of claims 1-6.

Images