A method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion

Abstract

The present application provides a coal seam roof and floor damage depth monitoring method based on multimodal AI fusion, which relates to the field of coal mine safety monitoring technology, and obtains multimodal parameters of the coal mine area; inputs the multimodal parameters and geological parameters of the coal mine area into the AI model to obtain the roof damage depth, floor damage depth and their confidence levels in the coal mine area, and calculates the correction coefficients of the roof and floor damage depths; calculates the damage risk index of the coal mine area based on the roof damage depth, floor damage depth, the correction coefficient of the roof damage depth, the correction coefficient of the floor damage depth and the geological parameters, and determines the damage risk level of the coal mine area; uses multimodal parameters to verify each other to improve monitoring accuracy, and combines geological parameters and AI model identification to obtain the roof damage depth and floor damage depth that meet the rock damage constitutive equation, and calculates the damage risk index through the correction coefficient to determine the damage risk level of the coal mine area, so as to further improve monitoring accuracy.

Description

Technical Field

[0001] The present application relates to the technical field of coal mine safety monitoring, and specifically to a method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion. Background Art

[0002] Coal mine subsidence monitoring and early warning are crucial components of coal mine safety production. They are primarily used to predict and prevent surface subsidence and geological disasters caused by coal mining activities. However, in actual coal mine subsidence monitoring, traditional roof failure monitoring relies on delamination meters or borehole stress gauges, which struggle to capture the dynamic expansion of roof cracks and failure depth in real time. This lack of integration with real-time geomechanical data leads to inaccurate water inrush risk assessments. Furthermore, the roof and floor monitoring systems operate independently, lacking the ability to correlate and analyze multi-source data, making it difficult to comprehensively assess the impact of mining. Therefore, a high-precision method for monitoring roof and floor failure depth is urgently needed. Summary of the Invention

[0003] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide a method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion.

[0004] The present application provides a method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion, comprising: using a multimodal sensor to obtain multimodal parameters of a coal mine area; inputting the multimodal parameters and geological parameters of the coal mine area into an AI model to obtain the roof damage depth and its confidence level, and the floor damage depth and its confidence level of the coal mine area; wherein the AI model includes an encoder module, a physical constraint module and an output module, wherein the encoder module extracts the long-term time series dependency of the multimodal parameters, the physical constraint module embeds the rock damage constitutive equation, and the output module Based on the long-term time-series dependency of multimodal parameters, an output result satisfying the rock damage constitutive equation is output; based on the confidence level of the roof damage depth and the confidence level of the bottom plate damage depth, a correction coefficient of the roof damage depth and a correction coefficient of the bottom plate damage depth are calculated; based on the roof damage depth, the bottom plate damage depth, the correction coefficient of the roof damage depth, the correction coefficient of the bottom plate damage depth and the geological parameters, a damage risk index of the coal mine area is calculated; based on the damage risk index of the coal mine area, the damage risk level of the coal mine area is determined.

[0005] In one embodiment, the use of multimodal sensors to obtain multimodal parameters of the coal mine area includes: using distributed optical fiber sensors, array acoustic emission sensors, pressure sensors, microseismic sensors, infrared thermal imagers and laser scanners to respectively obtain the delamination strain and microbending loss of the coal mine area, the acoustic wave signal of roof crack expansion, the vertical stress and pore water pressure gradient, the energy and frequency of floor rupture events, the top and bottom plate surface temperature fields and the top and bottom plate surface deformation point cloud data.

[0006] In one embodiment, the calculating the damage risk index of the coal mine area based on the roof damage depth, the floor damage depth, the correction coefficient of the roof damage depth, the correction coefficient of the floor damage depth, and the geological parameters includes:

[0007]

[0008] Where R is the damage risk index, Hb is the top plate damage depth, Hc is the bottom plate damage depth, α is the correction coefficient of the top plate damage depth, β is the correction coefficient of the bottom plate damage depth, Da is the distance from the bottom plate to the aquifer, and Lc is the maximum span of the top plate.

[0009] In one embodiment, determining the destruction risk level of the coal mine area based on the destruction risk index of the coal mine area includes: if the destruction risk index is greater than or equal to a first preset value and less than a second preset value, adjusting the coal mining machine propulsion speed and strengthening manual inspections.

[0010] In one embodiment, determining the damage risk level of the coal mine area based on the damage risk index of the coal mine area includes: if the damage risk index is greater than or equal to the second preset value, suspending mining and starting support reinforcement.

[0011] In one embodiment, the method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion further includes: aligning the multimodal parameters based on the acquisition time and acquisition position of the multimodal parameters.

[0012] In one embodiment, aligning the multimodal parameters based on the acquisition time and acquisition position of the multimodal parameters includes: aligning the multimodal parameters based on the acquisition time of the multimodal parameters to obtain a first parameter sequence; wherein the first parameter sequence includes the multimodal parameters aligned along the time dimension; and using interpolation to complete the parameter values of the non-completely corresponding acquisition positions in the first parameter sequence to obtain a second parameter sequence; wherein the non-completely corresponding acquisition position indicates that the parameters at the acquisition position cannot cover all modalities.

[0013] In one embodiment, the method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion also includes: constructing a loss function of the AI model; and training the AI model based on the loss function of the AI model.

[0014] In one embodiment, the loss function of constructing the AI model includes:

[0015]

[0016] Among them, Loss represents the loss function, is the predicted value The mean square error between the true value y and is the weight coefficient, is the stress field divergence, It is the external force per unit volume of rock mass.

[0017] In one embodiment, the calculating the correction coefficient of the top plate damage depth and the correction coefficient of the bottom plate damage depth based on the confidence level of the top plate damage depth and the confidence level of the bottom plate damage depth includes: determining the sign of the correction coefficient of the top plate damage depth and the sign of the correction coefficient of the bottom plate damage depth based on the top plate damage depth and the confidence level of the bottom plate damage depth; calculating the magnitude of the correction coefficient of the top plate damage depth and the magnitude of the correction coefficient of the bottom plate damage depth based on the confidence level of the top plate damage depth and the confidence level of the bottom plate damage depth; determining the correction coefficient of the top plate damage depth and the correction coefficient of the bottom plate damage depth based on the sign of the correction coefficient of the top plate damage depth, the sign of the correction coefficient of the bottom plate damage depth, the magnitude of the correction coefficient of the top plate damage depth and the magnitude of the correction coefficient of the bottom plate damage depth.

[0018] The present application provides a method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion, which obtains the multimodal parameters of the coal mine area by adopting a multimodal sensor; the multimodal parameters and the geological parameters of the coal mine area are input into the AI model to obtain the roof damage depth and its confidence level, the floor damage depth and its confidence level of the coal mine area; wherein the AI model includes an encoder module, a physical constraint module and an output module, the encoder module extracts the long-term dependency of the multimodal parameters, the physical constraint module embeds the rock damage constitutive equation, and the output module outputs the output result that satisfies the rock damage constitutive equation based on the long-term dependency of the multimodal parameters; based on the confidence level of the roof damage depth and the bottom damage depth, the output module generates a new model for the damage depth of the coal seam roof and the floor damage depth. The confidence level of the roof failure depth is calculated, and the correction coefficient of the roof failure depth and the correction coefficient of the bottom plate failure depth are calculated; based on the roof failure depth, the bottom plate failure depth, the correction coefficient of the roof failure depth, the correction coefficient of the bottom plate failure depth and the geological parameters, the damage risk index of the coal mine area is calculated; based on the damage risk index of the coal mine area, the damage risk level of the coal mine area is determined; multi-modal parameters are used to verify each other to improve the monitoring accuracy, and the roof failure depth and bottom plate failure depth that meet the rock damage constitutive equation are obtained by combining geological parameters and AI model identification, and the damage risk index is calculated by the correction coefficient to determine the damage risk level of the coal mine area, so as to further improve the monitoring accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0020] Figure 1 It is a structural schematic diagram of a coal seam roof and floor damage depth monitoring system based on multimodal AI fusion provided by an exemplary embodiment of the present application.

[0021] Figure 2 It is a flow chart of a method for monitoring the depth of coal seam roof and floor damage based on multimodal AI fusion provided by an exemplary embodiment of the present application.

[0022] Figure 3 It is a structural diagram of an electronic device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0023] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0024] Figure 1 This is a schematic diagram of the structure of a coal seam roof and floor damage depth monitoring system based on multimodal AI fusion provided by an exemplary embodiment of the present application. Figure 1 As shown, the coal seam roof and floor damage depth monitoring system based on multimodal AI fusion includes: a multimodal perception layer, an edge intelligence layer, a cloud analysis layer and an early warning feedback layer; wherein, the multimodal perception layer is set in the monitored coal mine area to monitor multiple modal parameters of the coal mine area; the edge intelligence layer is set in the underground substation in the coal mine area, and performs data preprocessing and real-time feature fusion by carrying an AI acceleration chip; the cloud analysis layer obtains the monitoring results of the coal mine area based on multimodal parameter identification by constructing an AI model; the early warning feedback layer is used to perform risk classification based on the monitoring results and can construct a dynamic cloud map of the roof and floor damage depth, superimpose geological structure data, and assist in decision-making.

[0025] Figure 2 This is a flow chart of a method for monitoring the depth of coal seam roof and floor damage based on multimodal AI fusion provided by an exemplary embodiment of the present application. The method for monitoring the depth of coal seam roof and floor damage based on multimodal AI fusion is applied to the above-mentioned system for monitoring the depth of coal seam roof and floor damage based on multimodal AI fusion, such as Figure 2 As shown in FIG, the coal seam roof and floor damage depth monitoring method based on multimodal AI fusion includes the following steps:

[0026] Step 110: Use a multimodal sensor to obtain multimodal parameters of the coal mine area.

[0027] This application uses sensors of multiple modes to simultaneously collect status parameters of the coal mine area, so as to monitor the status parameters of the coal mine area from multiple dimensions, so as to improve the monitoring reliability of the coal mine area.

[0028] Step 120: Input the multimodal parameters and the geological parameters of the coal mine area into the AI model to obtain the roof failure depth and its confidence level, and the floor failure depth and its confidence level of the coal mine area.

[0029] Among them, the AI model includes an encoder module, a physical constraint module and an output module. The encoder module extracts the long-term temporal dependency of multimodal parameters, the physical constraint module embeds the rock damage constitutive equation, and the output module outputs the output result that satisfies the rock damage constitutive equation based on the long-term temporal dependency of multimodal parameters. This application uses the physical constraint module to constrain the output of the AI model to conform to the laws of mechanics, thereby obtaining the top plate failure depth and its confidence level, and the bottom plate failure depth and its confidence level that conform to the laws of mechanics. Among them, multimodal parameters may include strain rate, acoustic emission energy, water pressure change, temperature gradient, and geological parameters may include rock strength, joint density, and aquifer location.

[0030] Step 130: Calculate the correction coefficient of the top plate failure depth and the correction coefficient of the bottom plate failure depth based on the confidence level of the top plate failure depth and the confidence level of the bottom plate failure depth.

[0031] After calculating the confidence levels of the top plate damage depth and the bottom plate damage depth, the present application calculates the correction coefficients of the top plate damage depth and the bottom plate damage depth based on the confidence levels of the top plate damage depth and the bottom plate damage depth.

[0032] Step 140: Calculate a damage risk index for the coal mine area based on the roof damage depth, the floor damage depth, the correction coefficient of the roof damage depth, the correction coefficient of the floor damage depth, and geological parameters.

[0033] After calculating the correction coefficient of the roof damage depth and the correction coefficient of the bottom plate damage depth, this application calculates the damage risk index of the coal mine area based on the roof damage depth, the bottom plate damage depth, the correction coefficient of the roof damage depth, the correction coefficient of the bottom plate damage depth and the geological parameters.

[0034] Step 150: Determine the damage risk level of the coal mine area based on the damage risk index of the coal mine area.

[0035] This application calculates the destruction risk index of the coal mine area and determines the destruction risk level of the coal mine area based on the destruction risk index of the coal mine area, thereby achieving monitoring of the coal mine area.

[0036] The present application provides a method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion, which obtains the multimodal parameters of the coal mine area by adopting a multimodal sensor; the multimodal parameters and the geological parameters of the coal mine area are input into the AI model to obtain the roof damage depth and its confidence level, the floor damage depth and its confidence level of the coal mine area; wherein the AI model includes an encoder module, a physical constraint module and an output module, the encoder module extracts the long-term dependency of the multimodal parameters, the physical constraint module embeds the rock damage constitutive equation, and the output module outputs the output result that satisfies the rock damage constitutive equation based on the long-term dependency of the multimodal parameters; based on the confidence level of the roof damage depth and the bottom damage depth, the output module generates a new model for the damage depth of the coal seam roof and the floor damage depth. The confidence level of the roof failure depth is calculated, and the correction coefficient of the roof failure depth and the correction coefficient of the bottom plate failure depth are calculated; based on the roof failure depth, the bottom plate failure depth, the correction coefficient of the roof failure depth, the correction coefficient of the bottom plate failure depth and the geological parameters, the damage risk index of the coal mine area is calculated; based on the damage risk index of the coal mine area, the damage risk level of the coal mine area is determined; multi-modal parameters are used to verify each other to improve the monitoring accuracy, and the roof failure depth and bottom plate failure depth that meet the rock damage constitutive equation are obtained by combining geological parameters and AI model identification, and the damage risk index is calculated by the correction coefficient to determine the damage risk level of the coal mine area, so as to further improve the monitoring accuracy.

[0037] In one embodiment, the specific implementation method of the above-mentioned step 110 can be: using distributed optical fiber sensors, array acoustic emission sensors, pressure sensors, microseismic sensors, infrared thermal imagers and laser scanners to respectively obtain the delamination strain and microbending loss of the coal mine area, the acoustic wave signal of roof crack expansion, the vertical stress and pore water pressure gradient, the energy and frequency of floor rupture events, the surface temperature field of the top and bottom plates, and the deformation point cloud data of the top and bottom plates.

[0038] Specifically, the present application can arrange monitoring units on the roof: distributed optical fiber sensors, array acoustic emission sensors, wherein the distributed optical fiber sensors are arranged along the direction of the roof rock layer, for example, one optical fiber is arranged every 8m along the direction of the working face roof, with a total of 5 optical fibers arranged, covering the fault zone and a range of 20m on both sides, and the sampling frequency of the distributed optical fiber sensor is 200Hz, which is used to monitor the separation strain and microbending loss. The present application can also install 8 array acoustic emission sensors on the roof anchor to form a three-dimensional monitoring network, and set the trigger threshold of the array acoustic emission sensor to 40dB to filter out the underground mechanical noise and capture the roof cracks. The acoustic wave signal of the gap expansion is used to locate the three-dimensional coordinates of the rupture source; the present application can also arrange monitoring units on the bottom plate: mining intrinsically safe pressure sensors and high-precision microseismic sensor networks, wherein the mining intrinsically safe pressure sensors are embedded in the bottom plate boreholes to monitor the vertical stress and pore water pressure gradient, and the high-precision microseismic sensor network is arranged in the bottom plate rock layer to identify the energy and frequency of bottom plate rupture events; the present application can also arrange environmental perception units: infrared thermal imagers and laser scanners, wherein the infrared thermal imagers scan the surface temperature field of the top and bottom plates in real time to identify abnormal water seepage areas, and the laser scanners periodically (for example, 30 minutes) obtain the surface deformation point cloud data of the top and bottom plates.

[0039] In one embodiment, the specific implementation of step 140 may be:

[0040]

[0041] Among them, R is the damage risk index, Hb is the top plate damage depth, and Hc is the bottom plate damage depth.

[0042] This application calculates the destruction risk index of the coal mine area based on the roof destruction depth Hb, the bottom plate destruction depth Hc, the correction coefficient α of the roof destruction depth, the correction coefficient β of the bottom plate destruction depth, the distance Da from the bottom plate to the aquifer and the roof limit span Lc, and evaluates the risk of the coal mine area based on the destruction risk index.

[0043] In one embodiment, the specific implementation of the above step 150 may be: if the damage risk index is greater than or equal to a first preset value and less than a second preset value, the shearer propulsion speed is adjusted and manual inspections are strengthened.

[0044] If the damage risk index is greater than or equal to the first preset value (for example, 0.5) and less than the second preset value (for example, 0.8), it means that there are certain safety hazards in the coal mine area at this time, but it will not cause a safety accident. In this case, the coal mining machine advancement speed will be adjusted and manual inspections will be strengthened to slow down the construction speed and strengthen manual inspections, strengthen safety monitoring, and ensure personnel safety.

[0045] In one embodiment, the specific implementation of the above step 150 may be: if the damage risk index is greater than or equal to the second preset value, then suspending mining and starting support reinforcement.

[0046] If the damage risk index is greater than or equal to the second preset value, it means that the coal mine area is likely to cause a safety accident at this time, so mining is suspended and support reinforcement is started to avoid safety accidents.

[0047] In one embodiment, the present application can arrange a distributed optical fiber every 10 meters on the roof of the coal mine area, with a sampling frequency ≥100Hz, and extract the optical fiber strain mutation point (for example, the strain change rate >0.1% / minute) at the edge intelligent layer to trigger the focused collection of the acoustic emission sensor. The cloud analysis layer combines the strain mutation position, the acoustic emission event energy (for example, greater than 50dB) and the rock strength parameters to calculate the roof damage depth Hc. If Hc exceeds 2 / 3 of the roof thickness, mining is suspended and support reinforcement is started.

[0048] In one embodiment, the present application can use microseismic sensors to monitor clusters of bottom plate rupture events. If there are ≥5 times within 10 minutes and the energy is >1×10³J, the pore water pressure drop rate is synchronously detected. The cloud-based analysis layer inputs the above characteristics and the aquifer distance Da, outputs Hb, and calculates the damage risk index. Corresponding protective operations are performed based on the damage risk index.

[0049] In one embodiment, the above-mentioned coal seam roof and floor damage depth monitoring method based on multimodal AI fusion may also include: aligning the multimodal parameters based on the acquisition time and acquisition position of the multimodal parameters.

[0050] After acquiring multimodal parameters using a multimodal sensor, the present application performs an alignment operation on the multimodal parameters based on the acquisition time and acquisition position to ensure that the multimodal parameters remain aligned in time and space, thereby improving the accuracy of the monitoring results.

[0051] In one embodiment, the above-mentioned coal seam roof and floor damage depth monitoring method based on multimodal AI fusion may also include: aligning multimodal parameters based on the acquisition time of the multimodal parameters to obtain a first parameter sequence; wherein, the first parameter sequence includes multimodal parameters aligned along the time dimension; using interpolation method to complete the parameter values of the non-completely corresponding acquisition positions in the first parameter sequence to obtain a second parameter sequence; wherein, the non-completely corresponding acquisition position indicates that the parameters at the acquisition position cannot cover all modes.

[0052] Specifically, after collecting multimodal parameters (time series parameters), the present application aligns the first parameter sequence along the time dimension to obtain the first parameter sequence. Since there may be a certain misalignment between the parameter points (collection points) of different modes after time alignment, the incomplete parameters at a certain moment are supplemented by interpolation. The specific supplementation method can be: obtain the number of parameters corresponding to all time points in the first parameter sequence, and select the time point where the number of parameters is greater than a preset value (such as a set value or a multiple of the number of parameter modes, such as 0.5 times) as the target time point, and perform difference estimation on the parameter sequence corresponding to the mode for which no parameters are collected at the target time point to obtain the parameter value of the target time point, thereby obtaining the parameter values under all modes corresponding to the target time point, so as to provide more meaningful values for the subsequent mutual calibration between different modal parameters, thereby improving the accuracy of the data.

[0053] After obtaining the first parameter sequence, the present application can also obtain the time series curve of the parameter sequence corresponding to each mode by fitting, and determine the mutation point or extreme point on the time series curve according to the slope, and align the first parameter sequence based on each mutation point and extreme point. For example, the extreme point on all first parameter sequences within the same time period can be used as a reference point, and the first parameter sequence can be aligned with the reference point as the standard to obtain the second parameter sequence, thereby avoiding the problem of inaccurate data alignment due to differences in the time when each sensor collects and transmits data, and further improving the accuracy of the data.

[0054] In one embodiment, the above-mentioned coal seam roof and floor damage depth monitoring method based on multimodal AI fusion may also include: constructing a loss function of the AI model; and training the AI model based on the loss function of the AI model.

[0055] This application constructs an AI model and trains the AI model to obtain a model whose recognition accuracy meets actual needs. During the training process, a loss function of the AI model is constructed and the AI model is trained with this loss function to ensure that its recognition accuracy meets the preset accuracy requirements.

[0056] In one embodiment, constructing the loss function of the AI model includes:

[0057]

[0058] Among them, Loss represents the loss function, is the predicted value The mean square error between the true value y and is the weight coefficient, is the stress field divergence, It is the external force per unit volume of rock mass.

[0059] Represents the model prediction value The difference between the model output and the true value y reflects the fitting accuracy of the data-driven part, ensuring that the model output is consistent with the measured damage depth, thereby improving the credibility of the data. Among them, y represents the measured top or bottom plate damage depth (such as the measured value of the borehole, the microseismic positioning inversion value, etc.). is the damage depth predicted by the model. It is used to adjust the contribution ratio of the data-driven term (MSE) in the total loss function. The larger its value is, the more the model tends to fit the measured data, but it will sacrifice the consistency of physical laws. Otherwise, it will emphasize physical constraints more. Represents the stress tensor (including normal stress and shear stress), which characterizes the internal force distribution per unit area inside the rock mass. It represents the divergence of the stress field and physically represents the net force per unit volume of rock mass (derived from the stress gradient). is the body force, such as gravity, seepage force, etc., which represents the external force on the unit volume of rock mass. represents a physical constraint that forces the stress field predicted by the model to ) satisfies the mechanical equilibrium equation ( ), It represents the weight coefficient of the physical constraint term, which controls the constraint strength of the mechanical equilibrium equation on the model training. The larger its value is, the more strictly the stress field predicted by the model satisfies the mechanical laws, but it will reduce the flexibility of fitting the measured data.

[0060] In one embodiment, the specific implementation method of the above-mentioned step 130 can be: based on the top plate damage depth and the bottom plate damage depth, determine the sign of the correction coefficient of the top plate damage depth and the sign of the correction coefficient of the bottom plate damage depth; based on the confidence level of the top plate damage depth and the confidence level of the bottom plate damage depth, calculate the size of the correction coefficient of the top plate damage depth and the size of the correction coefficient of the bottom plate damage depth; based on the sign of the correction coefficient of the top plate damage depth, the sign of the correction coefficient of the bottom plate damage depth, the size of the correction coefficient of the top plate damage depth and the size of the correction coefficient of the bottom plate damage depth, determine the correction coefficient of the top plate damage depth and the correction coefficient of the bottom plate damage depth.

[0061] Specifically, the present application uses the calculated top plate failure depth and bottom plate failure depth in combination with multimodal parameters and geological parameters to determine whether the correction of the current top plate failure depth and bottom plate failure depth is increased or decreased, that is, whether the modified top plate failure depth and bottom plate failure depth are based on the current top plate failure depth and bottom plate failure depth plus or minus the correction value, and based on the confidence of the top plate failure depth and the confidence of the bottom plate failure depth, calculates the size of the correction coefficient of the top plate failure depth and the size of the correction coefficient of the bottom plate failure depth. Specifically, the lower the confidence of the top plate failure depth and the confidence of the bottom plate failure depth, the larger the corresponding correction coefficient. For example, if the correction coefficient is determined to be positive and the confidence is low, the corresponding correction coefficient can be set to 1.2. If the correction coefficient is determined to be positive and the confidence is high, the corresponding correction coefficient can be set to 1.05.

[0062] Below, reference Figure 3 The electronic device according to the embodiment of the present application is described. The electronic device may be either or both of the first device and the second device, or a standalone device independent of them, and the standalone device may communicate with the first device and the second device to receive collected input signals from them.

[0063] Figure 3 A block diagram of an electronic device according to an embodiment of the present application is illustrated.

[0064] like Figure 3 As shown, the electronic device 10 includes one or more processors 11 and a memory 12 .

[0065] The processor 11 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0066] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), a hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may execute the program instructions to implement the methods of the various embodiments of the present application described above and / or other desired functions. Various contents such as input signals, signal components, and noise components may also be stored in the computer-readable storage medium.

[0067] In one example, the electronic device 10 may further include an input device 13 and an output device 14 , and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0068] When the electronic device is a stand-alone device, the input device 13 may be a communication network connector, configured to receive collected input signals from the first device and the second device.

[0069] In addition, the input device 13 may also include, for example, a keyboard, a mouse, and the like.

[0070] The output device 14 can output various information to the outside, including determined distance information, direction information, etc. The output device 14 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0071] Of course, to simplify, Figure 3 Only some of the components related to the present application in the electronic device 10 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, the electronic device 10 may further include any other appropriate components according to specific application scenarios.

[0072] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the method according to various embodiments of the present application described in the above-mentioned "Exemplary Method" section of this specification.

[0073] The computer program product may be written in any combination of one or more programming languages to implement the program code for performing the operations of the embodiments of the present application, including object-oriented programming languages such as Java, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0074] In addition, an embodiment of the present application may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, enable the processor to execute the steps of the method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0075] The computer-readable storage medium may be any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0076] The basic principles of the present application have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in this application are merely illustrative and not restrictive, and it should not be assumed that these advantages, strengths, and effects are required of each embodiment of this application. In addition, the specific details disclosed above are merely illustrative and facilitating understanding, and are not restrictive. The above details do not limit this application to necessarily being implemented using the above specific details.

[0077] The block diagrams of the devices, devices, equipment, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, or configured in any manner. Words such as "include," "comprise," "have," and the like are open-ended words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0078] It should also be noted that in the apparatus, device, and method of the present application, each component or each step can be decomposed and / or recombined, and such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0079] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0080] The above description has been provided for the purpose of illustration and description. Furthermore, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion, characterized in that: include: Use multimodal sensors to obtain multimodal parameters of the coal mine area; Inputting the multimodal parameters and the geological parameters of the coal mine area into an AI model to obtain the roof failure depth and its confidence level, and the floor failure depth and its confidence level of the coal mine area; wherein the AI model includes an encoder module, a physical constraint module, and an output module, the encoder module extracts the long-term temporal dependency of the multimodal parameters, the physical constraint module embeds a rock damage constitutive equation, and the output module outputs an output result that satisfies the rock damage constitutive equation based on the long-term temporal dependency of the multimodal parameters; Calculating a correction coefficient for the top plate failure depth and a correction coefficient for the bottom plate failure depth based on the confidence level of the top plate failure depth and the confidence level of the bottom plate failure depth; Calculating a damage risk index for the coal mine area based on the roof damage depth, the floor damage depth, a correction coefficient for the roof damage depth, the correction coefficient for the floor damage depth, and the geological parameters; determining a damage risk level of the coal mine area based on a damage risk index of the coal mine area; Calculating the damage risk index of the coal mine area based on the roof damage depth, the floor damage depth, the correction coefficient of the roof damage depth, the correction coefficient of the floor damage depth, and the geological parameters includes: ;in, R is the damage risk index, Hb is the top plate failure depth, Hc is the bottom plate damage depth, α is the correction factor for the top plate failure depth, β is the correction coefficient of the bottom plate failure depth, Da is the distance from the bottom plate to the aquifer, Lc is the maximum span of the top plate.

2. The method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion according to claim 1 is characterized in that: The method of obtaining multimodal parameters of a coal mine area by using a multimodal sensor includes: Distributed fiber optic sensors, array acoustic emission sensors, pressure sensors, microseismic sensors, infrared thermal imagers and laser scanners are used to obtain the delamination strain and microbending loss of the coal mine area, the acoustic wave signal of roof crack expansion, the vertical stress and pore water pressure gradient, the energy and frequency of floor rupture events, the surface temperature field of the roof and floor plates, and the deformation point cloud data of the roof and floor plates.

3. The method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion according to claim 1 is characterized in that: Determining the destruction risk level of the coal mine area based on the destruction risk index of the coal mine area includes: If the damage risk index is greater than or equal to the first preset value and less than the second preset value, the shearer advance speed is adjusted and manual inspections are strengthened.

4. The method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion according to claim 3 is characterized in that: Determining the destruction risk level of the coal mine area based on the destruction risk index of the coal mine area includes: If the damage risk index is greater than or equal to the second preset value, mining is suspended and support reinforcement is started.

5. The method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion according to claim 1 is characterized in that: The coal seam roof and floor damage depth monitoring method based on multimodal AI fusion also includes: The multimodal parameters are aligned based on the acquisition time and acquisition position of the multimodal parameters.

6. The method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion according to claim 5 is characterized in that: The aligning the multimodal parameters based on the acquisition time and acquisition position of the multimodal parameters includes: Aligning the multimodal parameters based on their acquisition time to obtain a first parameter sequence; wherein the first parameter sequence includes the multimodal parameters aligned along the time dimension; An interpolation method is used to complete the parameter values of the incompletely corresponding acquisition positions in the first parameter sequence to obtain a second parameter sequence; wherein the incompletely corresponding acquisition position indicates that the parameters at the acquisition position cannot cover all modes.

7. The method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion according to claim 1 is characterized in that: The coal seam roof and floor damage depth monitoring method based on multimodal AI fusion also includes: Constructing a loss function for the AI model; The AI model is trained based on the loss function of the AI model.

8. The method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion according to claim 7 is characterized in that: The loss function for constructing the AI model includes: ； in, Loss represents the loss function, MSE(y, ŷ ) is the predicted value and the true value y The mean square error between and is the weight coefficient, is the stress field divergence, It is the external force per unit volume of rock mass.

9. The method for monitoring the damage depth of coal seam roof and floor based on multimodal AI fusion according to claim 1 is characterized in that: The calculating of the correction coefficient of the top plate failure depth and the correction coefficient of the bottom plate failure depth based on the confidence level of the top plate failure depth and the confidence level of the bottom plate failure depth includes: Determining the signs of the correction coefficients for the top plate failure depth and the bottom plate failure depth based on the top plate failure depth and the bottom plate failure depth; Calculating the correction coefficients of the top plate failure depth and the bottom plate failure depth based on the confidence levels of the top plate failure depth and the bottom plate failure depth; The correction coefficient of the top plate failure depth and the correction coefficient of the bottom plate failure depth are determined based on the sign of the correction coefficient of the top plate failure depth, the sign of the correction coefficient of the bottom plate failure depth, the size of the correction coefficient of the top plate failure depth and the size of the correction coefficient of the bottom plate failure depth.

Images