CN-122014348-A - Geological disaster forecasting and early warning system and method for goaf of coal mine area

CN122014348ACN 122014348 ACN122014348 ACN 122014348ACN-122014348-A

Abstract

The invention belongs to the technical field of coal mining, and discloses a geological disaster forecasting and early-warning system and method for a goaf of a coal mine area. The method comprises the steps of deploying an uphole monitoring layer to calculate the surface subsidence gradient field in real time, deploying a downhole sensing layer to collect rock mass stress-acoustic emission signals in real time, and collecting the rock mass stress-acoustic emission signals The method comprises the steps of acquiring gas concentration data, acquiring a goaf fracture fractal dimension by using a dynamic fracture network evolution model, inverting fracture expansion trend by using microseismic data, acquiring goaf seepage-stress relation, inputting acquired information into a bee colony optimized neural network (BSO-NN), and acquiring grading early warning information of geological disasters of a goaf of a coal mining area by using a bionic intelligent early warning mechanism. The method and the device can be used for more effectively identifying and pre-warning the hidden danger of geological disasters of the coal mine, so that the safety and the sustainability of the coal mine production are ensured.

Inventors

CHEN WEICHONG
MENG ZIYUAN
LIU JIE
DONG LEI
LIU SONG
HOU JUN
LIU WENMING
LI ZEKUN
GAO YAOQUAN
LIN XUDONG

Assignees

国能榆林能源有限责任公司
西安煤科透明地质科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. A geological disaster forecasting and early warning method for a goaf in a coal mine area is characterized by comprising the following steps: S1, deploying an uphole monitoring layer to calculate an earth surface settlement gradient field in real time, deploying an downhole sensing layer to acquire rock mass stress-acoustic emission signals in real time and acquiring the rock mass stress-acoustic emission signals Gas concentration data; S2, acquiring a goaf fracture fractal dimension by using a dynamic fracture network evolution model, inverting fracture expansion trend by microseismic data, and acquiring a goaf seepage-stress relationship; s3, inputting the acquired information into a bee colony optimization neural network BSO-NN, and acquiring grading early warning information of geological disasters of the goaf of the coal mining area through a bionic intelligent early warning mechanism.
2. The method for forecasting and early warning geological disasters in a goaf of a coal mining area according to claim 1 is characterized in that in the step S1, in an uphole monitoring layer, an improved millimeter wave interference radar is operated at 80GHz; the real-time calculation of the surface subsidence gradient field by fusing Beidou/GNSS data comprises the steps of introducing deep learning YOLOv to automatically identify a subsidence funnel area, inverting nonlinear deformation by combining a two-dimensional Gaussian model, solving phase aliasing of large gradient subsidence, and calculating the surface subsidence gradient field.
3. The method for forecasting and warning the geological disaster of the goaf in the coal mine area according to claim 2, wherein the construction of the two-dimensional Gaussian model comprises the following steps: and (3) inverting the parameters of the sedimentation funnel, and constructing a two-dimensional Gaussian model by taking the angle deviation minimization as an objective function for each funnel area identified by YOLOv, wherein the expression is as follows: In the formula, Is the area of the funnel region, In the transverse direction, the direction is the transverse direction, In order to be longitudinal in direction, Is used as a center position of the device, To control the deformation range; and inverting the deformation gradient by nonlinear optimization fitting of the two-dimensional Gaussian model parameters.
4. The method for forecasting and pre-warning geological disasters in a goaf of a coal mine area according to claim 2, wherein the calculation of the surface subsidence gradient field comprises the following steps: Space-time interpolation and gridding, discretizing the fusion displacement data according to a space-time grid, and filling the gap by adopting Kriging interpolation or inverse distance weight IDW; Gradient field generation, comprising calculating sedimentation rate by grid unit based on time series displacement data, and the expression is: In the formula, In order for the time difference to be a function of the time difference, Is displacement difference; gradient vector extraction, and calculating a spatial gradient by applying a Sobel operator or a central difference method, wherein the expression is as follows: The spatially varying intensity and direction of the sedimentation rate are characterized.
5. The method for forecasting and early warning the geological disasters of the goaf of the coal mining area according to claim 1 is characterized in that in the step S1, the underground sensing layer comprises a bionic anchor rod group sensor, a simulated plant root system structure, a piezoelectric ceramic and fiber grating composite sensing unit is implanted, and rock stress-acoustic emission signals are collected in real time; The bionic anchor rod group sensor comprises: Strain-acoustic emission dual-mode sensing, wherein an optical fiber grating or a vibrating wire type miniature strain gauge is embedded in the anchor rod to monitor stress change, and a piezoelectric ceramic acoustic emission sensor is integrated on the surface of the rod body to capture high-frequency elastic waves of rock mass fracture; the grid layout is adopted, a plurality of groups of sensors are distributed along the axial direction of the anchor rod to form a space sensing network, and three-dimensional mapping of a stress field is realized; Designing a bionic structure, and designing a bionic anchor rod group sensor by referring to a biological cilia resonance principle; The real-time acquisition of rock mass stress-acoustic emission signals comprises: (1) The system adopts wavelet transformation and EMD decomposition to eliminate the interference of underground electromechanical equipment and retains the effective acoustic emission characteristics; (2) The edge intelligent processing is carried out, a lightweight LSTM network is deployed at the sensor nodes of the bionic anchor rod group, the acoustic emission event type is identified in real time, and stress mutation points are associated; (3) Establishing a stress-acoustic emission correlation model, wherein the stress is suddenly increased by a high-frequency acoustic emission event, and the local overload early warning of the rock mass is sent; The real-time correction by the Mie scattering model in the rock mass stress-acoustic emission signal is also acquired in real time, comprising the following steps: inputting parameters, dust concentration and particle size distribution, and carrying out an operation process: In the formula, For the historic scattering concentration of the dust, For the actual measured dust scattering concentration, Is the scattering coefficient of the dust and is used for the dust, For the optical path length of the lens, Is the dust concentration.
6. The method for forecasting and warning of geological disasters in a goaf of a coal mining area according to claim 1, wherein in step S2, the dynamic fracture network evolution model includes: Establishing a goaf fracture fractal dimension calculation model: In the formula, For the slit scanning accuracy, Inverting the crack expansion trend by microseismic data for the number of cracks; The coupled percolation-stress equation is: In the formula, To permeability, permeability And stress In relation to each other, In order to achieve the pore pressure, the gas is, Is saturation and is used for predicting water inrush/gas emission risk; As a vector differential operator of the downward gradient, In order for the porosity to be the same, The time is the moment; Inversion of fracture propagation trends by microseismic data includes: S201, data acquisition and system deployment, wherein a three-component detector array is arranged in a target area to form a microseismic monitoring network, and the system acquires P-wave and S-wave elastic wave signals generated by rock fracture in real time and records arrival time, amplitude and polarity; s202, preprocessing data, carrying out band-pass filtering on an original seismic record, and carrying out static correction and dynamic correction; S203, positioning a microseism event, namely calculating a focus coordinate by using P-wave first arrival time of at least 4 sensors and adopting a travel time inversion or grid search method; S204, seismic source mechanism and parameter inversion, judging waveform types through polarization analysis, extracting seismic source mechanism parameters including trend, inclination angle and sliding angle, synchronously optimizing spatial positions and seismic source mechanisms by using a multi-scanning joint waveform inversion algorithm, identifying shearing fracture directions, calculating microseism energy, visual stress and body deformation indexes, and reflecting fracture strength and range.
7. The method for predicting and warning geological disasters in a goaf of a coal mining area according to claim 6, wherein the step of determining the waveform type through polarization analysis comprises the steps of: Three-component data matrix construction, namely constructing a data matrix for three component data segments of vertical Z, north-south N and east-west E recorded by a seismometer: Covariance matrix calculation: in the superscript The transposed matrix is the number of sampling points in the time window; Solving eigenvalue decomposition, namely solving eigenvalue λ1, λ2, and corresponding eigenvectors v1, v2, and v3 of the covariance matrix, wherein the eigenvalue meets λ1 ≫ λ2 and λ3; the polarization direction, the characteristic vector v1 corresponding to the maximum characteristic value lambda 1 indicates the wave propagation direction, wherein the S wave judgment, the characteristic value meets lambda 1 and lambda 2> lambda 3; polarization complexity is high, combined with the axial ratio of the polarization ellipses, axial ratio = λ2/λ1.
8. The method for forecasting and pre-warning the geological disasters of the goaf of the coal mine area according to claim 7, wherein the seismic source mechanism parameters are obtained through inversion of the primary motion polarity or full waveform of the P waves of the multiple stations; a P-wave priming polarity method comprising: the method comprises the steps of identifying the primary motion direction of P waves, marking primary motion symbols of each station in the projection of a hypocenter sphere lower hemisphere, calculating parameters, enabling the primary motion symbols to be divided into four quadrants by searching two orthogonal joint planes, wherein the parameters are defined by the joint plane geometry: Azimuth angle of fault and horizontal plane intersection line relative to true north, range [0 °,360 °); the inclination delta, the included angle of the fault plane and the horizontal plane, the range (0 degree, 90 degrees; sliding angle lambda, the included angle between sliding direction and trend, range of [ -180, 180 ] or [0, 360 °); wherein, lambda is about 0 DEG, left-handed slip, lambda is about + -180 DEG, right-handed slip, lambda is about +90 DEG, reverse fault; A positive fault; The full waveform inversion method includes: Objective function minimization using observed waveforms And theoretical waveform Is solved by lattice search or iterative optimization: In the formula, As a parameter of the model, it is possible to provide, ; P wave and surface wave bands are separated by a CAP method, inversion is carried out respectively, and weighted fitting is carried out: In the formula, In order to separate the P wave band and the surface wave band and invert the weighted fitting value respectively, Are all the weight coefficients of the two-dimensional space model, To separate the observed waveform residuals of the P-wave inversion, To separate the theoretical waveform residuals of the P-wave inversion, To separate the observed waveform residuals of the face inversion, Theoretical waveform residuals for separated face wave inversion; Microseismic energy The calculation formula is based on a vibration velocity damping model: In the formula, The peak vibration speed of the measuring point is directly obtained by a sensor; for the distance of the source from the sensor, Is an energy characteristic coefficient and is related to the field medium characteristics; earthquake moment Physical quantities characterizing fault sliding intensity, calculation dependent waveform inversion: In the formula, In order to achieve a medium density of the material, For the velocity of the seismic wave, The low-frequency level of the seismic source spectrum is obtained through spectrum analysis; As a coefficient of the radiation pattern, Is a directivity correction factor; Correlation index, earthquake moment and moment-magnitude The relation of (2) is: Visual stress Description of average stress per unit area of fault sliding consumption: Body potential change Characterizing the volume change caused by fault sliding, and the expression is as follows: In the formula, Is the shear modulus.
9. The method for predicting and warning geological disasters in a goaf of a coal mining area according to claim 1, wherein in step S3, the swarm optimization neural network BSO-NN comprises: input layer, surface deformation, gas concentration, microseismic energy, fracture density multidimensional heterogeneous data; the hidden layer, dynamic weight adjustment module, simulate the bee foraging path optimization process; the implementation process of risk classification and early warning comprises the following steps: The method comprises the steps of step 1, data acquisition and feature extraction, multi-source monitoring data fusion, generation of time sequence features by integrating multi-dimensional data such as earth surface deformation, gas concentration, microseismic energy, fracture density multi-dimensional heterogeneous data and the like, key risk factor screening, analysis of index and disaster relevance by using Pearson correlation coefficients, and redundant parameter elimination; Step2, constructing a dynamic risk classification model, initializing weight distribution, generating initial weights by using a Tent chaotic map based on historical disaster data, and setting classification threshold values; Stage 3, real-time early warning and response, early warning triggering mechanism, triggering multi-stage response when the model output risk level rises, dynamic threshold adjustment, and recursively updating weight according to new monitoring data to realize model self calibration.
10. A geological disaster forecasting and early warning system for a goaf of a coal mine area, which is characterized in that the system implements the geological disaster forecasting and early warning method for the goaf of the coal mine area according to any one of claims 1 to 9, and the system comprises: The data acquisition module is used for deploying an uphole monitoring layer to calculate the surface subsidence gradient field in real time, deploying a downhole sensing layer to acquire rock mass stress-acoustic emission signals in real time and acquiring the rock mass stress-acoustic emission signals Gas concentration data; The goaf seepage-stress relation acquisition module is used for acquiring goaf fracture fractal dimension by using a dynamic fracture network evolution model, inverting fracture expansion trend by microseismic data and acquiring goaf seepage-stress relation; the grading early warning module is used for inputting the acquired information into the bee colony optimization neural network BSO-NN, and acquiring grading early warning information of geological disasters of the goaf of the coal mining area through a bionic intelligent early warning mechanism.

Description

Geological disaster forecasting and early warning system and method for goaf of coal mine area Technical Field The invention belongs to the technical field of coal mining, and particularly relates to a geological disaster forecasting and early-warning system and method for a goaf of a coal mine area. Background The coal mine is mainly underground mining, the balance of the original geologic body is destroyed during mining, so that geological disasters occur in the coal mine mining process, the occurrence of the geological disasters of the coal mine not only brings economic loss to coal mine enterprises, but also causes serious casualties, the social influence is extremely severe, the occurrence of the geological disasters of the coal mine is caused by the fact that the balance of the original geologic body is destroyed during the coal mine mining, and abnormal phenomena occur in the process of realizing new balance. In the prior art, typical natural disaster comprehensive detection, early warning and information management levels such as mine disasters, geological disasters, environmental disasters and the like cannot be simultaneously realized, dynamic monitoring and real-time forecasting and early warning of disaster risk hidden danger cannot be realized, so that the disaster detection, early risk identification and forecasting and early warning capabilities are poor, and how to improve the uniformity, accuracy and reliability of disaster data acquisition and management is an important direction of current research and development. In the field of coal mining, geological disasters are important factors restricting coal mine safety production and sustainable development. Due to the complexity of the geological environment of the coal mine and the continuous disturbance of mining activities, geological disasters such as ground subsidence and the like seriously threaten the life and property safety of coal mine staff and the normal production operation of the coal mine. Traditionally, monitoring and early warning of coal mine geological disasters mainly depend on manual inspection and simple monitoring equipment, and the method has a plurality of limitations. Firstly, manual inspection is difficult to realize comprehensive and real-time monitoring of the geological environment of the coal mine, and the conditions of missed detection and false detection occur. Secondly, simple monitoring equipment can only provide single geological index data, and occurrence risk and development trend of geological disasters cannot be comprehensively judged. Therefore, the traditional method has obvious defects in the aspects of accuracy and timeliness of geological disaster early warning. In recent years, with the rapid development of the technology of the internet of things, the technology of the internet of things is increasingly widely applied in various fields. The internet of things technology can combine various information sensing devices with the internet to form a huge network through communication sensing technologies such as intelligent sensing, recognition technology, pervasive computing and the like. The technology provides a new solution for monitoring and early warning of coal mine geological disasters. Moreover, the traditional monitoring relies on a single means (such as InSAR earth surface deformation monitoring and gas concentration analysis), so that the data cooperation between the underground and the uphole is not realized, early warning hysteresis is caused, the sensor is easy to be interfered by high humidity and high dust, the data distortion rate is high, and the early warning model is static and does not consider the dynamic coupling effect of mining stress and a geological structure. Disclosure of Invention In order to overcome the problems in the related art, the embodiment of the invention discloses a geological disaster forecasting and early warning system and method for a goaf of a coal mine area. The technical scheme is that the geological disaster forecasting and early warning method for the goaf of the coal mine area comprises the following steps: S1, deploying an uphole monitoring layer to calculate an earth surface settlement gradient field in real time, deploying an downhole sensing layer to acquire rock mass stress-acoustic emission signals in real time and acquiring the rock mass stress-acoustic emission signals Gas concentration data; S2, acquiring a goaf fracture fractal dimension by using a dynamic fracture network evolution model, inverting fracture expansion trend by microseismic data, and acquiring a goaf seepage-stress relationship; s3, inputting the acquired information into a bee colony optimization neural network BSO-NN, and acquiring grading early warning information of geological disasters of the goaf of the coal mining area through a bionic intelligent early warning mechanism. In the step S1, in an uphole monitoring layer, the working frequency band of the improved millime