CN-121881868-B - Quantitative characterization method for deep coal seam outburst risk based on multi-field coupling effect

CN121881868BCN 121881868 BCN121881868 BCN 121881868BCN-121881868-B

Abstract

The invention relates to the field of deep coal mining disaster prevention and control, in particular to a quantitative characterization method of the outburst risk of a deep coal seam based on a multi-field coupling effect, which comprises the following steps of S1, establishing a encodable heat-flow-solidification-four-field coupling control equation set; S2, constructing a chemical damage evolution model, S3, based on field variable distribution obtained by solving the control equation set, establishing an energy mutation instability criterion taking the second-order variation of the total potential energy of the system as a core, determining a critical energy threshold for outburst danger, S4, constructing and training a physical model driving neural network to realize quick prediction of the field variable, S5, utilizing the trained physical model to drive the neural network to carry out quick prediction, and realizing quantitative classification and engineering early warning of the deep coal seam outburst danger based on comparison of a prediction result and the critical energy threshold. The invention provides scientific, reliable and practical technical support for the accurate prevention and control of the deep coal mining outstanding disasters.

Inventors

LI QINGSONG
ZOU QUANLE
ZHU QUANJIE
HE LIN
ZHANG PENG
SHEN ZHENHUA

Assignees

贵州省矿山安全科学研究院有限公司
贵州省煤矿设计研究院有限公司

Dates

Publication Date: 20260512
Application Date: 20260318

Claims (9)

1. The quantitative characterization method of the deep coal seam outburst risk based on the multi-field coupling effect is characterized by comprising the following steps of: S1, establishing a encodable heat-flow-solidification-four-field coupling control equation set, wherein the control equation set corrects an effective stress equation based on occurrence characteristics of a deep coal seam by introducing a temperature stress item and a chemical adsorption stress item, and constructs a permeability dynamic evolution equation containing chemical damage variables so as to completely describe a strong coupling action mechanism among a temperature field, a seepage field, a stress field and a chemical field; S2, constructing a chemical injury evolution model based on chemical potential gradients, and quantitatively representing the injury degree of a chemical field to the intensity of a coal body by defining chemical injury variables and the degradation relation of the chemical injury variables to the coal physical parameters; Wherein, the chemical injury evolution model is constructed, specifically comprising: defining a chemical damage variable based on a chemical potential gradient, wherein the chemical damage variable is used for quantitatively representing the degradation degree of coal body caused by initiation and expansion of microcracks generated by shrinkage of a coal matrix in the gas desorption process, and the value range of the chemical damage variable is between no damage and complete damage; Calibrating a chemical damage coefficient, namely synchronously monitoring acoustic emission signals to represent the development degree of microcracks by carrying out adsorption/desorption experiments on coal samples under different gas pressures, and obtaining a calibration value of the chemical damage coefficient by fitting the relationship between the damage degree and the chemical potential gradient; Establishing a dynamic evolution relation of the chemical injury variable, and substituting a change rule of the gas desorption quantity along with time into a chemical potential gradient calculation formula to realize dynamic update of the chemical injury variable along with the exploitation disturbance time; S3, calculating the total potential energy of the coal-gas system by carrying out global integration on the elastoplastic deformation energy and the gas internal energy of the coal body based on field variable distribution obtained by solving the control equation set, establishing an energy mutation instability criterion taking the second-order variation of the total potential energy of the system as a core, and determining a critical energy threshold value for highlighting danger; S4, constructing and training a physical model driving neural network, wherein the network takes a high-order time discrete iteration function derived based on the control equation set as physical constraint, learns a space-time mapping rule from a current moment field variable and a boundary condition to a next moment field variable, and takes the space-time mapping rule as a proxy model of the control equation set so as to realize quick prediction of the field variable; S5, comprehensively applying the steps S1-S4, carrying out laboratory experiment calibration parameters, field data acquisition, numerical simulation to solve critical threshold values, utilizing a trained physical model to drive a neural network to carry out rapid prediction, and finally realizing quantitative classification and engineering early warning on the outburst risk of the deep coal seam based on comparison of a prediction result and the critical energy threshold value.
2. The quantitative characterization method for the outburst risk of the deep coal seam based on the multi-field coupling effect according to claim 1, wherein in the step S1, a encodable heat-flow-solidification-four-field coupling control equation set is established, and specifically comprises the following steps: constructing a coal body total strain constitutive equation, wherein the coal body total strain consists of four parts of elastic strain, plastic strain, thermal expansion strain and gas adsorption expansion strain so as to describe the deformation behavior of the coal body under the action of multi-field coupling; The modified effective stress equation is constructed by adding a temperature stress item and a chemical adsorption stress item on the basis of the traditional effective stress so as to quantitatively represent a regulation mechanism of temperature change and gas adsorption/desorption effect on the effective stress of the coal body; Constructing a permeability dynamic evolution equation, namely realizing dynamic updating of the permeability of the coal seam in the mining disturbance process by coupling the influences of the volumetric strain, the chemical damage variable and the temperature change of the coal body; And constructing a gas adsorption constitutive equation considering temperature influence, namely accurately describing the change rule of gas adsorption/desorption behaviors in a high-ground-temperature environment by introducing temperature to correct Langmuir adsorption parameters.
3. The quantitative characterization method for the outburst risk of the deep coal seam based on the multi-field coupling effect according to claim 1, wherein the step S2 further comprises: Constructing a coupling relation between chemical damage and coal strength, introducing the chemical damage variable into a strength parameter of the coal, and carrying out degradation correction on the original cohesion and the original internal friction angle of the coal to realize the coupling effect of a chemical field on a stress field, wherein the degradation correction is realized through a combination relation between the chemical damage variable and a strength degradation coefficient.
4. The quantitative characterization method for the outburst risk of the deep coal seam based on the multi-field coupling effect according to claim 1, wherein in the step S3, the calculating the total potential energy of the coal-gas system by performing global integration on the elastoplastic deformation energy and the gas internal energy of the coal comprises: Calculating the total potential energy of a system consisting of two parts of elastoplastic deformation energy of the coal body and internal energy of the gas by carrying out global integration on a research area; wherein the elastoplastic deformation energy of the coal body is calculated based on a stress field and an elastic strain field, and only an elastic strain part is counted, and the plastic deformation part is not counted as dissipation energy; The gas internal energy is calculated according to an ideal gas state equation and a thermodynamic internal energy formula based on a gas pressure field, a temperature field and gas physical parameters.
5. The quantitative characterization method for the outburst risk of the deep coal seam based on the multi-field coupling effect according to claim 4, wherein in the step S3, the establishing of the energy mutation instability criterion with the second order variation of the total potential energy of the system as a core, and the determining of the critical energy threshold of the outburst risk specifically comprise: taking the second-order variation of the total potential energy of the system as a quantitative judgment basis for the stability of the coal-gas system; when the second-order variation is larger than 0, the system is judged to be in a steady state, and the coal bed is free of outburst danger; when the second-order variation is equal to 0, judging that the system is in a critical instability state, and determining the total potential energy of the system at the moment as a critical energy threshold value of the coal seam outburst risk; When the second-order variation is smaller than 0, the system is judged to be in an unsteady state, and coal and gas outburst occurs in the coal bed.
6. The quantitative characterization method for the outburst risk of the deep coal seam based on the multi-field coupling effect according to claim 1, wherein in the step S4, the constructed and trained physical model drives a neural network, and a core of the neural network is positioned as a proxy model of a thermal-flow-solidification-four-field coupling control equation set, and the method specifically comprises the following steps: Constructing a time iteration scheme of an integer order heat-flow-solidification system based on a high-order time discrete format, and taking a space-time mapping rule contained in the iteration scheme as a learning target of the neural network; The higher order time discrete format adopts The time derivative term in the control equation set is discretized in a format, and an iteration function from the last moment field variable distribution and boundary condition to the next moment field variable distribution is constructed; the physical model drives the neural network to replace the traditional numerical simulation gradual solving process by learning the input-output mapping relation of the iterative function, so that the quick prediction of the field variable is realized.
7. The quantitative characterization method of deep coal seam outburst risk based on multi-field coupling effect according to claim 6, wherein the physical model driven neural network adopts an end-to-end network structure comprising an input layer, a hidden layer, a spatio-temporal mapping layer and an output layer, wherein: The input layer receives field variable characteristics and boundary conditions at the current moment, wherein the field variable characteristics simultaneously comprise macroscopic statistical characteristics for describing the overall level of field distribution and spatial variation characteristics for capturing damage localized destabilization precursor information; the hidden layer consists of a plurality of layers of fully-connected neurons and is used for nonlinear feature extraction and mapping; The space-time mapping layer adopts a convolution structure and is used for maintaining the iteration continuity of the field variable on a time sequence; And the output layer outputs the field variable characteristic of the next moment, and the dimension of the field variable characteristic is consistent with the dimension of the field variable characteristic of the input layer.
8. The quantitative characterization method of the deep coal seam outburst risk based on the multi-field coupling effect according to claim 6, wherein in the step S4, a physical model driving neural network is constructed and trained, and the method further comprises the following steps: Generating training data by adopting a mixed data strategy of combining a numerical simulation sample and a field actual measurement sample, wherein a main training set is generated by thermal-flow-solidification numerical simulation, and a calibration set is obtained by inversion of the field actual measurement data; constructing a composite loss function comprising field variable fitting loss and physical iteration constraint loss, wherein the physical iteration constraint loss is constructed based on the iteration function and is used for constraining a neural network prediction result to accord with a physical evolution rule of a heat-flow-solidification system; training a network by adopting an optimization algorithm, and performing parameter fine adjustment on the trained model through a calibration set to control the error of a model prediction result and field actual measurement data within a preset range; the physical model driving neural network after training and calibration is used for on-site rapid prediction, and the field variable characteristics output by the prediction are used as the input of subsequent energy mutation instability criterion calculation and risk classification judgment, so that a rapid prediction-quantitative judgment two-step engineering application mode is formed.
9. The quantitative characterization method for the outburst risk of the deep coal seam based on the multi-field coupling effect according to claim 1, wherein the step S5 comprises: Inputting field variable characteristics of the current moment monitored in real time on the site into a physical model driving neural network after training and calibration, and predicting field variable distribution of the next moment; selecting a corresponding critical energy threshold according to the geological subregion to which the predicted position belongs, or estimating the critical energy threshold through a correlation model, and calculating the ratio of the total potential energy of the real-time system to the critical energy threshold; And judging the outstanding risk grade of the coal seam according to the preset quantitative grading standard and the ratio, and formulating corresponding disaster prevention and control measures according to the outstanding risk grade, so as to realize the accurate early warning and prevention and control of the outstanding disaster.

Description

Quantitative characterization method for deep coal seam outburst risk based on multi-field coupling effect Technical Field The invention relates to the technical field of deep coal mining disaster prevention and control, in particular to a quantitative characterization method of deep coal seam outburst risk based on Thermal-fluid-solid-Chemical (THMC) multi-field coupling effect, which is suitable for accurate evaluation and real-time prediction of deep coal seam coal and gas outburst risk under high ground temperature, high ground stress and high gas pressure occurrence conditions. Background Coal and gas outburst is one of the most serious geological disasters in the deep coal exploitation process, mine safety production and worker life safety are directly threatened, along with the continuous increase of coal exploitation depth, geological conditions of coal bed occurrence are more complex, and the three-high characteristic of high ground temperature, high ground stress and high gas pressure makes THMC multi-field coupling effect more obvious, so that the THMC multi-field coupling effect becomes a core disaster causing factor for inducing coal and gas outburst. The existing deep coal seam outburst risk evaluation method still has the technical defects that a THMC physical model and an Artificial Intelligence (AI) algorithm are coupled to form a logic fault, the physical model lacks a closed and encodable control equation set, the AI algorithm has no physical constraint to cause the prediction result to be practically disjointed with the engineering, a core control equation is a conceptual framework, a parameter calibration method is undefined, key parameters in the engineering implementation process are difficult to obtain, engineering operability of the model is poor, multi-field coupling numerical simulation calculation complexity is high, efficiency is low, engineering requirements of on-site real-time prediction cannot be met, the outburst risk evaluation is mainly performed mainly by qualitative analysis, quantitative criteria lack of definite physical significance, universality of a critical threshold is poor, spatial variation of geological conditions is difficult to adapt, and information loss is easy to exist in field variable feature characterization, instability precursors such as coal damage localization cannot be effectively captured, and accuracy of evaluation results is insufficient. In the prior art, part of methods only consider the coupling effect of a fluid field and a solid field, neglect the regulation and control effect of a temperature field and a chemical field on coal seam outburst, and cannot reflect a disaster-causing mechanism of a deep 'three-high' environment, part of methods introduce an AI algorithm to carry out outburst prediction, but only rely on data driving, do not combine the physical rule of THMC multi-field coupling, have poor model generalization capability and low training stability, part of methods construct a multi-field coupling model, do not design an efficient proxy model, have low calculation efficiency and are difficult to apply in real time on the ground, and meanwhile, most of outburst risk criteria of the existing methods are based on single indexes, lack global criteria based on system energy evolution, and cannot accurately represent the mutation process of a coal body-gas system from steady state to unsteady state. In summary, the prior art is difficult to realize high-precision, quantitative and real-time evaluation of the outburst risk of the deep coal seam, and development of a logic closed loop, high feasibility and theoretical rigor and engineering practicability deep coal seam outburst risk quantitative characterization method is needed to solve the core technical problems of coupling of a physical model and an AI algorithm, implementation of model engineering, efficient prediction, quantitative criteria and the like. Disclosure of Invention Aiming at the technical defects that the THMC physical model and AI algorithm are coupled to form a logical fault, a core control equation lacks a closed encodable form, engineering implementation parameters are difficult to obtain and calibrate, an AI prediction target is disjointed with field variable evolution logic, field variable characteristics represent information loss, critical threshold universality is poor, calculation efficiency is low and the like in the existing deep coal seam outburst risk evaluation method, the invention provides a THMC multi-field coupling effect-based quantitative representation method for the deep coal seam outburst risk. The method comprises the steps of constructing a complete encodable THMC four-field coupling control equation set, determining calibration and acquisition methods of all parameters, improving engineering feasibility of a physical model, defining a chemical damage evolution model, quantitatively representing degradation effect of a chemical field on coal