Slope deformation dynamic early warning method for fire zone of open-pit mine based on multi-factor coupling driving

Abstract

The application discloses a kind of based on multi-factor coupling driving's open-pit mine fire burning area slope deformation dynamic early warning method, its method includes: obtaining the meteorological data of study area, engineering disturbance data, rock mass response data and respectively obtains characteristic index;With the single drive or / and two two coupling of characteristic index as driving item and add driving item weight jointly driven to construct dynamic early warning index;With the normalized acceleration of GNSS monitoring data as observation quantity, the observation relationship between weight and acceleration response is constructed;With the history sequence of dynamic early warning index ascending order sorting, adaptive threshold model is constructed, adaptive threshold model is according to quantile dynamic early warning index grade interval division to history sequence and determine the grade interval of dynamic early warning index belonging of current time and output slope risk level.The application realizes the dynamic early warning and determination of fire burning area slope risk state, can truly reflect the slope risk evolution process under complex working condition, improves adaptive capacity.

Description

Technical Field

[0001] This invention relates to the field of open-pit mine slope stability monitoring and risk early warning, and particularly to a dynamic early warning method for slope deformation in open-pit mine fire-affected areas based on multi-factor coupling. Specifically, it relates to a dynamic early warning method for slope deformation based on multi-factor coupling and dynamic weight inversion, which is designed for complex geological environments in open-pit mine fire-affected areas and integrates multi-source time-series information such as meteorological environment, engineering mining activities and rock mass damage. Background Technology

[0002] Slope stability in open-pit mines is a critical issue in mine safety. For fire-affected slopes formed after long-term underground coal fires, the rock mass is subjected to the combined effects of high-temperature burning, accumulated thermal damage, and subsequent weathering and water seepage. This significantly reduces structural integrity, enhances the development of fissures and pores, and continuously deteriorates permeability and mechanical properties, leading to a more abrupt and time-varying stability evolution process compared to general slopes, resulting in a higher risk of instability. Existing slope early warning methods are mostly designed for general slope scenarios, primarily relying on single monitoring indicators, empirical factors, fixed weights, or static thresholds for risk assessment. There is a lack of dynamic early warning methods specifically for fire-affected slopes in open-pit mines. In particular, existing methods struggle to uniformly characterize the multiple sources of disaster-causing factors and their coupling effects, such as meteorological changes, engineering disturbances, and the evolution of thermal damage and structural deterioration within the fire-affected slope area, and also find it difficult to dynamically adjust the contribution of each driving factor based on the slope's response state. Therefore, there is an urgent need to propose a dynamic early warning method for slopes in open-pit mine fire areas to solve the problems that existing technologies cannot uniformly characterize multiple disaster-causing factors and their coupling effects, cannot dynamically reflect the changes in the contribution of each driving factor, and are lagging in risk level identification. Summary of the Invention

[0003] The purpose of this invention is to address the complex, time-varying, and abruptly changing stability evolution of open-pit mine fire-affected slopes under the combined influence of multiple factors, including meteorological changes, engineering disturbances, and rock mass deterioration. Existing slope early warning technologies are mostly designed for general slope scenarios and lack a dynamic early warning method specifically for open-pit mine fire-affected slopes. Furthermore, existing technologies often rely on fixed thresholds and static weights, failing to reflect the coupling effects of multiple factors and the dynamic evolution of risks. This invention provides a dynamic early warning method for open-pit mine fire-affected slope deformation based on multi-factor coupling. Specifically targeting open-pit mine fire-affected slopes, it integrates multi-source time-series monitoring information from meteorological environment, engineering disturbances, and rock mass damage to construct a dynamic early warning index for slope deformation that considers the coupling effects of multiple factors. Based on the statistical distribution characteristics of this dynamic early warning index, a risk level discrimination mechanism that updates over time is established, enabling dynamic early warning and judgment of the risk status of fire-affected slopes. This improves the reliability and engineering applicability of the early warning results and provides a basis for the safety management of fire-affected slopes.

[0004] The objective of this invention is achieved through the following technical solution: A dynamic early warning method for slope deformation in open-pit mine fire-affected areas based on multi-factor coupling, the method comprising: S1. Obtain meteorological data, engineering disturbance data, and rock mass response data of the study area, and obtain characteristic indicators including meteorological environmental risk index, engineering disturbance load index, and rock mass damage degree in the burned area; obtain GNSS monitoring data of the study area. S2. A dynamic early warning index is constructed by using a single or / and pairwise coupled feature index as the driving term and adding driving term weights to jointly drive the construction; the driving term weights are continuously and smoothly processed in adjacent time based on the assumption of gradual change, and the covariance matrix of uncertainty is predicted and the Kalman gain matrix is ​​calculated. S3. Using the normalized acceleration of GNSS monitoring data as the observation, construct the observation relationship between weights and acceleration response; based on the deviation between the observation and the predicted response calculated from the observation relationship, correct the driving term weights to obtain the dynamic fusion weights for the current time, and calculate the dynamic early warning index for the current time. S4. Construct an adaptive threshold model using historical sequences sorted in ascending order of dynamic early warning index. The adaptive threshold model uses the comprehensive quantile, mean, and standard deviation of the historical sequences to obtain early warning thresholds and divide the dynamic early warning index into level intervals, setting the slope risk level corresponding to each level interval; determine the level interval to which the dynamic early warning index belongs at the current time and output the slope risk level.

[0005] To better implement this invention, in method S2, the expression for the dynamic early warning index is as follows: ,in, The dynamic early warning index is defined as time t. This indicates the weight of the driving term. For the data corresponding to the driving item, , , , For a single driving factor driven by a feature indicator, As an indicator of meteorological and environmental risk, This refers to the engineering disturbance load index. The degree of rock mass damage in the burned area. , , These are the driving terms that are coupled in pairs to the feature indicators.

[0006] Preferably, in method S2, the weights of the driving terms are continuously and smoothly processed in adjacent time intervals based on the assumption of gradual variation, as expressed below: , Assuming a gradual change, the weights of the driving terms at time t, Assuming a slowly varying hypothesis, the weights of the driving terms at time t-1, For small perturbation terms; assume that the prior estimate of the weight matrix at time t is related to time. If the weight matrices are equal, then the expression for the covariance matrix of uncertainty is as follows: , Let be the prior covariance matrix at time t. For time Posterior covariance matrix, The process noise covariance matrix is ​​given; the Kalman gain matrix expression is calculated as follows: , The data corresponding to the time-driven term t. To observe the noise covariance, for transpose, To perform the matrix inversion operation, This represents the variance of the prediction error.

[0007] Preferably, in method S3, the observed relationship between weights and acceleration response is constructed using the following expression: , The data corresponding to the time-driven term t. To detect noise, Kalman filtering is performed using the Kalman gain matrix, and its weight update expression is as follows: , ,in The actual values ​​of the driving term weights are adjusted and updated. The predicted values ​​before adjusting the weights of the driving terms. The acceleration response at time t, Let be the gain matrix at time t.

[0008] Preferably, in method S3, the actual values ​​of the driving term weights are corrected and updated. After normalization and with the sum of constraint weights set to 1, the dynamic early warning index for the current time is calculated using the actual values ​​of the driving term weights after normalization as the dynamic fusion weights for the current time. .

[0009] Preferably, in method S1, the temperature influence index for time t is obtained according to the following temperature influence function. : ,in The temperature at time t; It is the freezing point temperature; The soil and rock temperature sensitivity coefficient is used; the humidity influence index at time t is obtained according to the following humidity influence function. : , The normalized air humidity over time t; The humidity sensitivity coefficient is used; the pressure influence index is obtained according to the following pressure influence function. : , The normalized pressure value for time t; The rate of change of atmospheric pressure over time t. The pressure sensitivity coefficient is used; the impact index of rainfall intensity disturbance is obtained according to the following formula. : , The maximum rainfall intensity at time t; The maximum rainfall intensity recorded in the study area is used as a reference value; the cumulative effective rainfall impact index is obtained according to the following expression. : , Before time t Step length in a short period of time The amount of rainfall; Taking into account the duration of preceding rainfall, The rainfall attenuation coefficient represents the index of influence on effective rainfall. The effective rainfall impact index was obtained by normalization. Construct a meteorological environmental risk index for the impact of slope deformation in the study area. : , , , , , These are the weighting parameters, and the study area is an open-pit mine fire zone.

[0010] Preferably, in method S1, the study area is an open-pit mine fire zone. A blasting pulse response function is constructed to quantify the influence of instantaneous and cumulative impacts on slope deformation, and the blasting pulse response at time t... The expression is as follows: , For the first Peak velocity of vibration measured by secondary slope vibration sensors For the first The cumulative time of each blast This represents the total number of explosions. The blasting disturbance decay rate is given; the cumulative effect function of the mining face advance is constructed as follows to obtain the cumulative effect of the mining face advance at time t. : , Let t be the working face advance speed. Before time t The cumulative distance of advancement over a total time period , These are the corresponding influence coefficients; the following slope toe excavation unloading effect function is constructed to obtain the slope toe excavation unloading effect at time t. : , The depth of excavation on day t; Before time t The total scale of excavation over a given period; , The corresponding influence coefficients; constructing engineering disturbance load indices for the influence of slope deformation in the study area. : , , , These are the weight parameters.

[0011] Preferably, the study area is an open-pit mine fire zone. A micro-damage model is constructed to investigate the influence of slope deformation in the study area. The micro-damage model obtains micro-damage indices at time t. The expression is as follows: , The initial elastic modulus, The acoustic velocity of the rock mass on the slope of the burned area. The density of the rocks in the burned area. Given the Poisson's ratio of the rocks in the burned area; construct the following macroscopic deformation function to obtain the macroscopic deformation damage index at time t. : ,in for Deformation rate monitored on the slope of the burned area at all times; The critical deformation rate was used to construct the rock mass damage degree in the burned area affected by slope deformation in the study area. : , , These are the weighting parameters; during the slope creep stage At that time, microscopic cracks predominated, and the setting was... During the accelerated deformation stage At that time, macroscopic deformation was dominant, and the setting was... , This setting is based on historical monitoring data.

[0012] Preferably, in method S4, historical dynamic early warning index sequence data is obtained and sorted in ascending order to obtain a historical sequence, and the mean of the historical sequence is calculated. with standard deviation Select quantiles as The adjustment coefficient is obtained by conversion calculation. The warning threshold is obtained according to the following formula. , Select quantiles as The adjustment coefficient is obtained by conversion calculation. The warning threshold is obtained according to the following formula. , ; Dynamic early warning index based on current time t Slope risk level assessment is conducted, and slope risk levels are categorized into low, medium, and high. When the slope risk level is low, the output is low risk; when At that time, the slope risk level was output as medium risk; when At that time, the slope risk level was output as high risk.

[0013] Preferably, the weight update adopts a sliding time window mechanism; the preset sliding time window is a data window corresponding to the most recent M consecutive monitoring times up to the current time t, where M is a preset positive integer; the dynamic weight update mechanism corrects and updates the weight of the driving item based on the driving item data and acceleration response data within the data window.

[0014] Compared with the prior art, the present invention has the following advantages and beneficial effects: (1) This invention utilizes three core disaster-causing factors—meteorological environment, engineering disturbance, and rock mass damage evolution—to construct meteorological environment risk indicators, engineering disturbance load indicators, and rock mass damage degree indicators, respectively, to comprehensively quantify the influence of multi-dimensional sub-factors such as temperature, humidity, blasting vibration, mining advancement, and micro- and macro-scale rock mass damage. Moreover, it uses the single driving force and pairwise coupling of characteristic indicators as driving terms and adds driving term weights to jointly drive the construction of a dynamic early warning index. Based on the three basic indicators, it innovatively constructs driving terms that are pairedly coupled between meteorology and engineering, meteorology and rock mass, and engineering and rock mass, to accurately depict the synergistic amplification effect of different disaster-causing factors. It can more realistically reflect the evolution process of slope risk under complex working conditions, and can more comprehensively characterize the formation mechanism of slope risk in fire-affected areas, significantly improving the adaptability of the early warning method to the complex geological environment of fire-affected areas.

[0015] (2) This invention adopts a dynamic adaptive weight update method based on Kalman filtering. The normalized acceleration calculated from the slope GNSS displacement monitoring data is used as the observation. The weights of the basic indicators and coupled driving terms are estimated and updated in real time. The smoothness of the update is ensured by following the assumption of gradual weight change, while the rationality of the weights is ensured by non-negativity constraints and normalization processing. By introducing a dynamic weight update mechanism based on slope deformation response, the contribution of each driving factor can be adaptively adjusted according to the slope state changes at different time stages. This overcomes the problem that the fixed weight method is difficult to adapt to time-varying conditions. The dynamic weight update mechanism realizes the adaptive adjustment of the real-time response state of the slope, overcomes the defect of the fixed weight method being difficult to adapt to time-varying conditions, and makes the contribution of each driving factor accurately match the actual evolution stage of the slope.

[0016] (3) This invention abandons the fixed threshold mode (the traditional early warning method adopts a fixed threshold mode with fixed upper and lower limits, which cannot adapt to the risk characteristics of different stages and working conditions of the fire-affected slope, and is prone to misjudgment, missed judgment or delayed early warning). Based on the statistical characteristics (mean, standard deviation, empirical quantile) of the dynamic early warning index (DWI) within the sliding time window, an adaptive threshold model that updates over time is constructed. By selecting different quantile values ​​to determine the high and medium risk thresholds respectively, the dynamic division of risk levels is realized. It can fully reflect the sudden change and stage characteristics of the fire-affected slope risk, effectively avoid the misjudgment problem of fixed threshold under complex working conditions, and realize the real-time output of risk levels, thereby improving the timeliness and pertinence of early warning.

[0017] (4) The sub-factors of each basic indicator (i.e., meteorological environmental risk index, engineering disturbance load index, and rock mass damage degree in the fire area) of this invention are determined through experimental inversion (such as temperature, humidity, and pressure sensitivity coefficients), historical data fitting (such as rainfall attenuation coefficient), and mining area case calibration (such as critical deformation rate), ensuring the accuracy and scientific nature of the indicator quantification; the sub-factor weights of the basic indicators adopt entropy weight method coupled with grey relational analysis, multiple linear regression and other methods, and the top-level weights adopt Kalman filtering for dynamic updating, taking into account both objectivity and dynamic adaptability; three levels are clearly defined: green safety, yellow low risk, and red high risk, corresponding to different prevention and control measures, directly providing practical guidance for on-site safety management in mines, and improving the engineering applicability of the method. This invention specifically solves the core pain points of existing technologies, such as the difficulty in uniformly representing multiple disaster-causing factors and their coupling effects, the difficulty in dynamically reflecting the changes in the contribution of each driving factor, and the lag in risk level judgment, filling the technical gap in the field of dynamic early warning of slopes in open-pit mine fire areas, providing reliable technical support for the safety management of slopes in fire areas, and has significant engineering application value and promotion prospects. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method for dynamic early warning of slope deformation in open-pit mine fire areas according to the present invention; Figure 2 This is a schematic diagram illustrating the principle of the dynamic early warning method for slope deformation in open-pit mine fire areas according to the present invention. Figure 3 The following is a statistical result curve of the influence of temperature, humidity, pressure, rainfall intensity disturbance, and effective rainfall on the study area of ​​an open-pit mine fire area in Xinjiang, as an example in the embodiment; Figure 4 The example provided is a statistical result curve of the meteorological environmental risk index of a fire-affected area in an open-pit mine in Xinjiang, which serves as the study area. Figure 5 A statistical chart of peak blasting vibration velocity in the example study area; Figure 6 The curves of the influence functions of mining in three types of engineering indices in the engineering disturbance load index of the study area are shown as examples. Figure 7 The example provided is a statistical result curve of the engineering disturbance load index of a fire area in an open-pit mine in Xinjiang, which is used as the study area. Figure 8 A comparative statistical curve of micro-deformation damage indices and macro-deformation damage indices for the example study area; Figure 9 The example provided is a statistical result curve of rock mass damage index in a burned area of ​​an open-pit mine in Xinjiang, which serves as the study area. Figure 10 The example shown is a slope deformation early warning level map of a fire-affected area in an open-pit mine in Xinjiang, which serves as the study area. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to embodiments: Example like Figure 1 , Figure 2 As shown, a dynamic early warning method for slope deformation in open-pit mine fire areas based on multi-factor coupling is proposed, the method comprising: S1. Obtain GNSS monitoring data for the study area; acquire meteorological data, engineering disturbance data, and rock mass response data for the study area, and calculate characteristic indicators including meteorological environmental risk index, engineering disturbance load index, and rock mass damage degree in the burned area. The meteorological data, engineering disturbance data, and rock mass response data of the study area together constitute a multi-source dataset. This invention constructs a characteristic indicator system, which includes characteristic indicators of meteorological environmental risk index, engineering disturbance load index, and rock mass damage degree in the burned area. The meteorological environmental risk index, engineering disturbance load index, and rock mass damage degree in the burned area respectively characterize the comprehensive intensity of the disaster-causing factors. The study area of ​​this invention is an open-pit mine burned area. Among the factors affecting slope stability in the study area (open-pit mine burned area), the main effect of temperature on slope damage is reflected in the freeze-thaw cycle effect. The internal moisture of soil and rock undergoes repeated phase changes during freezing and thawing. Due to the volume difference between water and ice, the pore structure undergoes periodic expansion and contraction, thereby accelerating structural damage and strength deterioration of the soil and rock. To quantitatively assess the impact of freeze-thaw cycles on slope deformation in the study area (an open-pit mine fire zone), this invention constructs a temperature influence function sensitive to temperature fluctuations near the freezing point. The temperature influence index at time t is obtained according to the following temperature influence function. : ,in The temperature at time t, Freezing point temperature The temperature sensitivity coefficient for soil and rock (used to characterize the enhanced effect of temperature changes near the freezing point on soil and rock structure damage; in this embodiment, the preferred value range is 3-8, which can be determined by regression analysis based on the freeze-thaw cycle test results of representative soil and rock samples). The study area (an open-pit mine fire zone) has sparse vegetation, and the mineral and pore structures of the soil and rock have changed to varying degrees due to long-term coal fire burning, resulting in a significant decrease in the erosion resistance and moisture retention capacity of the slope soil and rock in the fire zone. This makes the slope more sensitive to air humidity input. Therefore, this invention specifically constructs a humidity influence function to quantitatively evaluate the impact of humidity on slope deformation. The humidity influence index at time t is obtained according to the following humidity influence function. : , The normalized air humidity over time t. The humidity sensitivity coefficient (used to characterize the sensitivity of soil and rock materials on fire-affected slopes to changes in air humidity; in this example, the preferred value range is 1.5-2.5, which can be obtained through regression analysis based on the mechanical property test results of representative soil and rock samples under different humidity conditions). The temperature at time t, freezing point temperature, soil and rock temperature sensitivity coefficient, normalized air humidity at time t, and humidity sensitivity coefficient are all derived from multi-source datasets or experimental measurements.

[0020] The soil and rock slopes in the study area (an open-pit mine burning area) have undergone significant enhancement in pore structure and fissure development under long-term coal fire burning, making the gas exchange between the soil and rock mass and the external atmosphere more sensitive. When the external air pressure changes rapidly, the gas in the pores inside the soil and rock mass cannot escape in time, easily forming a transient internal and external air pressure difference, which adversely affects the stability of the slope. This invention constructs a pressure influence function (used to quantitatively evaluate the influence of atmospheric pressure on slope deformation), and obtains the pressure influence index according to the following pressure influence function. : , The normalized pressure value at time t. The rate of change of atmospheric pressure over time t (i.e., the change in atmospheric pressure per unit time). It represents the amount of air pressure change; a triaxial testing apparatus, including an air pressure control system and a high-precision monitoring system, can be used to simulate multi-rate air pressure change processes and monitor the instantaneous changes in pore pressure inside the soil or rock mass in real time. The pressure sensitivity coefficient (in this example, the preferred value range is 0.5-2.0; in this example, multiple sets of data points can be experimentally measured and substituted into the damage model of the air pressure influence function. The damage model has a decay equation, and the optimal value that minimizes the error between the model prediction curve and all experimental data points is found as the pressure sensitivity coefficient). The normalized air pressure value at time t, the normalized air pressure change rate at time t, and the pressure sensitivity coefficient are all derived from multi-source datasets or experimental measurements.

[0021] Rainfall intensity in the study area (preferably an open-pit mine fire area) directly reflects the rapid disturbance of slopes by surface runoff and water infiltration over a short period of time, and is one of the key risk drivers for inducing shallow slope instability. This invention constructs a rainfall intensity disturbance impact index function (to quantitatively assess the rapid disturbance impact of rainfall), and obtains the rainfall intensity disturbance impact index according to the following formula (i.e., the rainfall intensity disturbance impact index function). : , The maximum rainfall intensity at time t. This represents the reference value for the maximum rainfall intensity recorded in the study area. In the study area (preferably an open-pit mine fire area), after early rainfall infiltrates the rock mass, it will drain or evaporate over time through fissure channels. Its impact on pore water pressure and slope deformation at the current moment shows a decreasing trend over time, and the impact of recent rainfall on slope stability is more significant. Therefore, this invention constructs an effective rainfall influence index function with a decay coefficient, and the cumulative effective rainfall influence index is obtained according to the following expression. : , Before time t Step length in a short period of time The amount of rainfall; The preceding rainfall time to be considered (e.g., selecting the number of days or time period, i.e., considering rainfall data that affects deformation up to time t, including rainfall time and corresponding rainfall amount). The rainfall attenuation coefficient (preferred value in this example is 0.8; water dissipation on the slopes of the burned area is relatively fast, and the value of the rainfall attenuation coefficient in this example is determined by fitting the historical rainfall-displacement response relationship); in some embodiments, the effective rainfall impact index is... The effective rainfall impact index was obtained by normalization. The effective rainfall impact index obtained exist Within the range, the effective rainfall impact index The normalization expression is as follows: , This represents the maximum rainfall threshold for the study area over a historical time period (e.g., seven days). The maximum rainfall intensity at time t and the maximum rainfall threshold for the study area over a historical time period are both derived from multi-source datasets. The reference value for the maximum rainfall intensity recorded in the study area is obtained statistically from the multi-source datasets.

[0022] This invention obtains temperature influence indicators Humidity Influence Index Pressure Influence Indicators Rainfall intensity disturbance impact indicators Effective rainfall impact indicators Subsequently, this invention constructs a meteorological environmental risk index for the impact of slope deformation in the study area. : , , , , , These are the weighting parameters (i.e., the temperature influence indicators). Humidity Influence Index Pressure Influence Indicators Rainfall intensity disturbance impact indicators Effective rainfall impact indicators (Weight parameters). This embodiment... , , , , The five weighting parameters can be determined using a combination of entropy weighting and grey relational analysis (i.e., determining the five weighting parameters). The entropy weighting method calculates the dispersion of each meteorological impact indicator within the historical monitoring period to calculate the objective weights. (Right now , , , , (Five weighting parameters); simultaneously, using the slope displacement rate or settlement rate sequence as the reference sequence, and the normalized sequences of each meteorological factor (each meteorological influence index) as the comparison sequence, grey relational analysis is used to calculate the correlation between each factor and the slope deformation response. The correlation weights reflecting the sensitivity to disaster response are obtained; finally, the objective weights and correlation weights are coupled, fused, and normalized to construct a joint weight. We obtained the weight coefficients of each part of the meteorological environment risk index (i.e., the meteorological environment driving index).

[0023] This embodiment uses the slope of a burned area in an open-pit mine in Xinjiang as the research object or research area. Monitoring data (including GNSS monitoring data) and engineering data (including meteorological data, engineering disturbance data, and rock mass response data) from January 1, 2024 to June 30, 2024 were acquired using multi-source sensors deployed on the slope and mining operation logs. The normalized weight matrix was calculated using the entropy weight method-grey relational analysis method. In the case study of a fire-affected mining area slope in an open-pit mine in Xinjiang, the meteorological environmental risk indicators are as follows: The slope of the burning area in an open-pit mine in Xinjiang was used as a temperature influence indicator for the study area. Humidity Influence Index Pressure Influence Indicators Rainfall intensity disturbance impact indicators Effective rainfall impact indicators The result is as follows Figure 3 As shown, meteorological environmental risk index The result is as follows Figure 4 As shown.

[0024] The impact of blasting vibration on the slope of the study area (preferably an open-pit mine fire area) is mainly reflected in the instantaneous impact and the cumulative effect of dynamic damage caused by it. The high-frequency dynamic load generated by blasting will accelerate the propagation of rock mass fissures and weaken the slope stability in a short period of time. This invention constructs a blasting impulse response function to quantitatively evaluate the impact of instantaneous and cumulative impacts on slope deformation. The blasting impulse response at time t is... The expression is as follows: , , For the first The peak vibration velocity measured by the secondary slope vibration sensor, in a case study of a fire-prone area in an open-pit mine in Xinjiang, is statistically analyzed as follows: Figure 5 As shown; For the first The cumulative time of each blast This represents the total number of explosions. The decay rate of the blasting disturbance (preferred in this example, the value range is 0.5-0.8) is used to determine the blasting disturbance characterization value for the current time period (e.g., the current day) if multiple blasts occur within the same time period (e.g., the same day). This characterization value is formed by superimposing or accumulating the responses of all blasting events within the current time period (e.g., the current day). In the case study of a fire-affected slope in an open-pit mine in Xinjiang, the normalized blasting impulse response is... The value is 0.763. Due to thermal damage reducing the strength of the soil and rock, the slopes in the study area (preferably the open-pit mine fire area in this invention) are more sensitive to stress disturbances, and the advancement of the working face during mining operations can easily induce large-area settlement or deep slippage. To accurately quantify the impact of mining operations on slope deformation, this invention constructs a cumulative effect function for the advancement of the mining working face by combining the advancement rate and the cumulative advancement amount. The cumulative effect of the working face advancement at time t is obtained by constructing the following cumulative effect function for the advancement of the mining working face. : , Let t be the working face advance speed. Before time t Total time (e.g.) Take 30 days, The cumulative advance distance is the cumulative advance distance over the 30 days prior to time t. , These are the corresponding influence coefficients and the cumulative effect of working face advancement. The value is 0.648. In a case study of a fire-affected slope in an open-pit mine in Xinjiang, the normalized value, in some embodiments, utilizes time-series data of the actual deep slope displacement rate acquired synchronously. Construct the following residual sum of squares equation: And then , By calculating the first-order partial derivatives and setting them to zero, we can obtain the optimal estimates of the two influence coefficients on the slope displacement (i.e., the calculated influence coefficients). , Continuous excavation activities remove the soil and rock mass at the toe of the slope, releasing the constraining stress of the overlying soil and rock mass. This leads to the slope in the fire-affected area being prone to collapse. To accurately quantify the impact of this engineering excavation on the deformation of the fire-affected slope, this invention constructs the following toe excavation unloading effect function to obtain the toe excavation unloading effect at time t. : , The depth of excavation on day t is the time t. Before time t The total scale of excavation over a long period of time , The corresponding influence coefficients, for , Solve and , The basic solution approach is the same: using the least squares method, we inversely calculate the sum of squared residuals between the actual monitoring data and the theoretical model by minimizing the sum of squared residuals. , Two influence coefficients. During actual open-pit mining operations, the slope of the study area (preferably the open-pit fire zone in this invention) will be affected by the construction process, mainly including blasting vibration effects, engineering mining propulsion effects, and slope toe excavation unloading effects. This invention obtains the blasting pulse response... Cumulative effect of working face advancement Slope toe excavation unloading effect Subsequently, this invention constructs an engineering disturbance load index for the influence of slope deformation in the study area. : , , , These are three types of engineering activities (i.e., blasting pulse response). Cumulative effect of working face advancement Slope toe excavation unloading effect The weight parameters corresponding to the three types of engineering activities. In some embodiments, for obtaining the three types of engineering activities, this embodiment can use the slope displacement rate or settlement rate as the dependent variable, and the normalized blasting disturbance characterization value, mining advancement characterization value, and slope toe excavation unloading characterization value as independent variables to establish a multiple linear regression model, and use the standardized regression coefficients to characterize the relative contribution intensity of different engineering activities to the slope deformation response; after normalizing each standardized regression coefficient, the engineering disturbance load index is obtained. The weight parameters of each engineering disturbance sub-factor are determined. Using the slope of a fire-affected area in an open-pit mine in Xinjiang as a case study, a normalized weight matrix is ​​obtained. Engineering disturbance load index of the study area (i.e., the open-pit mine fire area) Its explosive impulse response is 0.767. (Also known as the blasting vibration effect component EB), cumulative effect of working face advancement (Also known as the propulsion speed influence component EF), slope toe excavation unloading effect (Also known as the unloading disturbance influence component EE) are as follows: Figure 6 As shown; Engineering disturbance load index The statistical results are as follows Figure 7 As shown.

[0025] The rock mass in the study area (preferably an open-pit mine fire area) has been subjected to long-term coal fire burning, resulting in changes to its internal microstructure. To quantify the impact of these changes in rock mass microstructure on slope stability, this invention constructs a micro-damage model of slope deformation in the study area. The micro-damage model obtains micro-damage indices at time t. The expression is as follows: , The initial elastic modulus, The acoustic velocity of the rock mass on the slope of the burned area. The density of the rocks in the burned area. The Poisson's ratio of the rock in the burned area is used. The aforementioned micro-damage function of the micro-damage model is used to characterize the degree of material property degradation caused by thermal damage, crack propagation, and structural loosening of the rock mass in the burned area. In the study area (preferably an open-pit mine burned area), the macroscopic damage of the slope rock mass manifests as surface deformation and displacement. The displacement rate is the most direct and critical macroscopic indicator reflecting slope instability. This invention constructs the following macroscopic deformation function to obtain the macroscopic deformation damage index at time t. (By linking critical instability states, the extent of slope collapse stages can be quantified): ,in for Deformation rate monitored on the slope of the burned area at all times. The critical deformation rate is defined in this embodiment based on statistical results of historical instability cases in the mining area, long-term monitoring data, and on-site engineering criteria. Considering that the slope damage evolution in the study area (preferably an open-pit mine fire zone) exhibits both internal material structure degradation and external apparent deformation enhancement, this invention comprehensively evaluates the impact of rock mass damage on slope stability in the fire zone from both microscopic and macroscopic perspectives. Therefore, this invention constructs a rock mass damage degree for the fire zone, reflecting the influence of slope deformation on the study area. : , As an indicator of microscopic damage, As a macroscopic deformation damage index, , These are weighting parameters (also known as proportion coefficients, which dynamically switch according to the risk range of the monitored displacement rate). During the slope creep stage... At that time, microscopic cracks predominated, and the setting was... During the accelerated deformation stage At that time, macroscopic deformation was dominant, and the setting was... , This is set based on historical monitoring data (determined by historical monitoring data and / or instability cases). The critical deformation rate (or critical instability velocity of the slope). In a case study of a fire-affected slope in an open-pit mine in Xinjiang, the micro-damage index... The value is 0.145, representing the macroscopic deformation damage index. The value is 0.454, therefore the damage degree of the rock mass in the burned area at time t is calculated. The value is 0.262. The statistical curve comparing the microscopic deformation damage index and the macroscopic deformation damage index is shown below. Figure 8 As shown in the figure, the statistical results curve of the rock mass damage index in the study area is as follows: Figure 9 As shown. The data involved in step S1 of this embodiment are all from multi-source datasets (including meteorological data, engineering disturbance data, and rock mass response data) or obtained through statistical calculations of multi-source datasets.

[0026] S2. After obtaining the meteorological environmental risk index MERI, the engineering disturbance load index EDI, and the rock mass damage degree index RD, as follows: Figure 1 , Figure 2As shown, to further characterize the synergistic enhancement effect among different types of disaster-causing factors and comprehensively reflect the changing characteristics of slope risk status over time, a dynamic early warning index (DWI) is constructed using single-driven and / or pairwise coupled characteristic indicators as driving terms, with added driving term weights. The DWI consists of three basic driving indicators and their pairwise coupled driving terms. The three basic driving terms characterize the independent impacts of meteorological environment, engineering disturbance, and rock mass damage evolution on slope stability. The three coupled driving terms characterize the synergistic amplification effect of different types of disaster-causing factors acting simultaneously. Considering that slope risk in fire-affected areas is not linearly determined by a single factor and that the dominant mechanism changes significantly at different stages, the DWI is defined as follows: The formula can be expanded as follows: The first three terms in the formula are independent action terms of the three types of driving factors, used to characterize the direct impact of meteorology, engineering mining and rock mass damage on slope stability; the last three terms are coupling terms of the three types of factors, used to describe the mutual enhancement and amplification effects of the three types of driving factors at the same time. , , For the constructed The meteorological environmental risk index, engineering disturbance load index, and rock mass damage index of the fire area after time-normalization. , , , , , These represent the weight of each part; Characterizing the meteorological-engineering coupling effect, after rainfall infiltration, the pore water pressure of the burnt rocks in the fire zone increases and the effective stress decreases. If blasting vibration is superimposed at this time, the crack propagation rate will increase exponentially. Characterizing the meteorological-rock mass effect, as the cumulative damage to the rock mass increases, the penetration of fissures in the sintered rock increases, leading to deeper infiltration paths for atmospheric precipitation, which further accelerates the deterioration of the rock mass and forms a positive feedback loop. Characterizing the engineering-rock mass effect, continuous mining on the basis of existing rock mass damage significantly enhances the sensitivity to stress redistribution, accelerating slope settlement. Preferably, the dynamic early warning index expression is as follows: ,in, The dynamic early warning index is defined as time t. This indicates the weight of the driving term. ; For the data corresponding to the driving item, , , , The driving term is a single feature indicator driven by a single feature indicator (also known as a basic driving term; preferably, the driving term driven by a single feature indicator needs to be normalized). As an indicator of meteorological and environmental risk, This refers to the engineering disturbance load index. The degree of rock mass damage in the burned area. , , These are the driving terms that are paired and coupled to each other (also known as coupled driving terms, which need to be normalized).

[0027] The covariance matrix of the uncertainty is predicted by continuously and smoothly processing the driving term weights based on the slowly varying assumption between adjacent time intervals, and the Kalman gain matrix is ​​calculated. Preferably, the expression for continuously and smoothly processing the driving term weights based on the slowly varying assumption between adjacent time intervals is as follows: , Assuming a gradual change, the weights of the driving terms at time t, Assuming a slowly varying hypothesis, the weights of the driving terms at time t-1, This represents a small perturbation term. Assume the prior estimate of the weight matrix at time t is related to the time... If the weight matrices are equal, then the expression for the covariance matrix of uncertainty is as follows: , Let be the prior covariance matrix at time t. For time Posterior covariance matrix, Let be the process noise covariance matrix. The Kalman gain matrix is ​​calculated and expressed as follows: , The data corresponding to the time-driven term t. To observe the noise covariance, for transpose, To perform the matrix inversion operation, This represents the variance of the prediction error.

[0028] In some embodiments, to avoid distortion of the dynamic early warning index at different stages due to the use of fixed weights, a dynamic adaptive weight update method based on Kalman filtering is adopted. This method estimates and updates the six top-level weights corresponding to MERI, EDI, RD, and the three types of coupled driving terms in real time. The dynamic weight update method uses a six-dimensional weight vector as the state variable and the normalized acceleration calculated from slope GNSS displacement monitoring data as the observation. It adaptively corrects the contribution of each driving term in the top-level early warning model using real-time slope response information, thereby obtaining a dynamic weight sequence that changes over time. First, the dynamic early warning index of slope deformation in the fire-affected area is written as: ,in, , Set the initial weight matrix. and the initial covariance matrix ,in This is the identity matrix. Considering that slope structural conditions, rock mass thermal damage, and engineering methods change relatively smoothly over short timescales, the weight evolution follows the "gradual change assumption," meaning that the weights exhibit continuous and smooth changes between adjacent time points, conforming to the formula: ,in To reflect the small perturbation term of dynamic adjustment, based on the assumption of gradual change in weights, Prior estimate weight matrix at time step and The weight matrices at time points are equal, i.e. Simultaneously update the covariance matrix representing prediction uncertainty: ;in The process noise covariance matrix is ​​typically as follows: Then use the predicted weights to calculate Moment The formula is: Then calculate the Kalman gain matrix. The formula is: ;in, To observe the noise covariance and quantitatively characterize the random error level and data reliability of the real-time slope monitoring system during data acquisition, the calculation benchmark is determined by the squared ratio of instrument physical error to instability threshold. Then, adjustments are made step-by-step based on the smoothness and sensitivity of the inversion curve. Numerical values ​​ensure that the system can effectively filter out noise and quickly capture disaster signals.

[0029] S3. Using the normalized acceleration from GNSS monitoring data as the observation, construct the observation relationship between weights and acceleration response; based on the deviation between the observation and the predicted response calculated from the observation relationship, correct the driving term weights to obtain the dynamic fusion weights for the current time, and calculate the dynamic early warning index for the current time. Preferably, the expression for constructing the observation relationship between weights and acceleration response is as follows: , The data corresponding to the time-driven term t. To detect noise, Kalman filtering is performed using the Kalman gain matrix, and its weight update expression is as follows: , ,in The actual values ​​of the driving term weights are adjusted and updated. The predicted values ​​before adjusting the weights of the driving terms. The acceleration response at time t (preferably the normalized acceleration response at time t). Let be the gain matrix at time t; The normalized acceleration response is obtained based on the predicted weights and driving term data. The predicted response quantities are at the same normalized scale. Preferably, the weight update adopts a sliding time window mechanism; the preset sliding time window is a data window corresponding to the most recent M consecutive monitoring times up to the current time t, where M is a preset positive integer; the dynamic weight update mechanism corrects and updates the driving term weights based on the driving term data and acceleration response data within the data window. This embodiment can also adopt the following technical solution: the observation relationship between weights and acceleration response is constructed using a preset sliding time window for dynamic weight update mechanism, the preset sliding time window covers the time length of the characteristic index, and the dynamic weight update mechanism obtains the data from the preset sliding time window to correct and update the driving term weights. Preferably, the actual value after correcting and updating the driving term weights... After normalization and with the sum of constraint weights set to 1, the dynamic early warning index for the current time is calculated using the actual values ​​of the driving term weights after normalization as the dynamic fusion weights for the current time. , , These are the actual values ​​of the driving term weights after normalization constraint processing. This refers to the data corresponding to the time-driven term t.

[0030] In some embodiments, the normalized slope acceleration is calculated using GNSS monitoring data. As an observation, the observational relationship between the weighted state and the actual slope response is established: ;in To mitigate observation noise and characterize monitoring system errors and data uncertainties, the weights are constrained by introducing the actual slope response. Finally, the deviation between the actual and predicted values ​​is used to synchronously correct the six weights (using the normalized slope acceleration as the observed value, the six top-level weights are dynamically corrected). The updated weight vector is subjected to non-negativity constraints and normalization to ensure that each weight value is within [0,1] and the sum is 1. This yields the dynamic fusion weights of the six driving terms at the current time, which are then used to calculate the dynamic early warning index (DWI) at time t. Through this Kalman dynamic weight update process, the early warning model can automatically adjust the contribution of each driving factor and its coupling terms according to the real-time response state of the slope, thereby improving the adaptability of the early warning results to the staged evolution characteristics of the fire-affected slope.

[0031] S4. Construct an adaptive threshold model using historical sequences sorted in ascending order of dynamic early warning index. The adaptive threshold model uses the combined quantiles, mean, and standard deviation of the historical sequences to obtain early warning thresholds, and divides the dynamic early warning index into level intervals, setting the slope risk level corresponding to each level interval. It then determines the level interval to which the dynamic early warning index belongs at the current time and outputs the slope risk level. The method for obtaining the early warning threshold using the combined quantiles, mean, and standard deviation is as follows: Calculate the mean of the historical sequence... with standard deviation Set Q quantiles, and calculate the adjustment coefficient by converting the quantiles. Let q be the quantile index (the total number of indices is Q). The warning threshold for the quantile index q is obtained according to the following formula. A total of Q warning thresholds were obtained. These thresholds were then sorted by magnitude to sequentially divide the dynamic warning index into level intervals. Taking slope risk levels (low, medium, and high) as an example, the mean of the historical sequence was calculated. with standard deviation Select quantiles as The adjustment coefficient is obtained by conversion calculation. , , To prevent extremely small positive numbers with a denominator of zero, the warning threshold is obtained using the following formula. , Select quantiles as The adjustment coefficient is obtained by conversion calculation. , , To prevent extremely small positive numbers with a denominator of zero, the warning threshold is obtained using the following formula. , ; Dynamic early warning index based on current time t Slope risk level assessment is conducted, and slope risk levels are categorized into low, medium, and high. When the slope risk level is low, the output is low risk; when At that time, the slope risk level was output as medium risk; when At that time, the slope risk level was output as high risk.

[0032] In some embodiments, historical dynamic early warning index sequence data is acquired and sorted in ascending order to obtain a historical sequence. In the process of judging the slope risk level in the study area (preferably an open-pit mine fire area), the traditional fixed upper and lower threshold method is no longer used. Instead, the risk threshold is regarded as a dynamic statistic that changes with the evolution of the slope state. An adaptive threshold model is constructed based on the statistical characteristics of the recent dynamic early warning index to achieve real-time output of slope risk level and risk state. This method can fully reflect the abrupt and phased evolution characteristics of the slope under the influence of meteorological activities, engineering activities, and rock mass damage in the fire area, effectively avoiding the misjudgment problem caused by the fixed threshold method under complex working conditions. At any given time... Select the most recent Dynamic early warning index sequence at each time point: When the monitoring frequency, slope response period, or sample length changes, ( (The example is for 30 days) This can also be adjusted based on the stability and sensitivity analysis results of historical monitoring sequences. Based on this indicator sequence, calculate the current stage... mean of the sequence with standard deviation , , Secondly, the Dynamic Early Warning Index (DWI) sequences are sorted in ascending order of numerical value to obtain the empirical distribution sequence: The 90th percentile is used as a reference for high-risk identification, and an early warning threshold is obtained. The 65th percentile is used as a reference for identifying medium risk and a warning threshold is obtained. ; Construct early warning thresholds that change over time and , , ;in, , This is the adjustment coefficient for the current sliding window, and its calculated value range is typically [range missing]. , The dynamic early warning index at the current time t. Slope risk level assessment is conducted, and slope risk levels are categorized into low, medium, and high. At that time, the slope risk level output was low risk (also known as green safety status), indicating that the overall stability of the slope in the study area (i.e., the fire zone) was good, the combined effect of multiple disaster-causing factors had not yet caused a significant instability response, the slope was in a relatively stable and controllable risk stage, and routine monitoring and daily inspections could be maintained to continuously track changes in relevant monitoring indicators. At that time, the slope risk level output was medium risk (also known as yellow low risk status), indicating that the slope in the burned area had been affected by one or more adverse factors, such as rainfall infiltration, engineering disturbance, cumulative thermal damage to the rock mass, and enhanced fissure permeability. The slope stability was beginning to show an unfavorable evolution trend, with potential instability risk, but it had not yet developed into a high-risk stage. Monitoring frequency and on-site inspection intensity should be increased, focusing on changes in slope response, and timely implementation of targeted prevention and control measures such as drainage, unloading, local reinforcement, or operational adjustments to curb further risk development. When the slope risk level is output as high risk (also known as red high risk status), it indicates that the coupling effect of multiple disaster-causing factors has significantly increased, the slope stability has obviously decreased, and there is a risk of serious collapse, sliding, or overall instability. At this time, the high-risk early warning response mechanism should be activated immediately, and measures such as stopping work and evacuating personnel, sealing off dangerous areas, emergency response, and engineering reinforcement should be taken to reduce the safety risks to personnel and equipment caused by slope instability. In a case study of a fire-affected mining area slope in an open-pit mine in Xinjiang, the early warning threshold was obtained. , They are 0.312 and 0.393 respectively, when The date t is May 23, 2024, and the risk level is medium (also known as yellow low-risk status). Therefore, the slopes in the study area (i.e., the burned area) are determined to be in a state of potential low instability risk, requiring strengthened monitoring and targeted prevention and control measures. The slope deformation warning level for the study area is as follows: Figure 10 As shown.

[0033] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A dynamic early warning method for slope deformation in open-pit mine fire-affected areas based on multi-factor coupling, characterized in that: The methods include: S1. Obtain meteorological data, engineering disturbance data, and rock mass response data of the study area, and obtain characteristic indicators including meteorological environmental risk index, engineering disturbance load index, and rock mass damage degree in the burned area; obtain GNSS monitoring data of the study area. S2. A dynamic early warning index is constructed by using a single or / and pairwise coupled feature index as the driving term and adding driving term weights to jointly drive the construction; the driving term weights are continuously and smoothly processed in adjacent time based on the assumption of gradual change, and the covariance matrix of uncertainty is predicted and the Kalman gain matrix is ​​calculated. S3. Using the normalized acceleration of GNSS monitoring data as the observation, construct the observation relationship between weights and acceleration response; based on the deviation between the observation and the predicted response calculated from the observation relationship, correct the driving term weights to obtain the dynamic fusion weights for the current time, and calculate the dynamic early warning index for the current time. S4. Construct an adaptive threshold model using historical sequences sorted in ascending order of dynamic early warning index. The adaptive threshold model uses the comprehensive quantile, mean, and standard deviation of the historical sequences to obtain early warning thresholds and divide the dynamic early warning index into level intervals, setting the slope risk level corresponding to each level interval; determine the level interval to which the dynamic early warning index belongs at the current time and output the slope risk level.

2. The method for dynamic early warning of slope deformation in open-pit mine fire-affected areas based on multi-factor coupling driving as described in claim 1, characterized in that: In method S2, the expression for the dynamic early warning index is as follows: ,in, The dynamic early warning index is defined as time t. This indicates the weight of the driving term. For the data corresponding to the driving item, , , , For a single driving factor driven by a feature indicator, As an indicator of meteorological and environmental risk, This refers to the engineering disturbance load index. The degree of rock mass damage in the burned area. , , These are the driving terms that are coupled in pairs to the feature indicators.

3. The method for dynamic early warning of slope deformation in open-pit mine fire-affected areas based on multi-factor coupling driving as described in claim 1, characterized in that: In method S2, the weights of the driving terms are continuously and smoothly processed in adjacent time intervals based on the assumption of gradual variation, as shown in the following expression: , Assuming a gradual change, the weights of the driving terms at time t, Assuming a slowly varying hypothesis, the weights of the driving terms at time t-1, For small perturbation terms; assume that the prior estimate of the weight matrix at time t is related to time. If the weight matrices are equal, then the expression for the covariance matrix of uncertainty is as follows: , Let be the prior covariance matrix at time t. For time Posterior covariance matrix, The process noise covariance matrix; The Kalman gain matrix expression is calculated as follows: , The data corresponding to the time-driven term t. To observe the noise covariance, for transpose, To perform the matrix inversion operation, This represents the variance of the prediction error.

4. The method for dynamic early warning of slope deformation in open-pit mine fire-affected areas based on multi-factor coupling driving as described in claim 1 or 3, characterized in that: In method S3, the observed relationship between weights and acceleration response is constructed using the following expression: , The data corresponding to the time-driven term t. To detect noise, Kalman filtering is performed using the Kalman gain matrix, and its weight update expression is as follows: , ,in The actual values ​​of the driving term weights are adjusted and updated. The predicted values ​​before adjusting the weights of the driving terms. The acceleration response at time t, Let be the gain matrix at time t.

5. The method for dynamic early warning of slope deformation in open-pit mine fire-affected areas based on multi-factor coupling driving as described in claim 4, characterized in that: In method S3, the actual values ​​of the driving term weights are corrected and updated. After normalization and with the sum of constraint weights set to 1, the dynamic early warning index for the current time is calculated using the actual values ​​of the driving term weights after normalization as the dynamic fusion weights for the current time. .

6. The method for dynamic early warning of slope deformation in open-pit mine burning areas based on multi-factor coupling driving as described in claim 1, characterized in that: In method S1, the temperature influence index for time t is obtained according to the following temperature influence function. : ,in The temperature at time t; It is the freezing point temperature; The soil and rock temperature sensitivity coefficient is used; the humidity influence index at time t is obtained according to the following humidity influence function. : , The normalized air humidity over time t; The humidity sensitivity coefficient is used; the pressure influence index is obtained according to the following pressure influence function. : , The normalized pressure value for time t; The rate of change of atmospheric pressure over time t. The pressure sensitivity coefficient is used; the impact index of rainfall intensity disturbance is obtained according to the following formula. : , The maximum rainfall intensity at time t; The maximum rainfall intensity recorded in the study area is used as a reference value; the cumulative effective rainfall impact index is obtained according to the following expression. : , Before time t Step length in a short period of time The amount of rainfall; Taking into account the duration of preceding rainfall, The rainfall attenuation coefficient represents the index of influence on effective rainfall. The effective rainfall impact index was obtained by normalization. ; Constructing a meteorological environmental risk index for the impact of slope deformation in the study area : , , , , , These are the weighting parameters, and the study area is an open-pit mine fire zone.

7. The method for dynamic early warning of slope deformation in open-pit mine fire-affected areas based on multi-factor coupling driving as described in claim 1, characterized in that: In method S1, the study area is an open-pit mine fire zone. A blasting pulse response function is constructed to quantify the influence of instantaneous and cumulative impacts on slope deformation. The blasting pulse response at time t... The expression is as follows: , For the first Peak velocity of vibration measured by secondary slope vibration sensors For the first The cumulative time of each blast This represents the total number of explosions. The blasting disturbance decay rate is given; the cumulative effect function of the mining face advance is constructed as follows to obtain the cumulative effect of the mining face advance at time t. : , Let t be the working face advance speed. Before time t The cumulative distance of advancement over a total time period , These are the corresponding influence coefficients; the following slope toe excavation unloading effect function is constructed to obtain the slope toe excavation unloading effect at time t. : , The depth of excavation on day t; Before time t The total scale of excavation over a given period; , The corresponding influence coefficients; constructing engineering disturbance load indices for the influence of slope deformation in the study area. : , , , These are the weight parameters.

8. The method for dynamic early warning of slope deformation in open-pit mine burning areas based on multi-factor coupling driving as described in claim 1, characterized in that: The study area is an open-pit mine fire zone. A micro-damage model of the slope deformation effect in the study area was constructed, and the micro-damage model obtained the micro-damage index at time t. The expression is as follows: , The initial elastic modulus, The acoustic velocity of the rock mass on the slope of the burned area. The density of the rocks in the burned area. Given the Poisson's ratio of the rocks in the burned area; construct the following macroscopic deformation function to obtain the macroscopic deformation damage index at time t. : ,in for Deformation rate monitored on the slope of the burned area at all times; The critical deformation rate was used to construct the rock mass damage degree in the burned area affected by slope deformation in the study area. : , , These are the weighting parameters; during the slope creep stage At that time, microscopic cracks predominated, and the setting was... During the accelerated deformation stage At that time, macroscopic deformation was dominant, and the setting was... , This setting is based on historical monitoring data.

9. The method for dynamic early warning of slope deformation in open-pit mine fire-affected areas based on multi-factor coupling driving as described in claim 1, characterized in that: In method S4, historical dynamic early warning index sequence data is obtained and sorted in ascending order to obtain the historical sequence, and the mean of the historical sequence is calculated. with standard deviation Select quantiles as The adjustment coefficient is obtained by conversion calculation. The warning threshold is obtained according to the following formula. , ; Select quantiles as The adjustment coefficient is obtained by conversion calculation. The warning threshold is obtained according to the following formula. , ; Dynamic early warning index based on current time t Slope risk level assessment is conducted, and slope risk levels are categorized into low, medium, and high. When the slope risk level is low, the output is low risk; when At that time, the slope risk level was output as medium risk; when At that time, the slope risk level was output as high risk.

10. The method for dynamic early warning of slope deformation in open-pit mine fire-prone areas based on multi-factor coupling driving as described in claim 4, characterized in that: The weight update adopts a sliding time window mechanism; the preset sliding time window is the data window corresponding to the most recent M consecutive monitoring times up to the current time t, where M is a preset positive integer; the dynamic weight update mechanism corrects and updates the weight of the driving item based on the driving item data and acceleration response data in the data window.

Images