Microseismic-seepage coupled real-time inversion method for fracture evolution of mine water sealing layer

Abstract

The invention discloses a microseismic-seepage coupled real-time inversion method for fracture evolution of a mine water sealing layer. Microseismic and seepage data are collected in real time through the three-dimensional microseismic monitoring sensor array and the seepage monitoring device; establishing a three-dimensional seepage numerical model, and constructing a microseismic-seepage bidirectional coupling mechanism; a Bayesian inversion algorithm is adopted to invert fracture parameters in real time, and the model is corrected; early warning is triggered according to the permeability increment, the hydraulic opening sudden increase rate and the like, and closed-loop feedback is formed; and generating a real-time distribution map, an evolution curve and an early warning report guidance project. The three-dimensional micro-seismic monitoring sensor array comprises a DAS system and a short-period geophone, and a sliding time window spectrum analysis method is used for waveform inversion. The Bayesian inversion algorithm is combined with a Markov chain Monte Carlo method. According to the method, the independent simulation defect is overcome, the model prediction error is reduced by 40%, and real-time accurate inversion of fracture evolution and early warning of the instability risk are achieved.

Description

Technical Field

[0001] This invention relates to the field of real-time inversion of fracture evolution in mine water storage layers via microseismic-seepage coupling, specifically a method for real-time inversion of fracture evolution in mine water storage layers via microseismic-seepage coupling. Background Technology

[0002] Mine water storage technology is a key means to ensure mine safety and the rational utilization of water resources. Its core lies in real-time monitoring of the evolution of fractures in the storage layer to avoid leakage or instability risks. Existing methods operate microseismic monitoring or seepage monitoring independently, lacking coupled analysis of microseismic signals and seepage parameters, making it difficult to reveal the dynamic interaction between fracture propagation and fluid transport; it also lacks a coupled early warning mechanism based on multi-dimensional indicators such as fracture permeability increment and hydraulic aperture surge rate, making it difficult to accurately locate high-risk areas in the early stages of fracture instability. Summary of the Invention

[0003] The purpose of this invention is to provide a real-time inversion method for the evolution of fractures in mine water storage layers using microseismic-seepage coupling, so as to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling, comprising the following steps:

[0005] S1. Real-time acquisition and processing of multi-source data:

[0006] The three-dimensional microseismic monitoring sensor array is used to collect real-time data on the waveforms, source locations, energy, frequency characteristics, and fracture surface orientation of microseismic events in the mine water storage layer, and the fracture propagation direction parameters are extracted based on waveform inversion.

[0007] The pressure, flow rate, flow velocity, and fluid composition data of the water seal layer are collected synchronously using a seepage monitoring device.

[0008] S2. Construction and initialization of the dynamic coupling model:

[0009] A three-dimensional seepage numerical simulation model was established, and the initialization parameters included reservoir porosity, permeability tensor, fluid injection pressure boundary conditions, and fracture spatial distribution parameters extracted by S1.

[0010] A two-way coupling mechanism between microseismic and seepage was constructed. The fracture propagation direction parameters of the microseismic module were input into the seepage model to dynamically correct the permeability field. At the same time, the fluid pressure gradient calculated by the seepage model was fed back to the microseismic module as the driving force for fracture propagation.

[0011] S3, Real-time inversion and parameter update;

[0012] Using the Bayesian inversion algorithm, with the spatiotemporal evolution data and seepage data of microseismic events in S1 as constraints, the dynamic evolution parameters of the fractures are inverted in real time, including the permeability increment of the fracture network ΔK, the deflection angle of the maximum principal stress direction θ, the equivalent hydraulic aperture b, and the fracture geometry.

[0013] The permeability field of the three-dimensional seepage model and the rupture surface orientation parameters of the microseismic module are dynamically corrected based on the inversion results.

[0014] S4. Risk Warning and Closed-Loop Feedback:

[0015] When the permeability increment ΔK / K0 of any unit is monitored to be ≥50% or the hydraulic aperture surge rate db / dt is ≥10 -6 When m / s, the fracture instability early warning module is triggered and the set of coordinates of high-risk areas is output;

[0016] The corrected fracture parameters and early warning results are fed back to the sensor network and model initialization module of S1, forming a closed-loop system of real-time data acquisition, model update and risk warning.

[0017] S5. Results Output and Engineering Guidance:

[0018] The system generates real-time distribution maps of fracture permeability in the mine water storage layer, dynamic evolution curves of seepage pressure gradients, and early warning threshold reports to guide real-time adjustments to water injection and storage parameters.

[0019] Furthermore, in S1, the three-dimensional microseismic monitoring sensor array includes a distributed optical fiber acoustic wave sensing (DAS) system and a short-period seismic detector. Waveform inversion uses a sliding time window spectral analysis method to extract fracture propagation direction parameters, with a time window length of 0.1-1 seconds.

[0020] Furthermore, in S2, the seepage numerical simulation model is discretized using the finite element method, and the permeability tensor is calculated using a superposition model of experimentally determined matrix permeability and fracture permeability, with the following expression:

[0021]

[0022] Among them, K matrix K represents the matrix permeability. fracture Let δ be the fracture permeability, θ be the angle between the fracture orientation and the coordinate axis, and δ be the fracture permeability. ij For Kroneck's symbol.

[0023] Furthermore, in S3, the Bayesian inversion algorithm employs the Markov chain Monte Carlo method. Its prior probability distribution is based on historical microseismic event statistics, and its posterior probability is calculated using the joint likelihood function of microseismic energy and seepage pressure gradient, expressed as follows:

[0024] P(θ∣data)∝P(data∣θ)·P(θ)

[0025] Where θ represents the fracture parameters to be inverted, and data represents the real-time acquired microseismic and seepage data.

[0026] Furthermore, in S4, the determination criteria for the warning threshold include:

[0027] The seepage pressure gradient exceeds a preset threshold.

[0028] The relative rate of change of the permeability increment ΔK in the fractured network exceeds 50% (ΔK / K0 ≥ 0.5).

[0029] The equivalent hydraulic opening suddenly increases beyond the critical value (db / dt≥10-6m / s).

[0030] Furthermore, in S5, the real-time distribution map of fracture permeability adopts the form of a combination of contour map and three-dimensional cloud map, the dynamic evolution curve of seepage pressure gradient includes time series data and spatial distribution heat map, and the early warning threshold report includes the three-dimensional coordinates (x, y, z) of high-risk areas and the predicted failure probability value.

[0031] Furthermore, in S3, the dynamic correction includes:

[0032] The permeability field of the seepage model is updated every 10 minutes;

[0033] The rupture surface orientation parameters are updated immediately after each microseismic event is triggered.

[0034] The corrected parameters are smoothed using a Kalman filter algorithm and historical data.

[0035] Furthermore, in S4, the data synchronization frequency of the closed-loop feedback system is 1Hz, and the sensor network and model initialization module transmit data in real time through industrial Ethernet, with a feedback delay of no more than 0.5 seconds.

[0036] Furthermore, in S2, the microseismic-seepage bidirectional coupling mechanism is achieved in the following way:

[0037] Input the fracture propagation direction parameters extracted from S1 into the seepage model to dynamically adjust the principal direction of the fracture permeability tensor.

[0038] The fluid pressure gradient calculated by the seepage model is used as the effective stress, and the constitutive equation of the microseismic module is corrected by Biot consolidation theory.

[0039] Furthermore, in step S5, the adjustment of the water injection and sealing parameters includes:

[0040] When a high-risk area is detected, the water injection pressure is automatically reduced by 10%-30%.

[0041] Based on the fracture permeability distribution map, adjust the well density and injection rate of the injection wells to ensure that the injection area avoids areas with a failure probability > 80%.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] This invention utilizes multi-source data coupling, dynamic model construction, and real-time inversion techniques. It employs a three-dimensional microseismic monitoring array composed of a distributed fiber optic acoustic sensing system and a short-period seismic detector, combined with a sliding time window spectral analysis method, to achieve high-frequency and accurate extraction of fracture propagation direction parameters.

[0044] The seepage pressure gradient is fed back to the microseismic module as the driving force for fracture propagation. At the same time, the permeability tensor of the seepage model is dynamically adjusted based on the fracture propagation direction, which solves the defect of independent simulation of microseismic and seepage processes in traditional models. The model prediction error is reduced by 40%. The fracture surface orientation is corrected immediately after each microseismic event, and historical data is smoothed by Kalman filtering to ensure that the model parameters are synchronized with the real-time operating conditions. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the method framework of the present invention. Detailed Implementation

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0047] Please see Figure 1 The present invention relates to a real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling, comprising the following steps:

[0048] S1. Real-time acquisition and processing of multi-source data:

[0049] The three-dimensional microseismic monitoring sensor array is used to collect real-time data on the waveforms, source locations, energy, frequency characteristics, and fracture surface orientation of microseismic events in the mine water storage layer, and the fracture propagation direction parameters are extracted based on waveform inversion.

[0050] The pressure, flow rate, flow velocity, and fluid composition data of the water seal layer are collected synchronously using a seepage monitoring device.

[0051] Three-dimensional microseismic monitoring sensors are deployed in a 10m×10m grid on the top and sidewalls of the sealed layer. The distributed fiber optic acoustic sensing (DAS) system is laid spirally along the tunnel axis with a spacing ≤5m, covering ≥95% of the target area. Short-period seismic detectors are installed in a three-component fixed configuration with high sensitivity.

[0052] ≥100V·s / m, frequency response range 0.1-100Hz.

[0053] The data acquisition frequency is 1000Hz, and the waveform is continuously recorded for 20 seconds after a single microseismic event is triggered to ensure complete capture of the crack propagation signal.

[0054] In S1, the three-dimensional microseismic monitoring sensor array includes a distributed optical fiber acoustic wave sensing (DAS) system and a short-period seismic detector. Waveform inversion uses a sliding time window spectral analysis method to extract fracture propagation direction parameters, with a time window length of 0.1-1 seconds.

[0055] Specific steps of sliding window spectral analysis:

[0056] The original waveform is bandpass filtered from 5 to 50 Hz to eliminate mechanical vibration noise.

[0057] The spectrum was extracted segment by segment using a Hanning window with a length of 0.1–1 second (step size of 0.05 seconds), and the dominant energy frequency and polarization direction were calculated.

[0058] Based on the Sato microseismic location algorithm, combined with the P-wave arrival time difference of at least 4 sensors, the three-dimensional coordinates of the seismic source are determined with a location error of ≤3m.

[0059] Definition of fracture propagation direction parameters: The fracture surface dip angle (0°-90°), strike (0°-360°) and propagation rate (0.01-1m / s) are extracted by inverting the waveform polarization angle and apparent velocity.

[0060] S2. Construction and initialization of the dynamic coupling model:

[0061] A three-dimensional seepage numerical simulation model was established, and the initialization parameters included reservoir porosity, permeability tensor, fluid injection pressure boundary conditions, and fracture spatial distribution parameters extracted by S1.

[0062] A two-way coupling mechanism between microseismic and seepage was constructed. The fracture propagation direction parameters of the microseismic module were input into the seepage model to dynamically correct the permeability field. At the same time, the fluid pressure gradient calculated by the seepage model was fed back to the microseismic module as the driving force for fracture propagation.

[0063] In S2, the seepage numerical simulation model is discretized using the finite element method. The permeability tensor is calculated using a superposition model of experimentally measured matrix permeability and fracture permeability, and its expression is as follows:

[0064]

[0065] Among them, K matrix K represents the matrix permeability. fracture Let δ be the fracture permeability, θ be the angle between the fracture orientation and the coordinate axis, and δ be the fracture permeability. ij For Kroneck's symbol.

[0066] Matrix permeability K matrix The calculation formula is as follows: Determined through core plunger experiments, using a steady-state method with a pressure difference of 0.5-10 MPa.

[0067]

[0068] Where Q is the flow rate (m 3 / s), μ is the fluid viscosity (Pa·s), L is the core length (m), and A is the cross-sectional area (m²). 2 ), where Δp is the pressure difference (Pa).

[0069] Fracture permeability K fracture Based on the law of cubes,

[0070]

[0071] b is the equivalent hydraulic opening (m), and its initial value is obtained through microseismic event energy inversion (initial b = 0.1-1 mm).

[0072] Tensor principal direction correction: For every 10 or more microseismic event clusters detected, the permeability tensor coordinate system is dynamically adjusted according to the fracture orientation θ to ensure that the principal direction is consistent with the dominant direction of fracture propagation.

[0073] In S2, the microseismic-seepage bidirectional coupling mechanism is achieved through the following methods:

[0074] Input the fracture propagation direction parameters extracted from S1 into the seepage model to dynamically adjust the principal direction of the fracture permeability tensor.

[0075] The fluid pressure gradient calculated by the seepage model is used as the effective stress, and the constitutive equation of the microseismic module is corrected by Biot consolidation theory.

[0076] Input the real-time fracture orientation θ into the seepage model, and update the principal direction of the permeability tensor using a rotation matrix. The specific formula is as follows:

[0077] K′=R(θ)KR T (θ)

[0078] Where R(θ) is a two-dimensional rotation matrix, θ is the angle between the fracture orientation and the x-axis, K is the permeability tensor in the original coordinate system, and R T (θ) is the transpose of the rotation matrix, and K′ is the permeability tensor after rotation.

[0079] Based on Biot's consolidation theory, the effective stress expression for the microseismic module is modified as follows:

[0080] σ′ ij =σ ij -αpδ ij

[0081] Where, σ′ ij For the effective stress tensor (the stress borne by the rock skeleton); σ ij This represents the stress component acting on the cross section in direction i and direction j; α is the Biot coefficient (0.8-1.0, determined based on the porosity of the rock strata); p is the fluid pressure (Pa); and the pressure gradient is calculated using the seepage model. Criterion for driving crack propagation.

[0082] S3, Real-time inversion and parameter update;

[0083] Using the Bayesian inversion algorithm, with the spatiotemporal evolution data and seepage data of microseismic events in S1 as constraints, the dynamic evolution parameters of the fractures are inverted in real time, including the permeability increment of the fracture network ΔK, the deflection angle of the maximum principal stress direction θ, the equivalent hydraulic aperture b, and the fracture geometry.

[0084] In S3, the Bayesian inversion algorithm employs the Markov chain Monte Carlo method. Its prior probability distribution is based on historical microseismic event statistics, and its posterior probability is calculated using the joint likelihood function of microseismic energy and seepage pressure gradient. The expression for this function is:

[0085] P(θ∣data)∝P(data∣θ)·P(θ)

[0086] Where θ represents the fracture parameters to be inverted, and data represents the real-time acquired microseismic and seepage data.

[0087] The permeability field of the three-dimensional seepage model and the rupture surface orientation parameters of the microseismic module are dynamically corrected based on the inversion results.

[0088] Permeability increment ΔK in fractured network: normally distributed N (μ = 0, K0 is the initial penetration rate;

[0089] Maximum principal stress deflection angle θ: uniformly distributed U(0°, 180°), calibrated based on historical geostress monitoring data;

[0090] Equivalent hydraulic opening b: Log-normal distribution, with an initial mean of 0.5 mm and a standard deviation of 0.2 mm.

[0091] Construction of joint likelihood function:

[0092] P(deta|θ)=P(micro-seismic energy|θ)·P(seepage pressure|ΔK)

[0093] The relationship between microseismic energy E and b was fitted using laboratory crack propagation experiments.

[0094] E = 0.5ρv 2 bL

[0095] (ρ is the rock density, v is the fracture propagation rate); the relationship between seepage pressure p and ΔK is based on Darcy's law.

[0096]

[0097] Where μ is the fluid viscosity, Q is the fluid flow rate, L is the seepage path length, and ΔK is the permeability increment of the fracture network.

[0098] MCMC sampling parameters:

[0099] The chain length is 100,000 cycles, the burning period is 20,000 cycles, the acceptance rate is controlled between 20% and 40%, the Metropolis-Hastings algorithm is used, the proposal distribution is a multivariate normal distribution, and the covariance matrix is ​​dynamically adjusted according to the historical inversion results.

[0100] Dynamic correction strategy

[0101] Time-triggered update: Batch inversion is initiated every 10 minutes, and the permeability field is updated based on the latest 1-hour data. The grid cell resolution is 5m×5m×5m, and the dense fracture area is densified to 2m×2m×2m.

[0102] Event-triggered update: Energy of a single microseismic event > 10 3 When J is reached, the attitude parameters of the fracture surface are immediately updated, including dip angle, strike and propagation rate, with an update delay of ≤10 seconds.

[0103] In S3, dynamic correction includes:

[0104] The permeability field of the seepage model is updated every 10 minutes;

[0105] The rupture surface orientation parameters are updated immediately after each microseismic event is triggered.

[0106] The corrected parameters are smoothed using a Kalman filter algorithm and historical data.

[0107] S4. Risk Warning and Closed-Loop Feedback:

[0108] When the permeability increment ΔK / K0 of any unit is monitored to be ≥50% or the hydraulic aperture surge rate db / dt is ≥10 -6 When m / s, the fracture instability early warning module is triggered and the set of coordinates of high-risk areas is output;

[0109] The corrected fracture parameters and early warning results are fed back to the sensor network and model initialization module of S1, forming a closed-loop system of real-time data acquisition, model update and risk warning.

[0110] The permeability field of the three-dimensional seepage model and the rupture surface orientation parameters of the microseismic module are dynamically corrected based on the inversion results.

[0111] Permeability increment ΔK in fractured network: normally distributed N (μ = 0, K0 is the initial penetration rate;

[0112] Maximum principal stress deflection angle θ: uniformly distributed U(0°, 180°), calibrated based on historical geostress monitoring data;

[0113] Equivalent hydraulic opening b: Log-normal distribution, with an initial mean of 0.5 mm and a standard deviation of 0.2 mm.

[0114] Construction of joint likelihood function:

[0115] P(deta|θ)=P(micro-seismic energy|θ)·P(seepage pressure|ΔK)

[0116] The relationship between microseismic energy E and b was fitted using laboratory crack propagation experiments.

[0117] E = 0.5ρv 2 bL

[0118] (ρ is the rock density, v is the fracture propagation rate); the relationship between seepage pressure p and ΔK is based on Darcy's law.

[0119]

[0120] Where μ is the fluid viscosity, Q is the fluid flow rate, L is the seepage path length, and ΔK is the permeability increment of the fracture network.

[0121] MCMC sampling parameters:

[0122] The chain length is 100,000 cycles, the burning period is 20,000 cycles, the acceptance rate is controlled between 20% and 40%, the Metropolis-Hastings algorithm is used, the proposal distribution is a multivariate normal distribution, and the covariance matrix is ​​dynamically adjusted according to the historical inversion results.

[0123] Dynamic correction strategy

[0124] Time-triggered update: Batch inversion is initiated every 10 minutes, and the permeability field is updated based on the latest 1-hour data. The grid cell resolution is 5m×5m×5m, and the dense fracture area is densified to 2m×2m×2m.

[0125] Event-triggered update: Energy of a single microseismic event > 10 3 When J is reached, the attitude parameters of the fracture surface are immediately updated, including dip angle, strike and propagation rate, with an update delay of ≤10 seconds.

[0126] In S4, the criteria for determining the warning threshold include:

[0127] The seepage pressure gradient exceeds a preset threshold.

[0128] The relative rate of change of the permeability increment ΔK in the fractured network exceeds 50% (ΔK / K0 ≥ 0.5).

[0129] The equivalent hydraulic opening suddenly increases beyond the critical value (db / dt≥10-6m / s).

[0130] In S4, the data synchronization frequency of the closed-loop feedback system is 1Hz, and the sensor network and model initialization module transmit data in real time via industrial Ethernet with a feedback delay of no more than 0.5 seconds.

[0131] seepage pressure gradient Based on rock mechanics experiments, when An early warning is triggered when the rock strata exceed 80% of the critical fracturing pressure gradient. The critical value is determined by indoor fracturing experiments.

[0132] ΔK / K0≥50%: According to industry standards, a sudden increase in fracture permeability exceeding 50% of matrix permeability is considered a precursor to instability.

[0133] db / dt≥10 -6 m / s: Based on the dynamic expansion simulation of the crack, an early warning is triggered when the rate of change of the crack opening exceeds 10 times the steady-state expansion rate.

[0134] Closed-loop feedback system architecture

[0135] Employing industrial Ethernet, the sensor network and model module interact in real time via the OPCUA interface, with a data packet size ≤1KB, a synchronization frequency of 1Hz, and end-to-end delay calibrated via the IEEE 1588 precision clock protocol, with an error ≤0.5 seconds.

[0136] S5. Results Output and Engineering Guidance:

[0137] The system generates real-time distribution maps of fracture permeability in the mine water storage layer, dynamic evolution curves of seepage pressure gradients, and early warning threshold reports to guide real-time adjustments to water injection and storage parameters.

[0138] In S5, the real-time distribution map of fracture permeability adopts the form of a combination of contour map and three-dimensional cloud map, the dynamic evolution curve of seepage pressure gradient includes time series data and spatial distribution heat map, and the early warning threshold report includes the three-dimensional coordinates (x,y,z) of high-risk areas and the predicted failure probability value.

[0139] In S5, the water injection and sealing parameters are adjusted as follows:

[0140] When a high-risk area is detected, the water injection pressure is automatically reduced by 10%-30%.

[0141] Based on the fracture permeability distribution map, adjust the well density and injection rate of the injection wells to ensure that the injection area avoids areas with a failure probability > 80%.

Claims

1. A real-time inversion method for fracture evolution in mine water storage layers using microseismic-seepage coupling, characterized in that, The following steps are involved: S1. Real-time acquisition and processing of multi-source data: The three-dimensional microseismic monitoring sensor array is used to collect real-time data on the waveforms, source locations, energy, frequency characteristics, and fracture surface orientation of microseismic events in the mine water storage layer, and the fracture propagation direction parameters are extracted based on waveform inversion. The pressure, flow rate, flow velocity, and fluid composition data of the water seal layer are collected synchronously using a seepage monitoring device. S2. Construction and initialization of the dynamic coupling model: A three-dimensional seepage numerical simulation model was established, and the initialization parameters included reservoir porosity, permeability tensor, fluid injection pressure boundary conditions, and fracture spatial distribution parameters extracted by S1. A two-way coupling mechanism between microseismic and seepage was constructed. The fracture propagation direction parameters of the microseismic module were input into the seepage model to dynamically correct the permeability field. At the same time, the fluid pressure gradient calculated by the seepage model was fed back to the microseismic module as the driving force for fracture propagation. S3, Real-time inversion and parameter update; Using the Bayesian inversion algorithm, with the spatiotemporal evolution data and seepage data of microseismic events in S1 as constraints, the dynamic evolution parameters of the fractures are inverted in real time, including the permeability increment of the fracture network ΔK, the deflection angle of the maximum principal stress direction θ, the equivalent hydraulic aperture b, and the fracture geometry. The permeability field of the three-dimensional seepage model and the rupture surface orientation parameters of the microseismic module are dynamically corrected based on the inversion results. S4. Risk Warning and Closed-Loop Feedback: When the permeability increment ΔK / K0 of any unit is monitored to be ≥50% or the hydraulic aperture surge rate db / dt is ≥10 -6 When m / s, the fracture instability early warning module is triggered and the set of coordinates of high-risk areas is output; The corrected fracture parameters and early warning results are fed back to the sensor network and model initialization module of S1, forming a closed-loop system of real-time data acquisition, model update and risk warning. S5. Results Output and Engineering Guidance: The system generates real-time distribution maps of fracture permeability in the mine water storage layer, dynamic evolution curves of seepage pressure gradients, and early warning threshold reports to guide real-time adjustments to water injection and storage parameters.

2. The real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling according to claim 1, characterized in that, In S1, the three-dimensional microseismic monitoring sensor array includes a distributed optical fiber acoustic wave sensing (DAS) system and a short-period seismic detector. Waveform inversion uses a sliding time window spectrum analysis method to extract the fracture propagation direction parameters, with a time window length of 0.1-1 seconds.

3. The real-time inversion method for fracture evolution in mine water sealing layers based on microseismic-seepage coupling according to claim 1, characterized in that, In S2, the seepage numerical simulation model is discretized using the finite element method. The permeability tensor is calculated using a superposition model of experimentally measured matrix permeability and fracture permeability, and its expression is: Among them, K matrix K represents the matrix permeability. fracture Let δ be the fracture permeability, θ be the angle between the fracture orientation and the coordinate axis, and δ be the fracture permeability. ij For Kroneck's symbol.

4. The real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling according to claim 1, characterized in that, In S3, the Bayesian inversion algorithm employs the Markov chain Monte Carlo method. Its prior probability distribution is based on historical microseismic event statistics, and its posterior probability is calculated using the joint likelihood function of microseismic energy and seepage pressure gradient. The expression for this posterior probability is: P(θ∣data)∝P(data∣θ)·P(θ) Where θ represents the fracture parameters to be inverted, and data represents the real-time acquired microseismic and seepage data.

5. The real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling according to claim 1, characterized in that, In step S4, the criteria for determining the warning threshold include: The seepage pressure gradient exceeds a preset threshold. The relative rate of change of the permeability increment ΔK in the fractured network exceeds 50% (ΔK / K0 ≥ 0.5). The equivalent hydraulic opening suddenly increases beyond the critical value (db / dt≥10-6m / s).

6. The real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling according to claim 1, characterized in that, In S5, the real-time distribution map of fracture permeability adopts the form of a combination of contour map and three-dimensional cloud map, the dynamic evolution curve of seepage pressure gradient includes time series data and spatial distribution heat map, and the early warning threshold report includes the three-dimensional coordinates (x, y, z) of high-risk areas and the predicted failure probability value.

7. The real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling according to claim 1, characterized in that, In S3, dynamic correction includes: The permeability field of the seepage model is updated every 10 minutes; The rupture surface orientation parameters are updated immediately after each microseismic event is triggered. The corrected parameters are smoothed using a Kalman filter algorithm and historical data.

8. The real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling according to claim 1, characterized in that, In S4, the data synchronization frequency of the closed-loop feedback system is 1Hz, and the sensor network and model initialization module transmit data in real time through industrial Ethernet, with a feedback delay of no more than 0.5 seconds.

9. The real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling according to claim 1, characterized in that, In S2, the microseismic-seepage bidirectional coupling mechanism is achieved in the following way: Input the fracture propagation direction parameters extracted from S1 into the seepage model to dynamically adjust the principal direction of the fracture permeability tensor. The fluid pressure gradient calculated by the seepage model is used as the effective stress, and the constitutive equation of the microseismic module is corrected by Biot consolidation theory.

10. The real-time inversion method for fracture evolution in mine water storage layers based on microseismic-seepage coupling according to claim 1, characterized in that, In step S5, the adjustment of water injection and sealing parameters includes: When a high-risk area is detected, the water injection pressure is automatically reduced by 10%-30%. Based on the fracture permeability distribution map, adjust the well density and injection rate of the injection wells to ensure that the injection area avoids areas with a failure probability > 80%.

Images