Intelligent mine physical property parameter real-time inversion and coal mining simulation real-time optimization method

Abstract

The invention relates to the technical field of intelligent mine mining, in particular to an intelligent mine physical property parameter real-time inversion and coal mining simulation real-time optimization method. The method specifically comprises the steps of arranging monitoring points, and monitoring physical property parameter stress data sigma and displacement data u in real time; establishing a mathematical model composed of a geometric equation, an equilibrium equation, a constitutive equation and initial-boundary conditions; a solver is used for simulating the coal mining process, and simulated stress data sigma simulation and displacement data u simulation of each monitoring point are obtained; comparing the real-time simulation stress data sigma simulation and the displacement data u simulation of each monitoring point with the field real-time monitoring stress data sigma monitoring and the displacement data u monitoring, and constructing a loss function; and by utilizing an optimization algorithm, the physical property parameters of the stratum are determined while the coal mining simulation is more accurate. According to the method, the accuracy and optimization effect of coal mining simulation can be improved, so that the safety and resource utilization rate of mine mining are improved.

Description

Technical Field

[0001] The present invention relates to the field of smart mine mining technology, and in particular to a method for using on-site monitoring data to invert smart mine stratum physical property parameters in real time and optimize coal seam mining simulation results. Background Art

[0002] With the continuous development of mining technology, smart mines have become an important part of modern mining. Smart mines mainly use advanced information technology, Internet of Things, big data, artificial intelligence and other technical means to carry out intelligent transformation of various links such as mine production, management, and safety, so as to achieve efficient, safe, environmentally friendly and sustainable development of mine production. Smart mines rely on a large amount of real-time monitoring data and numerical simulation technology to achieve efficient mining and management of mine resources.

[0003] Coal mining simulation refers to the virtual simulation of the coal mining process through computer simulation technology to evaluate and optimize the mining plan and improve mining efficiency and safety. Coal mining simulation involves geological modeling, mining process simulation, environmental impact assessment and other aspects, and is an important tool in modern mining engineering. The first is geological data collection, mainly through drilling, logging and other means to obtain geological data of coal seams, including coal seam thickness, inclination, faults, joints and other information; then use these geological data to establish a three-dimensional geological model of the coal seam to accurately describe the morphology and distribution of the coal seam.

[0004] Traditional coal mining simulation is usually based on static formation physical parameters, which come from historical data or prior models. However, the formation physical parameters measured by current geophysical exploration technology have problems such as large discreteness, high noise and many outliers, high data processing costs and insufficient accuracy. In addition, due to the complexity of geological conditions and the dynamic changes in formation characteristics during coal mining, the simulation results that rely solely on static models are difficult to accurately reflect the actual mining situation, thus affecting the safety and economy of coal mining. Therefore, how to use field monitoring data to invert formation physical parameters in real time and optimize mining simulation results has become an urgent problem to be solved in the current smart mining field. Summary of the invention

[0005] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a method for real-time inversion of smart mine physical property parameters and real-time optimization of coal seam mining simulation results using field monitoring data, so as to improve the accuracy and optimization effect of coal seam mining simulation, thereby improving the safety of mine mining and resource utilization.

[0006] To solve the problems of the prior art, the present invention discloses a real-time inversion method for smart mine physical property parameters and real-time optimization method for coal seam mining simulation. The first step is to arrange monitoring points and monitor the physical property parameters in real time: M strata are established above the coal mining seam, and Young's modulus E and Poisson's ratio v of each stratum are assigned values, N monitoring points are set in the tunnel of the coal mining seam, and rock formation stress data σ and displacement data u at each monitoring point are collected in real time.

[0007] Preferably, M strata are established above the coal mining seam, where M is determined based on actual conditions and M≥25.

[0008] Preferably, the N monitoring points set in the tunnel of the coal mining seam are arranged at equal intervals, and N ≥ 3. More preferably, two adjacent monitoring points are arranged at an interval of 10 meters.

[0009] Preferably, the Young's modulus E and Poisson's ratio v of each stratum are in the range of: 10 8 ≤E≤10 11 , 0.1≤v≤0.4.

[0010] Step 2: Modeling the coal seam mining area:

[0011] The geometric equation of the model is expressed as formula (1), Where ε is the strain tensor, u represents the displacement,

[0012] The equilibrium equation of this model is expressed as formula (2), ▽·σ+f=0(2), where σ is the stress tensor and f is the body force.

[0013] The relationship between σ and ε in equations (1) and (2), σ(ε), is called the constitutive equation. For viscoelastic materials, stress is a function of the strain rate.

[0014] Assuming no-slip boundary conditions on both sides and the bottom of the tunnel, the fixed constraint and normal displacement constraint are expressed as formula (3) and formula (4) respectively:

[0015] u| Γ =0 (3), where u represents displacement, Γ represents the boundary of the lane,

[0016] u·n Γ′ =0 (4), where u represents the displacement, n is the unit vector of the external normal at the boundary, and Γ' represents the bottom of the computational domain with fixed constraints.

[0017] Preferably, setting the two sides and the bottom of the roadway as no-slip boundary conditions is to fix the displacement of the two sides and the bottom of the roadway to 0.

[0018] More preferably, in a specific three-dimensional case, the displacement field is u(x, y, z), then the displacement components in the x, y, and z directions are all zero, and the fixed constraint condition is:

[0019] u x |Γ= 0 (9),u y |Γ=0 (10),u z |Γ=0 (11).

[0020] The stress and normal stress boundary conditions are expressed as formula (5) and formula (6), respectively.

[0021] n·σ Γ″ =t0(5), where n is the unit vector of the external normal at the boundary, t0 is the surface force vector of the rock mass at the boundary, Γ” is the boundary area where the stress is applied,

[0022] n·σ·n Γ″ =-p0(6), where n is the unit vector of the external normal at the boundary, p0 is the pressure on the rock mass at the boundary, and Γ” is the boundary area where the stress is applied.

[0023] The gravity field added to the model is expressed as formula (7), f g =ρg(7), where ρ is the density of the formation, unit: kg / m 3 , g is the gravitational acceleration vector, in three dimensions g = (0,0,-9.81) m / s 2 .

[0024] Preferably, the gravity acceleration vector g is in the range of (0, 0, -9.81) m / s 2 Fixed value in range.

[0025] Step 3: Use the solver to simulate the coal mining process and obtain the simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 :The established model formulas (1)-(7) are input into the solver, and according to the actual coal seam mining situation, the initial conditions and boundary conditions of the simulation are set, including mining depth, mining speed, tunnel shape and size, geological structure and external force, to obtain the simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 .

[0026] Preferably, the solver used is the open source software OpenGeoSys.

[0027] Step 4: Compare the real-time simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 Real-time monitoring of stress data on site 监测and displacement data u 监测 , construct the loss function:

[0028] The real-time simulated stress data σ 模拟 , real-time simulation displacement data 模拟 The stress data of the real-time monitoring on site σ 监测 And real-time monitoring of displacement data on site 监测 Alignment is performed to ensure that the two sets of data match in space and time. Based on the aligned two sets of data, the loss function is constructed by taking the sum of the squares of the differences between the actual monitoring data on site and the corresponding simulation data. In the case of the isotropic linear elastic constitutive equation, the expression of the loss function L is formula (8):

[0029]

[0030] Wherein, N is the total number of monitoring points in step (S1), including k stress monitoring points and Nk displacement monitoring points, i is each natural number less than or equal to N, M is the total number of strata divided in step (S1), j is each natural number less than or equal to M, E j is the Young's modulus of the jth layer, v j is the Poisson's ratio of the jth stratum.

[0031] Step 5: Use optimization algorithm to determine formation physical parameters:

[0032] Set the initial physical property parameters of each layer, Young's modulus E 初始 and Poisson's ratio v 初始 , use different optimization methods to optimize the loss function, and determine the optimization algorithm after comparison.

[0033] The optimization algorithm is used to optimize the loss function shown in formula (8) so that the simulated stress data σ obtained in the solver is 模拟 And on-site real-time monitoring of stress dataσ 监测 , simulated displacement data u 模拟 And real-time monitoring of displacement data on site 监测 The difference between them becomes smaller and smaller until the value of the loss function L calculated according to formula (8) is within the threshold range, and the formation physical property parameters are determined.

[0034] Preferably, the determined optimization algorithm is a sequential quadratic programming method.

[0035] The present invention combines field monitoring data with numerical simulation to invert the stratum physical parameters of the smart mine in real time and optimize the coal seam mining simulation results, which has the following significant beneficial effects:

[0036] 1. Accuracy and safety: This application constructs a closed-loop feedback system, which realizes real-time optimization of the coal mining process by continuously optimizing the loss function and dynamically updating the formation physical property parameters, ensuring a high degree of consistency between the simulation results and the actual mining situation. At the same time, the present invention can reflect the stress and displacement changes of the formation during coal mining in real time, and by accurately inverting the formation physical property parameters, potential dangerous areas can be identified in advance, thereby enhancing the safety of coal mining.

[0037] 2. Low cost and universality: The present invention can replace traditional methods to obtain more accurate formation physical parameters, reduce data processing costs, and improve data processing efficiency and accuracy. Since the formation physical parameters can be inverted and updated in real time, the method of the present invention can adapt to complex geological conditions and dynamic changes in formation characteristics during coal mining, ensuring the wide adaptability of the model in practical applications.

[0038] 3. While the present invention introduces new technologies into the field of transparent geology, it also provides sufficient data support for the field of smart mining. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the two-dimensional structure of the present invention;

[0040] Figure 2 This is a comparison chart of optimization errors using three line search techniques;

[0041] Figure 3 It is a line graph of the errors of the three line search techniques between each iteration step. DETAILED DESCRIPTION

[0042] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the structure of the present invention.

[0043] A method for real-time inversion of physical property parameters of smart mines and real-time optimization of coal seam mining simulation. The first step is to arrange monitoring points to monitor physical property parameters in real time. Figure 1 As shown, M strata are established above the coal mining seam, and Young's modulus E and Poisson's ratio v are assigned to each stratum. N monitoring points are set in the tunnel of the coal mining seam, and the rock stratum stress data σ and displacement data u at each monitoring point are collected in real time.

[0044] There are M strata established above the coal mining seam, where M is determined based on actual conditions and M≥25.

[0045] The N monitoring points set in the tunnel of the coal mining seam are set at equal intervals, and N≥3. More preferably, two adjacent monitoring points are set at an interval of 10 meters.

[0046] The Young's modulus E and Poisson's ratio v of each stratum have a range of values: 108 ≤E≤10 11 , 0.1≤v≤0.4.

[0047] The Young's modulus E and Poisson's ratio v of each stratum are different. In this embodiment, we set M to 32. Then, according to the situation, we set the Young's modulus of the 1st to 25th strata to 1×10 10 , and the Young's modulus of the 26th to 30th strata are 2.8×10 10 , 1.5×10 9 , 6×10 9 , 6×10 9 , 6×10 9 , the Young's modulus of the 31st-32nd stratum is 1×10 10 ; Set the Poisson's ratio of the 1st to 32nd layers to 0.4.

[0048] Step 2: Model the coal seam mining area. The mechanical mathematical model of rock deformation consists of geometric equations, equilibrium equations, constitutive equations and initial-boundary conditions.

[0049] The geometric equation of the model is expressed as formula (1), Where ε is the strain tensor, u represents the displacement,

[0050] The equilibrium equation of this model is expressed as formula (2), ▽·σ+f=0(2), where σ is the stress tensor and f is the body force.

[0051] The relationship between σ and ε in equations (1) and (2), σ(ε), is called the constitutive equation. For viscoelastic materials, stress is a function of the strain rate.

[0052] Regarding the setting of initial-boundary conditions, the cross section of the tunnel is set to be rectangular, and the two sides and the bottom are set to no-slip boundary conditions. The fixed constraint and normal displacement constraint are expressed as formula (3) and formula (4) respectively:

[0053] u| Γ =0 (3), where u represents displacement, Γ represents the boundary of the lane,

[0054] u·n Γ′ =0 (4), where u represents the displacement, n is the unit vector of the external normal at the boundary, and Γ' represents the bottom of the computational domain with fixed constraints.

[0055] Setting the two sides and the bottom of the roadway as no-slip boundary conditions is to fix the displacement of the two sides and the bottom of the roadway to 0. In the specific three-dimensional case, the displacement field is u(x, y, z), then the displacement components in the x, y, and z directions are all zero, and the fixed constraint conditions are:

[0056] u x|Γ= 0 (9),u y |Γ=0 (10),u z |Γ=0 (11).

[0057] Its stress and normal stress boundary conditions are expressed as formula (5) and formula (6), respectively.

[0058] n·σ Γ″ =t0(5), where n is the unit vector of the external normal at the boundary, t0 is the surface force vector of the rock mass at the boundary, Γ” is the boundary area where the stress is applied,

[0059] n·σ·n Γ″ =-p0(6), where n is the unit vector of the external normal at the boundary, p0 is the pressure on the rock mass at the boundary, and Γ” is the boundary area where the stress is applied.

[0060] Considering the influence of vertical pressure on the formation, in order to fully ensure the reliability of the model, we add the gravity field into the model. The gravity field is expressed as formula (7), f g =ρg(7), where ρ is the density of the formation, unit: kg / m 3 , g is the gravitational acceleration vector, in three dimensions g = (0,0,-9.81) m / s 2 .

[0061] The gravity acceleration vector g is between (0,0,-9.81) m / s 2 Fixed value in range.

[0062] Step 3: Use the solver to simulate the coal mining process and obtain the simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 :The established model formulas (1)-(7) are input into the OpenGeoSys solver, and according to the actual coal seam mining situation, the initial conditions and boundary conditions of the simulation are set, including mining depth, mining speed, tunnel shape and size, geological structure and external force, to obtain the simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 When the mining vehicle moves forward in the coal seam area, the simulated stress data σ of the monitoring point is calculated once for each mining time step. 模拟 and displacement data u 模拟 , and finally the multiple sets of measurements at multiple time steps are collated and summarized.

[0063] Step 4: Compare the real-time simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 Real-time monitoring of stress data on site 监测 and displacement data u 监测, construct the loss function:

[0064] The real-time simulated stress data σ 模拟 , real-time simulation displacement data 模拟 The stress data of the real-time monitoring on site σ 监测 And real-time monitoring of displacement data on site 监测 Alignment is performed to ensure that the two sets of data match in space and time. Based on the aligned two sets of data, the loss function is constructed by taking the sum of the squares of the differences between the actual monitoring data on site and the corresponding simulation data. In the case of the isotropic linear elastic constitutive equation, the expression of the loss function L is formula (8):

[0065]

[0066] Wherein, N is the total number of monitoring points in step (S1), including k stress monitoring points and Nk displacement monitoring points, i is each natural number less than or equal to N, M is the total number of strata divided in step (S1), j is each natural number less than or equal to M, E j is the Young's modulus of the jth layer, v j is the Poisson's ratio of the jth stratum.

[0067] Step 5: Use optimization algorithm to determine formation physical parameters:

[0068] Set the initial physical property parameters of each layer, Young's modulus E 初始 and Poisson's ratio v 初始 , use different optimization methods to optimize the loss function, and determine the optimization algorithm after comparison.

[0069] The optimization algorithm is used to optimize the loss function shown in formula (8) so that the simulated stress data σ obtained in the solver is 模拟 And on-site real-time monitoring of stress dataσ 监测 , simulated displacement data u 模拟 And real-time monitoring of displacement data on site 监测 The difference between them becomes smaller and smaller until the value of the loss function L calculated according to formula (8) is within the threshold range, and the formation physical property parameters are determined.

[0070] When we set the Young's modulus of the 1st to 25th layers to 1×10 10 , and the Young's modulus of the 26th to 30th strata are 2.8×10 10 , 1.5×10 9 , 6×10 9 , 6×10 9 , 6×10 9 , the Young's modulus of the 31st-32nd stratum is 1×10 10; Set the Poisson's ratio of the 1st to 32nd strata to 0.4. Then use different optimization algorithms to optimize the loss function formula (8). Here we determine the sequential quadratic programming method as the optimization algorithm. During the optimization, in each iteration, the updated parameter values ​​are applied to the numerical simulation, and the real-time simulated stress data σ of the monitoring point is recalculated. 模拟 , real-time simulation displacement data 模拟 , and calculate the loss function L formula (8) again. This process continues until the value of the loss function L converges to a smaller error range, indicating that the formation parameters are determined.

[0071] like Figure 2 As shown in the figure, in the sequential quadratic programming algorithm, three different inexact linear search techniques, namely Armijo criterion, Wolfe criterion and Fletcher criterion, are used for optimization. It can be seen that:

[0072] 1. With the same number of iterations, the optimization effect of the Wolfe criterion is better. After 25 iterations, the root mean square error of the loss function is only 6.2855;

[0073] 2. When the optimization effect is the same, the Wolfe criterion takes the shortest time;

[0074] 3. Overall effect: Wolfe criterion is better than Fletcher criterion and better than Armijo criterion.

[0075] and Figure 3 The figure shows the relative error of Young's modulus E between each iteration step after optimization using three different inexact linear search techniques: Armijo criterion, Wolfe criterion, and Fletcher criterion. The change in E at each iteration step can also be seen intuitively from the figure, indicating that the above optimization algorithm is effective and further verifies the convergence of the algorithm.

[0076] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for real-time inversion of smart mine physical property parameters and real-time optimization of coal seam mining simulation, characterized in that: The following steps are involved: (S1) Arrange monitoring points and monitor physical parameters in real time: establish M strata above the coal mining seam, assign values ​​to the Young's modulus E and Poisson's ratio v of each stratum, set N monitoring points in the tunnel of the coal mining seam, and collect stress data σ and displacement data u of the overburden stratum at each monitoring point in real time; (S2) Modeling of coal seam mining areas: The geometric equation of the model is expressed as formula (1), Where ε is the strain tensor, u represents the displacement, The equilibrium equation of this model is expressed as formula (2): Where σ is the stress tensor, f is the body force, The relationship between σ and ε in formula (1) and formula (2) is called the constitutive equation. For viscoelastic materials, stress is a function of strain rate. Assuming no-slip boundary conditions on both sides and the bottom of the tunnel, the fixed constraint and normal displacement constraint are expressed as formula (3) and formula (4) respectively: u| Γ =0 (3), where u represents displacement, Γ represents the boundary of the lane, u·n| Γ′ =0 (4), where u represents displacement, n is the unit vector of the external normal at the boundary, Γ' represents the bottom of the computational domain with fixed constraints, The stress and normal stress boundary conditions are expressed as formula (5) and formula (6), respectively. n·σ| Γ″ =t0 (5), where n is the unit vector of the external normal at the boundary, t0 is the surface force vector of the rock mass at the boundary, Γ” is the boundary area where the stress is applied, n·σ·n| Γ″ =-p0 (6), where n is the unit vector of the external normal at the boundary, p0 is the pressure on the rock mass at the boundary, Γ” is the boundary area where the stress is applied, The vertical gravity field added to the model is expressed as formula (7): f g =ρg (7), where ρ is the density of the formation, unit: kg / m 3 , g is the gravitational acceleration vector, in three dimensions g = (0, 0, -9.81) m / s 2 ; (S3) Use the solver to simulate the coal seam mining process and obtain the simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 :The established model formulas (1)-(7) are input into the solver, and according to the actual coal seam mining situation, the initial conditions and boundary conditions of the simulation are set, including mining depth, mining speed, tunnel shape and size, geological structure and external force, to obtain the simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 ; (S4) Compare the real-time simulated stress data σ of each monitoring point 模拟 and displacement data u 模拟 Real-time monitoring of stress data on site 监测 and displacement data u 监测 , construct the loss function: The real-time simulated stress data σ 模拟 , real-time simulation displacement data 模拟 The stress data of the real-time monitoring on site σ 监测 And real-time monitoring of displacement data on site 监测 Alignment is performed to ensure that the two sets of data match in space and time. Based on the aligned two sets of data, the loss function is constructed by taking the sum of the squares of the differences between the actual monitoring data on site and the corresponding simulation data. In the case of the isotropic linear elastic constitutive equation, the expression of the loss function L is formula (8): Wherein, N is the total number of monitoring points in step (S1), including k stress monitoring points and Nk displacement monitoring points, i is each natural number less than or equal to N, M is the total number of strata divided in step (S1), j is each natural number less than or equal to M, E j is the Young's modulus of the jth layer, v j is the Poisson’s ratio of the jth stratum; (S5) Determine the formation physical property parameters using an optimization algorithm: Set the initial physical property parameters of each layer, Young's modulus E 初始 and Poisson's ratio v 初始 , use different optimization methods to optimize the loss function, and determine the optimization algorithm after comparison; The optimization algorithm is used to optimize the loss function shown in formula (8) so that the simulated stress data σ obtained in the solver is 模拟 And on-site real-time monitoring of stress dataσ 监测 , simulated displacement data u 模拟 And real-time monitoring of displacement data on site 监测 The difference between them becomes smaller and smaller until the value of the loss function L calculated according to formula (8) is within the threshold range, and the formation physical property parameters are determined.

2. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 1 is characterized by: The fixed constraint condition shown in formula (3) in step (S2) is that in a specific three-dimensional case, the displacement field is u(x, y, z), then the displacement components in the x, y, and z directions are all zero, and the fixed constraint condition is: in x |Γ=0 (9), in y |Γ=0 (10), in z |F=0 (11).

3. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 1 is characterized by: The gravity acceleration vector g in formula (7) in step (S2) is a value between (0, 0, -9.81) m / s 2 Fixed value in range.

4. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 1 is characterized by: In step (S1), M strata are established above the coal mining seam, where M is determined based on actual conditions and M≥25.

5. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 1 is characterized by: In step (S1), N monitoring points are set in the tunnel of the coal mining seam at equal intervals, and N≥3.

6. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 5 is characterized by: The two adjacent monitoring points in step (S1) are set at an interval of 10 meters.

7. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 1 is characterized by: The solver used in step (S3) is the open source software OpenGeoSys.

8. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 1 is characterized by: The optimization algorithm determined in step (S5) is a sequential quadratic programming method.

9. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 1 is characterized by: The Young's modulus E and Poisson's ratio v of each stratum in step (S1) have a value range of: 10 8 ≤E≤10 11 , 0.1≤v≤0.

4.

10. The method for real-time inversion of intelligent mine physical property parameters and real-time optimization of coal seam mining simulation according to claim 1, characterized in that: Step (S1) sets the two sides and the bottom of the tunnel as no-slip boundary conditions to fix the displacement of the two sides and the bottom of the tunnel to 0.

Images