Mining area mining surface subsidence inversion method and device, electronic equipment and storage medium

Abstract

The invention provides a mining area mining surface subsidence inversion method, a mining area mining surface subsidence inversion device, electronic equipment and a storage medium, and relates to the technical field of mining area mining. Four second-order oscillation factors are introduced in the oscillation stage of the improved two-stage oscillation particle swarm algorithm, so that the global search capability and the local search capability of the algorithm can be balanced, and the precision of an inversion result is improved; moreover, in combination with the optimization of the fitness function by a first norm method, the influence of field measurement errors (gross errors) can be weakened, the problem of poor inversion stability caused by the gross errors is improved, and thus a more accurate inversion result is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of mining area mining, and in particular to a method, device, electronic equipment and storage medium for mining area mining surface subsidence inversion. Background Art

[0002] With the rapid development of China's economy, the amount of coal resources mined has increased year by year, but large-scale coal mining has also brought about a series of social and environmental problems such as surface subsidence, building damage, ground fissures and heavy metal pollution. Surface subsidence is one of the most common disasters in mining areas, which poses a direct threat to surface buildings and infrastructure, and may endanger life and property safety. In order to achieve a balance between resource development and ecological protection, subsidence reduction technologies such as strip mining and backfill mining have been proposed, and accurate prediction of surface subsidence has become the key to the successful implementation of these technologies. Predicting surface subsidence caused by underground mining in advance is not only helpful in formulating protection measures, reasonably arranging the scope of coal pillar retention or the relocation plan of coal-pressing villages, but also has a wide range of applications in land reclamation and coal mining subsidence management. Accurate prediction of mining subsidence is of great significance to the development of mining areas and the sustained growth of the national economy, and is particularly critical for guiding the mining practice of "three under" (under railways, buildings and water bodies) working faces.

[0003] The probability integral method is an influence function method based on the normal distribution function. It describes the surface subsidence basin in an integral form and is suitable for conventional surface movement and deformation calculations. As the main method specified in China's "Regulations on Coal Pillar Retention and Compressed Coal Mining for Buildings, Water Bodies, Railways and Main Wells and Lanes", the probability integral method has become the most widely used method in the field of surface subsidence prediction after years of research. Particle Swarm Optimization (PSO) is a swarm intelligence optimization algorithm with the advantages of simple operation, fast convergence speed, few parameters and easy implementation. It is widely used in the field of engineering optimization. The probability integral method model is a complex nonlinear model. The swarm intelligence algorithm has better results in processing such models. However, the real geological conditions are often complex, and the traditional particle swarm optimization algorithm is also very prone to random instability or premature convergence problems, resulting in poor accuracy and precision of the inversion results. Summary of the invention

[0004] The object of the present invention is to provide a method, device, electronic equipment and storage medium for inversion of surface subsidence in mining area, so as to improve the accuracy and precision of the inversion result.

[0005] In a first aspect, the present invention provides a method for inversion of surface subsidence in mining areas, comprising:

[0006] Acquire model building data of a first sub-area and inversion basic data of a second sub-area in the target mining area; wherein the model building data includes first engineering data, observed subsidence data and parameter search ranges of model parameters, and the inversion basic data includes second engineering data;

[0007] According to the model construction data, the target probability integral model is established using the improved second-order oscillation particle swarm optimization algorithm and the linear norm method; the speed update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function;

[0008] According to the inversion basic data and the target probability integral model, the surface subsidence inversion is carried out on the second sub-area to obtain the target inversion result.

[0009] Furthermore, the first engineering data and the second engineering data both include the designed advancement length of the working face strike, the designed length of the inclination, the average inclination angle of the mined coal seam and the average thickness of the coal seam; the observed subsidence data include the observed surface settlement values ​​of multiple observation points; the model parameters include the subsidence coefficient, the tangent value of the main influencing angle, the mining influence propagation angle, the strike inflection point offset distance, the uphill direction inflection point offset distance and the downhill direction inflection point offset distance.

[0010] Furthermore, according to the model construction data, the target probability integral model is established using the improved second-order oscillating particle swarm optimization algorithm and the first-order norm method, including:

[0011] Using the probability integral method, an initial probability integral model is established, and the initial probability integral model is used to characterize the corresponding relationship between the model parameters and the subsidence data;

[0012] According to the improved second-order oscillating particle swarm optimization algorithm and the linear norm method, the position update formula, the speed update formula and the target fitness function are determined; wherein the target fitness function is obtained based on the optimization of the linear norm method;

[0013] According to the model construction data, the initial probability integral model is trained using the position update formula, the speed update formula and the target fitness function to obtain the trained target probability integral model.

[0014] Furthermore, the corresponding relationship between the model parameters and the subsidence data includes:

[0015]

[0016] Wherein, W(x, y) represents the expected surface settlement value of the observation point (x, y), u and v represent the coordinates of the mining unit in strike and dip, respectively, U1 and U2 represent the minimum and maximum values ​​in the value range of u, V1 and V2 represent the minimum and maximum values ​​in the value range of v, V1=O0, V2=T1-O0, U1=O1·sin(θ+α) / sinθ, U2=T2-O2·sin(θ+α) / sinθ, r=t / tanβ, T1 represents the designed advancement length of the working face strike, T2 represents the designed length of dip, α represents the average inclination angle of the mined coal seam, t represents the average thickness of the coal seam, q represents the subsidence coefficient, tanβ represents the tangent value of the main influencing angle, θ represents the mining influence propagation angle, O0 represents the strike inflection point offset, O1 represents the uphill inflection point offset, and O2 represents the downhill inflection point offset.

[0017] Furthermore, the location update formula includes:

[0018]

[0019] The velocity update formula includes:

[0020]

[0021] The target fitness function includes:

[0022]

[0023] in, represents the position of the i-th particle at the k-th iteration, represents the position of the i-th particle at the k-1th iteration, represents the position of the i-th particle at the k-2th iteration, represents the current optimal position of the i-th particle, m gbest represents the current optimal position of the population, represents the velocity of the i-th particle at the k-th iteration, represents the velocity of the ith particle at the k-1th iteration, ω represents the preset inertia weight, c1 and c2 represent the preset individual learning factor and social learning factor respectively, r1 and r2 represent two independent uniform random numbers in the range of [0, 1], ξ1, ξ2, ξ3, ξ4 represent four second-order oscillation factors; f i k represents the fitness value of the i-th particle at the k-th iteration, represents the weight of the nth observation point of the ith particle at the kth iteration, Indicates that at the kth iteration, the i-th particle m iThe residual square of the nth observation point, N represents the total number of observation points, fitness(m) represents the original fitness function of the improved second-order oscillatory particle swarm optimization algorithm under the model parameter m, S represents the observed surface settlement value of each observation point, W(m) represents the mapping function between the model parameter m and the expected surface settlement value W of each observation point, It represents the difference between the observed surface settlement value and the expected surface settlement value corresponding to the ith particle at the kth iteration, and σ represents the preset value.

[0024] Furthermore, when the training is in the early oscillation convergence stage, the four second-order oscillation factors satisfy the following formula:

[0025]

[0026] When the training is in the late stage of asymptotic convergence, the four second-order oscillation factors satisfy the following formula:

[0027] ξ1<ξ2-1, ξ3<ξ4-1, 0<ξ2<1, 0<ξ4<1.

[0028] Furthermore, after performing surface subsidence inversion on the second sub-area based on the inversion basic data and the target probability integral model and obtaining the target inversion result, the method further includes:

[0029] The target inversion result is visualized to obtain a target image; wherein the target image includes a settlement contour map and / or a three-dimensional settlement map.

[0030] In a second aspect, the present invention further provides a mining surface subsidence inversion device, comprising:

[0031] An acquisition module is used to acquire model building data of a first sub-area and inversion basic data of a second sub-area in the target mining area; wherein the model building data includes first engineering data, observed subsidence data and parameter search ranges of model parameters, and the inversion basic data includes second engineering data;

[0032] Establishing a module for constructing data according to the model, and establishing a target probability integral model using an improved second-order oscillation particle swarm optimization algorithm and a linear norm method; wherein the speed update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function;

[0033] The inversion module is used to perform surface subsidence inversion on the second sub-area according to the inversion basic data and the target probability integral model to obtain the target inversion result.

[0034] In a third aspect, the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the mining area surface subsidence inversion method of the first aspect is implemented.

[0035] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the mining surface subsidence inversion method of the first aspect is executed.

[0036] The mining surface subsidence inversion method, device, electronic device and storage medium provided by the embodiment of the present invention can obtain model construction data of the first sub-area and inversion basic data of the second sub-area in the target mining area; wherein the model construction data includes the first engineering data, the observed subsidence data and the parameter search range of the model parameters, and the inversion basic data includes the second engineering data; according to the model construction data, a target probability integral model is established using an improved second-order oscillation particle swarm optimization algorithm and a linear norm method; wherein the speed update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function; according to the inversion basic data and the target probability integral model, surface subsidence inversion is performed on the second sub-area to obtain a target inversion result. In this way, the improved secondary oscillation particle swarm algorithm is combined with the first norm to robustly estimate the model parameters of surface subsidence. The improved secondary oscillation particle swarm algorithm introduces four second-order oscillation factors in the oscillation stage, which can balance the global search ability and local search ability of the algorithm, thereby improving the accuracy of the inversion result; and, combined with the optimization of the fitness function by the first norm method, the influence of field measurement errors (gross errors) can be weakened, and the problem of poor inversion stability caused by gross errors can be improved, thereby obtaining more accurate inversion results. Therefore, the mining surface subsidence inversion method, device, electronic device and storage medium provided by the embodiment of the present invention improve the accuracy and precision of the inversion results. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 A schematic diagram of a process flow of a method for inversion of surface subsidence in mining areas provided by an embodiment of the present invention;

[0039] Figure 2A schematic diagram of the structure of a mining area surface subsidence inversion device provided by an embodiment of the present invention;

[0040] Figure 3 A schematic diagram of the structure of another mining area surface subsidence inversion device provided by an embodiment of the present invention;

[0041] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] Intelligent optimization algorithms can be divided into two parts: the first part is exploration, which is to globally search for potential feasible solutions; the second part is development, which is to further search for the optimal solution based on the first part. How to reasonably balance these two parts has always been a problem that intelligent optimization algorithms need to solve.

[0044] At present, since the probability integral method is a highly nonlinear problem, the least squares gradient algorithm is often very sensitive to the initial model parameters and gross errors, resulting in distortion of the inversion results and deviation from the actual situation. Particle swarm algorithm, as an intelligent optimization algorithm inspired by bird groups in nature, has been widely used in the inversion problem in the engineering field. Although the particle swarm algorithm is easy to implement and converges quickly, its standard version still has the risk of premature convergence, and the particle position update only relies on inertia, individual optimal position and global optimal position, and does not take into account the influence of individual particles and particle position changes on itself, so that the particle update does not make full use of effective information. Based on this, a method, device, electronic device and storage medium for inversion of surface subsidence in mining area provided by the embodiment of the present invention adopts an improved secondary oscillation particle swarm algorithm, which can further improve the problems existing in the standard particle swarm algorithm, and combines the first norm method to reduce the influence of gross errors, and successfully applies it to the probability integral method inversion of mining surface subsidence parameters to obtain more accurate inversion results.

[0045] To facilitate understanding of this embodiment, a method for inversion of surface subsidence in mining areas disclosed in an embodiment of the present invention is first introduced in detail.

[0046] The embodiment of the present invention provides a method for inversion of mining surface subsidence, which can be executed by an electronic device with data processing capability. Figure 1The flowchart of a method for inversion of surface subsidence in mining area shown in FIG. 1 mainly includes the following steps S110 to S130:

[0047] Step S110, obtaining model building data of the first sub-area and inversion basic data of the second sub-area in the target mining area; wherein the model building data includes the first engineering data, the observed subsidence data and the parameter search range of the model parameters, and the inversion basic data includes the second engineering data.

[0048] The target mining area is a mining area where surface subsidence estimation is required, and the target mining area may include a first sub-area and a second sub-area; wherein the first sub-area may be an area where a surface movement and deformation observation station is set up, and the surface movement and deformation observation station can measure and obtain observed subsidence data; the second sub-area may be the area in the target mining area excluding the first sub-area, or it may be the entire target mining area.

[0049] The model construction data of the first sub-area can be used to construct a target probability integral model, which, combined with the inversion basic data of the second sub-area, can realize the surface subsidence inversion of the second sub-area. The model construction data includes the first engineering data, the observed subsidence data and the parameter search range of the model parameters. The inversion basic data includes the second engineering data. Both the first engineering data and the second engineering data can include the designed advancement length (T1) of the working face, the designed length of inclination (T2), the average mined coal seam inclination (α) and the average coal seam thickness (t). The observed subsidence data can include the observed surface settlement values ​​of multiple observation points; the surface settlement value is the vertical displacement before and after mining. The surface settlement value actually measured is called the observed surface settlement value, and the surface settlement value subsequently predicted by the model is called the expected surface settlement value. The model parameters can include the subsidence coefficient (q), the tangent value of the main influence angle (tanβ), the mining influence propagation angle (θ), the strike inflection point offset (O0), the uphill inflection point offset (O1) and the downhill inflection point offset (O2). The value range of the model parameters (i.e., the parameter search range of the model parameters) can be estimated through continuous settlement monitoring data of the mining area and actual geological conditions. For example, a possible parameter search range is: q (0.5-0.99), tanβ (1.5-3.0), θ (80°-90°), O0, O1, O2 (10m-50m).

[0050] In specific implementation, when obtaining model construction data, the geological conditions, mining conditions and surrounding geological data of each working face or mining area can be obtained; wherein, a mining area can include one or more working faces, and the geological conditions can include coal seam dip (α), average coal seam thickness (t) and lithological characteristics (whether there are faults or folds), etc. If the observation point is located near the fault, its prediction error may be large, so the observed surface settlement value of the observation point has a small weight in the model; mining conditions can include working face strike advancement length (T1), dip advancement length (T2), mining method and mining depth, etc.; surrounding geological data can include surface morphology, groundwater conditions, observed subsidence data (S) and surface cracks, etc. The working face strike advancement length (T1), dip advancement length (T2), coal seam dip (α) and coal seam thickness (t) in the geological conditions and mining conditions can be used as the first engineering data, and the first engineering data is the input of the model. The observed subsidence data in the surrounding geological data is used as the observed subsidence data in the model construction data, and is the output associated with the first engineering data during model training. Constraint data such as lithological characteristics, mining methods, mining depth, and surrounding geological data will constrain model parameters and indirectly affect the inversion results; the parameter search range of model parameters can be determined based on these constraint data. Subsequently, the first engineering data and the parameter search range of model parameters can be used as inputs and constraints of the model, and the model can be trained in combination with the observed subsidence data to obtain the target probability integral model. Then, the second engineering data and the coordinates (x, y) of each point in the work area (i.e., the second sub-area) are input into the target probability integral model for calculation, and the corresponding expected surface settlement value W (x, y) can be obtained.

[0051] Step S120, based on the model construction data, a target probability integral model is established using an improved second-order oscillation particle swarm optimization algorithm and a linear norm method; wherein the speed update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function.

[0052] The embodiment of the present invention proposes a new method for inversion prediction of mining surface subsidence parameters, that is, using an improved second-order oscillating particle swarm algorithm combined with a first-order norm to robustly estimate the surface subsidence parameters. The surface subsidence prediction model (i.e., the target probability integral model) in the inversion is based on the most widely used probability integral method, which plays an important role in reducing surface subsidence disasters in mining areas and reducing economic losses. In some possible embodiments, the above step S120 may include the following sub-steps:

[0053] Sub-step 121, using the probability integration method, establishes an initial probability integration model, and the initial probability integration model is used to characterize the corresponding relationship between the model parameters and the subsidence data.

[0054] Optionally, the corresponding relationship between the above model parameters and the subsidence data includes:

[0055]

[0056] Wherein, W(x, y) represents the expected surface settlement value of the observation point (x, y), u and v represent the coordinates of the mining unit in strike and dip, respectively, U1 and U2 represent the minimum and maximum values ​​in the value range of u, V1 and V2 represent the minimum and maximum values ​​in the value range of v, V1=O0, V2=T1-O0, U1=O1·sin(θ+α) / sinθ, U2=T2-O2·sin(θ+α) / sinθ, r=t / tanβ, T1 represents the designed advancement length of the working face strike, T2 represents the designed length of dip, α represents the average inclination angle of the mined coal seam, t represents the average thickness of the coal seam, q represents the subsidence coefficient, tanβ represents the tangent value of the main influencing angle, θ represents the mining influence propagation angle, O0 represents the strike inflection point offset, O1 represents the uphill inflection point offset, and O2 represents the downhill inflection point offset.

[0057] Sub-step 122, determining a position update formula, a velocity update formula and a target fitness function according to an improved second-order oscillating particle swarm optimization algorithm and a linear norm method; wherein the target fitness function is obtained by optimization based on the linear norm method;

[0058] In this embodiment, the sum of the square of the residual of the predicted surface settlement value W and the observed surface settlement value S corresponding to each observation point and the product of the observation point weight is used as the target fitness function.

[0059] Optionally, the above position update formula includes:

[0060]

[0061] The velocity update formula includes:

[0062]

[0063] The target fitness function includes:

[0064]

[0065] in, represents the position of the i-th particle at the k-th iteration, represents the position of the i-th particle at the k-1th iteration, represents the position of the i-th particle at the k-2th iteration, represents the current optimal position of the i-th particle, m gbest represents the current optimal position of the population, represents the velocity of the i-th particle at the k-th iteration, represents the velocity of the ith particle at the k-1th iteration, ω represents the preset inertia weight, c1 and c2 represent the preset individual learning factor and social learning factor respectively, r1 and r2 represent two independent uniform random numbers in the range of [0, 1], ξ1, ξ2, ξ3, ξ4 represent four second-order oscillation factors; f i k represents the fitness value of the i-th particle at the k-th iteration, represents the weight of the nth observation point of the ith particle at the kth iteration, Indicates that at the kth iteration, the i-th particle m i The residual square of the nth observation point, N represents the total number of observation points, fitness(m) represents the original fitness function of the improved second-order oscillatory particle swarm optimization algorithm under the model parameter m, S represents the observed surface settlement value of each observation point, W(m) represents the mapping function between the model parameter m and the expected surface settlement value W of each observation point, It represents the difference between the observed surface settlement value and the expected surface settlement value corresponding to the ith particle at the kth iteration, and σ represents the preset value.

[0066] Sub-step 123, based on the model construction data, the initial probability integral model is trained using the position update formula, the speed update formula and the target fitness function to obtain a trained target probability integral model.

[0067] The improved second-order oscillating particle swarm optimization algorithm and the first-order norm method are used to optimize the fitness function so that the fitness value decreases continuously, that is, the settlement curve predicted by the model is gradually similar to the observed settlement curve; in the iterative process, the local optimal solution is continuously jumped out, and finally the optimal inversion result within the given parameter search range is obtained; the six model parameter values ​​obtained by inversion are used to establish the probability integral model of the entire target mining area.

[0068] During the training process, the convergence speed can be used to determine whether the current phase is in the early oscillation convergence phase or the late progressive convergence phase; the above four second-order oscillation factors are randomly selected at each iteration. Optionally, in order to improve the global convergence performance and improve the calculation accuracy, when the training is in the early oscillation convergence phase, the four second-order oscillation factors satisfy the following formula:

[0069]

[0070] When the training is in the late stage of asymptotic convergence, the four second-order oscillation factors satisfy the following formula:

[0071] ξ1<ξ2-1, ξ3<ξ4-1, 0<ξ2<1, 0<ξ4<1.

[0072] Step S130, performing surface subsidence inversion on the second sub-region according to the inversion basic data and the target probability integral model to obtain a target inversion result.

[0073] The inversion basic data is used as the input of the target probability integral model. The corresponding settlement value is calculated for each coordinate point inside the mining area using the established target probability integral model and the target inversion result is obtained.

[0074] In order to facilitate users to intuitively view the target inversion results, the above-mentioned surface subsidence inversion is performed on the second sub-area based on the inversion basic data and the target probability integral model. After obtaining the target inversion results, the above-mentioned method also includes: visualizing the target inversion results to obtain a target image; wherein the target image includes a settlement contour map and / or a three-dimensional settlement map.

[0075] ArcGIS and other tools can be used to generate settlement contour maps and three-dimensional settlement maps. Users can visually view the distribution of settlement in the mining area and further propose corresponding settlement reduction measures based on the target inversion results, optimize the mining plan of the mining area, and ensure safe production and sustainable development of the mining area.

[0076] The mining surface subsidence inversion method provided by the embodiment of the present invention can obtain model construction data of the first sub-area and inversion basic data of the second sub-area in the target mining area; wherein the model construction data includes the first engineering data, the observed subsidence data and the parameter search range of the model parameters, and the inversion basic data includes the second engineering data; according to the model construction data, a target probability integral model is established using an improved second-order oscillation particle swarm optimization algorithm and a linear norm method; wherein the velocity update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function; according to the inversion basic data and the target probability integral model, surface subsidence inversion is performed on the second sub-area to obtain a target inversion result. In this way, the improved secondary oscillation particle swarm algorithm is combined with the first norm to robustly estimate the model parameters of surface subsidence. The improved secondary oscillation particle swarm algorithm introduces four second-order oscillation factors in the oscillation stage, which can balance the global search ability and local search ability of the algorithm, thereby improving the accuracy of the inversion result; and, combined with the optimization of the fitness function by the first norm method, the influence of the field measurement error (gross error) can be weakened, and the problem of poor inversion stability caused by the gross error can be improved, thereby obtaining a more accurate inversion result. Therefore, the mining surface subsidence inversion method provided by the embodiment of the present invention improves the accuracy and precision of the inversion result.

[0077] To facilitate understanding, the inversion method for surface subsidence in the above-mentioned mining area is introduced in detail below.

[0078] Step 1: Obtain the first engineering data of the target mining area.

[0079] It includes: the designed advancement length of the working face (T1), the designed length of inclination (T2), the average inclination angle of the mined coal seam (α), the average coal seam thickness (t), and the observed subsidence data (S) measured by the surface movement and deformation observation station; among them, the mean square error of the height difference per kilometer of the observed subsidence data is no more than 10 mm, the units of T1, T2, and D are all meters (m), and the unit of S is millimeters (mm).

[0080] Step 2: Determine the model parameters and the parameter variation search range.

[0081] There are six model parameters, namely, subsidence coefficient (q), tangent value of main influencing angle (tanβ), mining influence propagation angle (θ), strike inflection point offset (O0), uphill inflection point offset (O1) and downhill inflection point offset (O2). The value ranges of the above six parameters are estimated by continuous settlement monitoring data of the mining area and actual geological conditions. The parameter search range used in the embodiment of the present invention is: q (0.5-0.99), tanβ (1.5-3.0), θ (80°-90°), O0, O1, O2 (10m-50m).

[0082] Step 3: Forward process and fitness function.

[0083] The estimated surface settlement value W is calculated according to the probability integration method. The model parameters of the probability integration method are inverted, that is, given the observed settlement data S, the model parameters m are estimated, m = (q, tanβ, θ, O0, O1, O2); the probability integration method establishes the mathematical relationship between the model parameter m and the surface settlement, which constitutes the forward process of the inversion. The expression of the estimated surface settlement value at any point on the surface is as follows:

[0084]

[0085] Among them, W(x, y) represents the expected surface settlement value of the observation point (x, y), u and v represent the coordinates of the mining unit in strike and dip, respectively, and the coordinate units are all meters (m); U1 and U2 represent the minimum and maximum values ​​in the value range of u, V1 and V2 represent the minimum and maximum values ​​in the value range of v, V1=O0, V2=T1-O0, U1=O1·sin(θ+α) / sinθ, U2=T2-O2·sin(θ+α) / sinθ, r=t / tanβ.

[0086] The fitness function is the inversion objective function, which means that the L2 norm square of the residual between the observed surface settlement value S and the expected surface settlement value W is the smallest, S = (S1, S2, ... S n ), W=(W1,W2,……W n), n represents the nth observation point of surface subsidence, and N represents the total number of observation points:

[0087]

[0088] Where W(m) represents the mapping function between the model parameter m and the expected surface settlement value W at each observation point. The intelligent optimization algorithm can be used to fit and converge the formula (2).

[0089] In this embodiment, the square of the residual of the predicted surface settlement value W and the observed surface settlement value S corresponding to each observation point and the product of the observation point weight are used as the target fitness function. For details, see the subsequent fifth step.

[0090] Step 4: Clarify the population dimension of the inversion problem and set the algorithm parameters.

[0091] A population includes multiple particles, and the location of each particle is a potential solution to the fitness function. The fitness function is shown in the third step. It is necessary to determine the location of the six settlement prediction model parameters and calculate the estimated surface settlement value through the second step, and calculate the fitness value. By continuously changing the particle position, the fitness value is continuously reduced until convergence, and the optimal model parameter position is obtained. Each particle has a dimension of 6 and can predict the settlement value.

[0092] Setting the algorithm parameters includes: setting the appropriate number of particles (i.e., the population size). The more particles there are in the population, the stronger the corresponding global search capability, but the longer the calculation time required. Randomly initialize a group of particles in the search space (i.e., randomly determine the initial population in the search space), and the total number can be ten times the particle dimension; setting the inertia weight, learning factor (including individual learning factor and social learning factor), update speed parameter, and number of iterations.

[0093] Step 5: Apply optimization algorithm for inversion.

[0094] The improved second-order oscillating particle swarm optimization algorithm and the first norm method are used to optimize the fitness function. First, the application of the traditional particle swarm algorithm in estimating sedimentation parameters is introduced. The particle swarm algorithm speed update formula is as follows:

[0095]

[0096] Where i=1, 2, ... l represents the i-th particle among a total of l particles, k represents the number of iterations, represents the current optimal position of the i-th particle, m gbestrepresents the current optimal position of the population, ω represents the inertia weight, c1 and c2 represent the acceleration coefficients that control the maximum update size, also known as individual learning factors and social learning factors, r1 and r2 represent two independent uniform random numbers in the range of [0, 1]. After r1 and r2 are selected, they remain unchanged in different iterations. The position update of the i-th particle at the k-th iteration is as follows (model parameter update equation, that is, position update formula):

[0097]

[0098] The second-order oscillation particle swarm optimization algorithm improves the speed update formula of the standard particle swarm optimization algorithm by introducing a second-order oscillation mechanism, thereby enhancing the diversity of particles and improving the global search capability. This improved method performs well in the search area of ​​the position function. After the introduction of the second-order oscillation mechanism, the algorithm shows a strong global search capability in the early stage and gradually converges through oscillation; in the later stage, it focuses on local search and gradually reaches an accurate convergence state. The evolutionary formula for its speed update is as follows:

[0099]

[0100] When the algorithm's early oscillation converges:

[0101]

[0102] When the algorithm converges gradually in the later stage:

[0103]

[0104] The improved second-order oscillating particle swarm algorithm analyzes the relationship between these parameters and their value range by setting the parameters to four different values, thereby improving the diversity of the group. This method effectively improves the global convergence performance of the algorithm and improves the calculation accuracy. Its speed update formula is as follows:

[0105]

[0106] The algorithm with different parameters in the oscillation link is called the improved second-order oscillating particle swarm algorithm.

[0107] The following are the ranges of the four parameters when the algorithm converges in the early oscillation stage:

[0108]

[0109] When the algorithm converges gradually in the later stage:

[0110] ξ1<ξ2-1, ξ3<ξ4-1, 0<ξ2<1, 0<ξ4<1. (10)

[0111] ξ1, ξ2, ξ3, ξ4 are called second-order oscillation factors, which are random values ​​that satisfy equations (9) and (10) (randomly selected in each iteration). Parameters can be continuously iterated and updated using equation (4), and the fitness value continues to decrease and converge. However, when the measurement error of the measured data is large, it will cause difficulty in parameter inversion and deviate from the true value. Therefore, the embodiment of the present invention combines the first-order norm robust estimation method with the improved second-order oscillation particle swarm algorithm to obtain more accurate and robust estimated parameters. First, it is assumed that each observation point n has the same accuracy and weight, that is, Pn = 1; after one iteration of the optimization algorithm, the residual between the expected value of each observation point (that is, the expected surface settlement value) and the measured value (that is, the observed surface settlement value) is calculated, and the weight of each observation point is adjusted according to the size of the residual, thereby introducing the influence of weight in the original fitness function calculation.

[0112]

[0113] represents the weight of the nth observation point of the ith particle at the kth iteration, represents the difference between the observed surface settlement value and the expected surface settlement value corresponding to the ith particle at the kth iteration, and σ represents the preset value, which is a very small number to prevent the denominator from being zero, for example, 2.22×10 -16 .

[0114] After adding the weight, the fitness value of the i-th particle at the k-th iteration can be expressed as:

[0115]

[0116] Indicates that at the kth iteration, the i-th particle m i The squared residual of the nth observation point is used to multiply the squared residual of each observation point, thereby adjusting the influence of these observation points on the inversion results. In this way, the robust estimation reduces the influence of outliers on the inversion results, making the inversion process more stable and accurate. The fitness value f after adding the weight is optimized by the optimization algorithm. i k It keeps decreasing until the appropriate convergence accuracy is met or the preset number of iterations is reached, and then the iteration stops.

[0117] Step 6: Jump out of local optimum.

[0118] The improved second-order oscillation particle swarm algorithm has a strong exploration capability in the early stage and can screen out suitable particles; in the later stage, it improves the development capability and accelerates the convergence speed of the algorithm. The algorithm increases the diversity of the group by selecting different ξ1, ξ2, ξ3, ξ4 and accurately determining their value ranges, thereby optimizing the global convergence. In the process of gradually decreasing fitness values, the second-order oscillation factor guides the movement direction of the particles towards a more optimal angle range, which helps the particles to jump out of the local optimal solution and improve the ability of global search.

[0119] Step 7: Establish a probability integral model to complete the prediction of mining area subsidence.

[0120] After estimating q, tanβ, θ, O0, O1, and O2, the target probability integral model can be established, and the geological conditions and mining conditions of each working face or mining area can be used as the input of the model. The corresponding settlement value is calculated for each coordinate point inside the mining area using the established target probability integral model using a numerical method to complete the settlement prediction, thereby achieving a comprehensive prediction and evaluation of the settlement of the entire mining area, which is helpful to formulate a reasonable mining plan and reduce the impact of surface settlement on buildings and infrastructure.

[0121] The advantages of the embodiments of the present invention are that the least squares gradient algorithm can effectively avoid the problem of over-reliance on the initial model and falling into the local minimum, and can also improve the problem of poor inversion stability caused by field measurement errors (gross errors), and can obtain more accurate prediction results. The new optimization algorithm improves the second-order oscillation particle swarm optimization method, and introduces different parameter settings in the oscillation stage to balance the global search and local search capabilities of the algorithm, thereby improving the accuracy of the inversion results, and introduces a norm after each iteration to weight the prediction error of each observation point, thereby reducing the impact of data gross errors.

[0122] Corresponding to the above-mentioned mining area surface subsidence inversion method, the embodiment of the present invention also provides a mining area surface subsidence inversion device. Figure 2 The schematic diagram of the structure of a mining area surface subsidence inversion device is shown, and the device comprises:

[0123] An acquisition module 201 is used to acquire model building data of a first sub-area and inversion basic data of a second sub-area in a target mining area; wherein the model building data includes first engineering data, observed subsidence data and parameter search ranges of model parameters, and the inversion basic data includes second engineering data;

[0124] Establishing module 202, for establishing a target probability integral model according to model construction data by using an improved second-order oscillation particle swarm optimization algorithm and a linear norm method; wherein the speed update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function;

[0125] The inversion module 203 is used to perform surface subsidence inversion on the second sub-area according to the inversion basic data and the target probability integral model to obtain a target inversion result.

[0126] The mining surface subsidence inversion device provided by the embodiment of the present invention can obtain model construction data of the first sub-area and inversion basic data of the second sub-area in the target mining area; wherein the model construction data includes the first engineering data, the observed subsidence data and the parameter search range of the model parameters, and the inversion basic data includes the second engineering data; according to the model construction data, a target probability integral model is established using an improved second-order oscillation particle swarm optimization algorithm and a linear norm method; wherein the speed update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function; according to the inversion basic data and the target probability integral model, surface subsidence inversion is performed on the second sub-area to obtain a target inversion result. In this way, the improved secondary oscillation particle swarm algorithm is combined with the first norm to perform robust estimation of the model parameters of surface subsidence. The improved secondary oscillation particle swarm algorithm introduces four second-order oscillation factors in the oscillation stage, which can balance the global search ability and local search ability of the algorithm, thereby improving the accuracy of the inversion result; and, combined with the optimization of the fitness function by the first norm method, the influence of the field measurement error (gross error) can be weakened, and the problem of poor inversion stability caused by the gross error can be improved, thereby obtaining a more accurate inversion result. Therefore, the mining surface subsidence inversion device provided by the embodiment of the present invention improves the accuracy and precision of the inversion result.

[0127] Furthermore, the above-mentioned first engineering data and second engineering data include the designed advancement length of the working face strike, the designed length of the inclination, the average inclination angle of the mined coal seam and the average thickness of the coal seam; the observed subsidence data include the observed surface settlement values ​​of multiple observation points; the model parameters include the subsidence coefficient, the tangent value of the main influencing angle, the mining influence propagation angle, the strike inflection point offset distance, the uphill direction inflection point offset distance and the downhill direction inflection point offset distance.

[0128] Furthermore, the above-mentioned establishment module 202 is specifically used for: using the probability integration method to establish an initial probability integration model, the initial probability integration model is used to characterize the corresponding relationship between model parameters and sinking data; according to the improved second-order oscillation particle swarm optimization algorithm and the linear norm method, the position update formula, the speed update formula and the target fitness function are determined; wherein the target fitness function is obtained based on the optimization of the linear norm method; according to the model construction data, the initial probability integration model is trained using the position update formula, the speed update formula and the target fitness function to obtain the trained target probability integration model.

[0129] Furthermore, the corresponding relationship between the above model parameters and the subsidence data includes:

[0130]

[0131] Wherein, W(x, y) represents the expected surface settlement value of the observation point (x, y), u and v represent the coordinates of the mining unit in strike and dip, respectively, U1 and U2 represent the minimum and maximum values ​​in the value range of u, V1 and V2 represent the minimum and maximum values ​​in the value range of v, V1=O0, V2=T1-O0, U1=O1·sin(θ+α) / sinθ, U2=T2-O2·sin(θ+α) / sinθ, r=t / tanβ, T1 represents the designed advancement length of the working face strike, T2 represents the designed length of dip, α represents the average inclination angle of the mined coal seam, t represents the average thickness of the coal seam, q represents the subsidence coefficient, tanβ represents the tangent value of the main influencing angle, θ represents the mining influence propagation angle, O0 represents the strike inflection point offset, O1 represents the uphill inflection point offset, and O2 represents the downhill inflection point offset.

[0132] Furthermore, the above position update formula includes:

[0133]

[0134] The velocity update formula includes:

[0135]

[0136] The target fitness function includes:

[0137]

[0138] in, represents the position of the i-th particle at the k-th iteration, represents the position of the i-th particle at the k-1th iteration, represents the position of the i-th particle at the k-2th iteration, represents the current optimal position of the i-th particle, m gbest represents the current optimal position of the population, represents the velocity of the i-th particle at the k-th iteration, represents the velocity of the ith particle at the k-1th iteration, ω represents the preset inertia weight, c1 and c2 represent the preset individual learning factor and social learning factor respectively, r1 and r2 represent two independent uniform random numbers in the range of [0, 1], ξ1, ξ2, ξ3, ξ4 represent four second-order oscillation factors; f i k represents the fitness value of the i-th particle at the k-th iteration, represents the weight of the nth observation point of the ith particle at the kth iteration, Indicates that at the kth iteration, the i-th particle m i The residual square of the nth observation point, N represents the total number of observation points, fitness(m) represents the original fitness function of the improved second-order oscillatory particle swarm optimization algorithm under the model parameter m, S represents the observed surface settlement value of each observation point, W(m) represents the mapping function between the model parameter m and the expected surface settlement value W of each observation point, It represents the difference between the observed surface settlement value and the expected surface settlement value corresponding to the ith particle at the kth iteration, and σ represents the preset value.

[0139] Furthermore, when the training is in the early oscillation convergence stage, the four second-order oscillation factors satisfy the following formula:

[0140]

[0141] When the training is in the late stage of asymptotic convergence, the four second-order oscillation factors satisfy the following formula:

[0142] ξ1<ξ2-1, ξ3<ξ4-1, 0<ξ2<1, 0<ξ4<1.

[0143] Further, see Figure 3 The schematic diagram of another structure of the mining surface subsidence inversion device is shown in FIG. Figure 2 On the basis of, the above device also includes:

[0144] The visualization module 301 is used to perform visualization processing on the target inversion result to obtain a target image; wherein the target image includes a settlement contour map and / or a three-dimensional settlement map.

[0145] The mining area surface subsidence inversion device provided in this embodiment has the same implementation principle and technical effects as those of the aforementioned mining area surface subsidence inversion method embodiment. For the sake of brief description, for matters not mentioned in the mining area surface subsidence inversion device embodiment, reference may be made to the corresponding contents in the aforementioned mining area surface subsidence inversion method embodiment.

[0146] like Figure 4 As shown, an electronic device 400 provided by an embodiment of the present invention includes: a processor 401, a memory 402 and a bus, the memory 402 stores a computer program that can be run on the processor 401, when the electronic device 400 is running, the processor 401 and the memory 402 communicate through the bus, and the processor 401 executes the computer program to implement the above-mentioned mining surface subsidence inversion method.

[0147] Specifically, the memory 402 and the processor 401 can be general-purpose memories and processors, which are not specifically limited here.

[0148] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the mining surface subsidence inversion method in the above method embodiment is executed. The computer-readable storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a RAM, a magnetic disk or an optical disk, and other media that can store program codes.

[0149] The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the term "at least one" herein represents any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can represent including any one or more elements selected from the set consisting of A, B, and C.

[0150] In all examples shown and described herein, any specific values ​​should be interpreted as merely exemplary and not as limiting, and thus other examples of the exemplary embodiments may have different values.

[0151] The flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and a part of a module, a program segment or a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of the boxes in the block diagram and / or the flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0152] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or modules can be electrical, mechanical or other forms.

[0153] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0154] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for inversion of surface subsidence in mining areas, characterized in that: include: Acquire model building data of a first sub-area and inversion basic data of a second sub-area in the target mining area; wherein the model building data includes first engineering data, observed subsidence data and parameter search ranges of model parameters, and the inversion basic data includes second engineering data; According to the model construction data, a target probability integral model is established using an improved second-order oscillation particle swarm optimization algorithm and a linear norm method; wherein the speed update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function; According to the inversion basic data and the target probability integral model, surface subsidence inversion is performed on the second sub-area to obtain a target inversion result.

2. The method according to claim 1, characterized in that The first engineering data and the second engineering data both include the designed advancement length of the working face strike, the designed length of the inclination, the average inclination of the mined coal seam and the average thickness of the coal seam; the observed subsidence data include the observed surface settlement values ​​of multiple observation points; the model parameters include the subsidence coefficient, the tangent value of the main influencing angle, the mining influence propagation angle, the strike inflection point offset distance, the uphill direction inflection point offset distance and the downhill direction inflection point offset distance.

3. The method according to claim 2, characterized in that The method of constructing data according to the model and establishing a target probability integral model using an improved second-order oscillating particle swarm optimization algorithm and a linear norm method includes: Using the probability integration method, an initial probability integration model is established, wherein the initial probability integration model is used to characterize the corresponding relationship between the model parameters and the subsidence data; According to the improved second-order oscillation particle swarm optimization algorithm and the linear norm method, a position update formula, a speed update formula and a target fitness function are determined; wherein the target fitness function is obtained based on the optimization of the linear norm method; According to the model construction data, the initial probability integral model is trained using the position update formula, the speed update formula and the target fitness function to obtain the trained target probability integral model.

4. The method according to claim 3, characterized in that The corresponding relationship between the model parameters and the subsidence data includes: Wherein, W(x, y) represents the expected surface settlement value of the observation point (x, y), u and v represent the coordinates of the mining unit in strike and dip, respectively, U1 and U2 represent the minimum and maximum values ​​in the value range of u, V1 and V2 represent the minimum and maximum values ​​in the value range of v, V1=O0, V2=T1-O0, U1=O1·sin(θ+α) / sinθ, U2=T2-O2·sin(θ+α) / sinθ, r=t / tanβ, T1 represents the designed advancement length of the working face strike, T2 represents the designed length of dip, α represents the average inclination angle of the mined coal seam, t represents the average thickness of the coal seam, q represents the subsidence coefficient, tanβ represents the tangent value of the main influencing angle, θ represents the mining influence propagation angle, O0 represents the strike inflection point offset, O1 represents the uphill inflection point offset, and O2 represents the downhill inflection point offset.

5. The method according to claim 3, characterized in that: The position update formula includes: The speed update formula includes: The target fitness function includes: in, represents the position of the i-th particle at the k-th iteration, represents the position of the i-th particle at the k-1th iteration, represents the position of the i-th particle at the k-2th iteration, represents the current optimal position of the i-th particle, m gbest represents the current optimal position of the population, represents the velocity of the i-th particle at the k-th iteration, represents the velocity of the ith particle at the k-1th iteration, ω represents the preset inertia weight, c1 and c2 represent the preset individual learning factor and social learning factor respectively, r1 and r2 represent two independent uniform random numbers in the range of [0, 1], ξ1, ξ2, ξ3, ξ4 represent the four second-order oscillation factors; represents the fitness value of the i-th particle at the k-th iteration, represents the weight of the nth observation point of the ith particle at the kth iteration, Indicates that at the kth iteration, the i-th particle m i The residual square of the nth observation point, N represents the total number of observation points, fitness(m) represents the original fitness function of the improved second-order oscillating particle swarm optimization algorithm under the model parameter m, S represents the observed surface settlement value of each observation point, W(m) represents the mapping function between the model parameter m and the expected surface settlement value W of each observation point, It represents the difference between the observed surface settlement value and the expected surface settlement value corresponding to the ith particle at the kth iteration, and σ represents the preset value.

6. The method according to claim 5, characterized in that When the training is in the early oscillation convergence stage, the four second-order oscillation factors satisfy the following formula: When the training is in the late progressive convergence stage, the four second-order oscillation factors satisfy the following formula: ξ1<ξ2-1, ξ3<ξ4-1, 0<ξ2<1, 0<ξ4<1.

7. The method according to claim 1, characterized in that After performing surface subsidence inversion on the second sub-area according to the inversion basic data and the target probability integral model to obtain a target inversion result, the method further includes: The target inversion result is visualized to obtain a target image; wherein the target image includes a settlement contour map and / or a three-dimensional settlement map.

8. A mining area surface subsidence inversion device, characterized in that: include: An acquisition module, used to acquire model building data of a first sub-area and inversion basic data of a second sub-area in a target mining area; wherein the model building data includes first engineering data, observed subsidence data and parameter search ranges of model parameters, and the inversion basic data includes second engineering data; Establishing a module, used to construct data according to the model, and establish a target probability integral model using an improved second-order oscillation particle swarm optimization algorithm and a linear norm method; wherein the speed update formula of the improved second-order oscillation particle swarm optimization algorithm includes four second-order oscillation factors, and the linear norm method is used to optimize the fitness function; An inversion module is used to perform surface subsidence inversion on the second sub-area according to the inversion basic data and the target probability integral model to obtain a target inversion result.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, the mining surface subsidence inversion method described in any one of claims 1-7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the mining surface subsidence inversion method described in any one of claims 1 to 7 is executed.

Images