Coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment

Abstract

The invention discloses a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment, and relates to the technical field of coal mine working face reconstruction under incomplete projection conditions. According to the sum of ray intercept in each network, performing multi-scale grid division through search space pruning; on the basis of considering the requirements of population diversity and rapid convergence of the multi-population genetic algorithm, a multi-scale reconstruction target function is constructed based on the divided multi-scale grids, and solving is carried out through the dynamic multi-population genetic algorithm. Therefore, by adopting the coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment, the number of populations can be dynamically adjusted and optimized, so that the diversity of the populations can be maintained, the search efficiency is improved, and the stability and convergence speed of working face reconstruction operation are improved.

Description

Technical Field

[0001] The invention relates to the technical field of coal mine working face reconstruction under incomplete projection conditions, and in particular to a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment. Background Art

[0002] Intelligent mining of coal mines is a trend of high-quality development in the industry. The refined reconstruction and prediction of the working face is the basis and prerequisite for reducing the number of people in the mine, or even eliminating them. As a non-contact detection method, electromagnetic wave CT technology for the working face has low detection cost and high efficiency, and is widely used in coal mining enterprises. However, due to the limitations of the closed space underground in coal mines and the narrow and long distribution of the working face, the scanning angle of the working face exploration and observation space is limited, and the effective CT projection data collected is often incomplete and highly sparse. This reconstruction technology under incomplete projection conditions has always been a research hotspot in the industry.

[0003] At present, when performing CT reconstruction calculations on coal mine working faces, it is necessary to discretize the detection and reconstruction area into regular rectangular units. The scale of the divided grid units directly affects the efficiency and resolution of the inversion operation. In general, the scale of the grid division is often determined according to the scale of the exploration area, the distribution of the emission points and the receiving points. The appropriate grid division scale can more accurately reflect the occurrence state of the coal seams in the working face. Due to the limitations of exploration conditions and underground structural conditions, the density of the emission points, receiving points and rays is usually unevenly covered, making it difficult to select a grid scale suitable for the entire calculation area. In 2001, E. Kisslin et al. proposed a three-grid method, which uses a smaller grid scale for ray tracing, selects the most suitable scale as the velocity structure of the detection medium, and uses a larger scale for inversion calculation as the best model parameterization method. In 2003, Hu Wei used multiple grid scales from small to large to perform inversion calculations in the selection of grid scales for seismic travel time inversion, and took the average of the calculation results of each grid scale as the final result according to the set weights. This method solves the problem of uneven ray coverage to a certain extent, but increases the calculation time. In 2005, Ma Detang et al. also proposed a dual-grid seismic tomography method, but for the convenience of calculation, the imaging grid scale was required to be an integer multiple of the ray tracing grid.

[0004] Existing research has found that the key to achieving good reconstruction results is to determine the optimal grid scale for forward and inversion before calculation and perform calculations at this scale. However, in practical problems, the selection of the optimal scale for forward and inversion is not easy, and how to finely match the grid subdivision scale to meet the needs of reconstruction of different working face observation systems has not been effectively improved.

[0005] Therefore, it is necessary to provide a multi-scale grid adaptive partitioning strategy to solve the problem of grid scale partitioning by pruning the reconstruction search space, so as to make full use of the ray coverage information of the working surface CT exploration observation system. Summary of the invention

[0006] The purpose of the present invention is to provide a coal mine working face reconstruction method based on search space pruning and dynamic adjustment of multiple populations, which can perform fine grid division on the working face exploration area, and at the same time obtain the optimal solution in terms of comprehensive reconstruction accuracy, stability and efficiency, so as to realize effective reconstruction of the coal mine working face under the conditions of incomplete CT projection.

[0007] To achieve the above object, the present invention provides a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment, comprising the following steps:

[0008] S1. Perform preliminary grid division on the exploration area of ​​the working face, and perform multi-scale grid division by search space pruning based on the sum of ray intercepts in each network;

[0009] S2. Based on the divided multi-scale grids, a multi-scale reconstruction objective function is constructed, and the population is eliminated and dynamically adjusted through a dynamic multi-population genetic algorithm to obtain a super population. Then, a refined search is performed on the super population, and the multi-scale reconstruction objective function is iteratively solved to obtain the electromagnetic wave absorption coefficient corresponding to the multi-scale grid.

[0010] Preferably, step S1 includes: first, preliminarily dividing the exploration area of ​​the working face into square grids according to the size of the working face; second, dividing each preliminarily divided square grid into four grids; then, determining the grids that need to be divided again according to the ray intercepts in the grids after the four-division, and iterating in sequence until the minimum scale limit is reached.

[0011] Preferably, in step S2, the multi-scale reconstruction objective function is expressed as follows:

[0012]

[0013] Where A is the total number of rays, B is the number of grids, X = (x 1 ,x 2 ,···,x j ,···,x B ) is the absorption coefficient vector of B grids to be inverted, H i is the electromagnetic wave field strength data collected by the i-th ray, r i is the length of the i-th ray, H 0 is the initial field strength of the launch, d ij is the intercept of the i-th ray at the j-th grid.

[0014] Preferably, the solution is obtained by a dynamic multi-population genetic algorithm, including:

[0015] S21, initialize the population, each population independently performs selection, crossover and mutation operations, explores different areas of the solution space, and records the elite individuals of all populations;

[0016] S22, regularly introduce the best individuals from various populations into the target population through the immigration operator, so that the populations can co-evolve;

[0017] S23, set a fixed evaluation period, calculate the average fitness of each sub-population, screen the sub-populations according to the fitness ranking, and merge the screened sub-populations and their corresponding elite individuals to obtain a super population;

[0018] S24. Use smaller crossover probability and mutation probability to conduct a detailed search on the super population until the preset maximum number of iterations is reached.

[0019] Preferably, different control parameters are used for each population.

[0020] Therefore, the present invention adopts the above-mentioned coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment, which has the following technical effects:

[0021] (1) The present invention effectively integrates the grids that are not passed through by the observed rays and reconstructs the effective pruning of the search space, making it suitable for exploration scales of different working faces and ray distribution characteristics.

[0022] (2) The present invention is based on a dynamic multi-population genetic algorithm, which significantly improves the accuracy, stability and efficiency of the reconstruction results by dynamically adjusting the number of populations.

[0023] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a numerical model of a working face CT exploration system in an embodiment of a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment;

[0025] Figure 2 It is a flow chart of a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment;

[0026] Figure 3 It is a schematic diagram of traditional grid division in an embodiment of a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment;

[0027] Figure 4It is a distribution diagram of the number of rays at different scales in traditional grid division in an embodiment of a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment;

[0028] Figure 5 This is a stability analysis diagram of reconstruction operations at different scales of traditional grid division in an embodiment of a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment. DETAILED DESCRIPTION

[0029] The present invention can be explained in more detail by the following examples. The purpose of disclosing the present invention is to protect all changes and improvements within the scope of the present invention. The present invention is not limited to the following examples.

[0030] like Figure 1 As shown in the figure, the numerical model of the CT exploration and observation system of the working face has a length L of 200m and a width W of 100m. There are two common geological anomaly areas in the numerical model of the CT exploration and observation system of the working face, and the electromagnetic wave absorption coefficient of the abnormal area is β 1 =1.0dB / m, the electromagnetic wave absorption coefficient of the rest of the normal area is β 2 =0.5dB / m. In the observation system layout, the spacing between the transmitting points is 30m, the corresponding spacing between the receiving points is 10m, and the total number of rays GS = 420.

[0031] like Figure 2 As shown, the present invention provides a coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment, comprising the following steps:

[0032] S1. Due to the complex environment of coal mine working faces and the characteristics of electromagnetic wave propagation, electromagnetic wave CT projection data is often incomplete, resulting in errors and uncertainties in the CT reconstruction results of the working face. In addition, the grid division of the working face reconstruction area is an important step in the reconstruction calculation, and the grid division scale directly affects the efficiency and accuracy of the inversion operation. Figure 3 As shown in the figure, before the work surface reconstruction calculation, the target detection area needs to be grid discretized and divided into B square grids. Each grid is called a pixel and is represented by x. j Represents the true absorption coefficient of the j-th grid, j = 1, 2, 3, …, B.

[0033] At present, uniform grid scale division is usually adopted in the reconstruction calculation of working face CT. However, in order to match the dense ray distribution, this method often results in dense grid division and time-consuming calculation. In addition, there are many network areas where no rays pass through, which will increase the stability and speed of solving the problem. If the grid spacing is simply increased and the number of grids is reduced, the inversion resolution will be reduced. Specifically, the CT exploration size of the working face is 200m*100m. According to the traditional uniform grid scale division method, the uniform grid division scales are set to 25m, 20m, 12.5m, 10m, 6.25m, 3.125m, 2m, and 1.5625m respectively. It is found that as the uniform grid division scale becomes smaller, the number of grids that are not passed through by the observed rays increases almost exponentially, such as Figure 4 As shown in the figure, these grids that are not traversed by the rays do not carry any valid information of the projection data and also need to participate in the reconstruction operation process, which not only reduces the efficiency of reconstruction, but also increases the difficulty of reconstructing and searching for the optimal solution.

[0034] In order to reduce the number of grids that are not passed by observed rays, invalid grids need to be integrated, that is, the scale of the grid is reduced in the area with dense ray coverage, and the scale of the grid is appropriately increased in the area with low ray coverage coefficient, so that each grid is passed by rays and carries effective observation information. This embodiment adopts a multi-scale adaptive grid division method, which effectively prunes the target search space of the working surface reconstruction to finely match the distribution characteristics of different working surface exploration sizes and different densities of rays, and fully utilizes the effective information in the incomplete projection data of the working surface CT exploration observation system, reduces the number of invalid grids, and improves the efficiency and accuracy of reconstruction, as follows:

[0035] First, according to the CT exploration size of the working surface of 200m*100m, the greatest common divisor is determined to be 100m, the side length of the largest square divided is ensured, and the number of squares is determined, and finally the working surface is divided into 2 squares of 100m*100m. The specific operation is: starting from one corner of the rectangle, according to the calculated square side length, squares are divided inside the rectangle in sequence; after each division, the boundaries of the remaining rectangles are updated, and this process is repeated until the entire rectangle is completely divided.

[0036] Secondly, according to the intercept distribution of the observation ray in the grid, the two squares are iteratively quadrisectioned, and the sum of the intercepts of all observation rays passing through each square grid after quadrisection is calculated. If any of the four grids after quadrisection has an intercept of 0, that is, no observation ray passes through the grid, the grid will no longer be subjected to subsequent smaller-scale quadrisection processing.

[0037] Then, according to the size of the on-site exploration of the working face, the efficiency and accuracy requirements of the reconstruction operation are fully considered, and the minimum scale of the four-division is restricted as the iteration stop condition to form the multi-scale grid division result of the working face.

[0038] In order to verify the influence of the minimum resolution of multi-scale division on the reconstruction performance, three minimum division scales of 6.25m, 3.125m and 1.5625m were set respectively. As shown in Table 1, the number of grids is effectively reduced by performing multi-scale division according to the above rules, especially when the minimum scale is 1.5625m, the number of grids is reduced from 8192 after uniform scale division to 1790, which realizes the effective pruning of the search space for reconstruction of the working face, which lays a good foundation for the subsequent reconstruction operation.

[0039] Table 1. Number of grids at different minimum scales

[0040]

[0041] S2. In the calculation of CT reconstruction of the working face, the core problem is to solve an ill-conditioned matrix equation. After the multi-scale division of the working face reconstruction grid, the ill-conditioned matrix solution problem is transformed into a functional extreme value solution problem, and a multi-scale reconstruction objective function is constructed. Then, the objective function is solved by a dynamic multi-population genetic algorithm with strong global search capability.

[0042] The objective function is defined as follows:

[0043]

[0044] Where A is the total number of rays, B is the number of grids, X = (x 1 ,x 2 ,···,x j ,···,x B ) is the absorption coefficient vector of B grids to be inverted, H i is the electromagnetic wave field strength data collected by the i-th ray, r i is the length of the i-th ray, H 0 is the initial field strength of the launch, d ij is the intercept of the i-th ray at the j-th grid.

[0045] The geological anomaly tomography inversion model can be defined as: find X'∈C, so that f(X')=minf(X); where X'=(x' 1 ,x' 2 ,···,x' j ,···,x' B ) is the absorption coefficient value corresponding to the inversion of each grid when the objective function f(X) reaches the minimum value.

[0046] In this embodiment, a dynamic multi-population genetic algorithm (DMPGA) is used to optimize the reconstruction of the working face, which specifically includes:

[0047] S21, initialize the population, including the number of populations MP = 10, the number of individuals in each population NIND = 20, the number of binary bits of the variable PRECI = 20, and the generation gap GGAP = 0.95; each population independently performs selection, crossover and mutation operations, and each population has different control parameters, such as the crossover probability P c ∈[0.75,0.95], mutation probability P m ∈[0.001,0.05], the maximum number of iterations MAXGEN = 1500, explore different areas of the solution space, and record the elite individuals of all populations.

[0048] S22. The best individuals from various populations are regularly introduced into the target population through the immigration operator to achieve information exchange and enable co-evolution among populations.

[0049] S23, Population elimination and dynamic adjustment: Set a fixed evaluation period, calculate the average fitness of each sub-population, sort by fitness, and directly remove all individuals from several sub-populations with the lowest ranking. When the population size is reduced to a certain threshold, merge all remaining sub-populations and recorded elite individuals into a super population with higher diversity and better gene combination.

[0050] S24. Use smaller crossover probability and mutation probability to conduct a fine search for the super population to further improve the quality and accuracy of the solution, maintain the stability of the population while accelerating convergence until the preset maximum number of iterations is reached. The reconstruction operation is completed and the optimal solution of the working surface reconstruction objective function and the electromagnetic wave absorption coefficient corresponding to the multi-scale grid of the working surface are output.

[0051] Through the above strategy, we can gradually reduce the population size while ensuring population diversity and avoiding premature convergence, and finally form a single population containing the optimal individuals. This strategy helps to improve the search efficiency and optimization performance of the genetic algorithm.

[0052] In order to verify the effectiveness of the method of this embodiment, two groups of comparative experiments were conducted. First, based on the multi-population genetic algorithm (MPGA), the effects of uniform scale and multi-scale grid division on the reconstruction operation of the working surface were verified; second, based on MPGA, the effects of dynamic adjustment of multiple populations (i.e., DMPGA) on the efficiency and accuracy of the reconstruction of the working surface were verified.

[0053] (1) When the grid is divided into uniform scales, the stability of the reconstruction operation of the working surface at different scales is compared (as shown in Table 2), and the corresponding relationship between the final average convergence result and the operation efficiency is compared (as shown in Table 2). Figure 5). It can be seen that when the uniform division scale of the working surface is reduced from 25m to 6.25m, after 1500 generations of reconstruction iterative search evolution, the convergence result shows a significant decrease, from 2.7925dB / m to 0.5210dB / m; and when the uniform division scale of the working surface is further reduced from 6.25m to 1.5625m, the average convergence result increases from 0.5210dB / m to 5.0449dB / m. At the same time, as the grid scale is reduced from 25m to 1.5625m, the maximum change in the search results of repeated reconstructions increases from 0.0953dB / m to 5.6625dB / m, and the average operation time also increases from 17.0324s to 4642.7151s. This shows that as the scale of the reconstructed mesh of the working surface decreases, the stability and efficiency of the reconstruction result gradually decrease. Therefore, under the condition of incomplete projection of the working surface, the reasonable design of the reconstruction mesh scale has an important influence on the performance of the reconstruction operation.

[0054] Table 2 Results of uniform scale grid division and reconstruction of working surface

[0055]

[0056] In addition, by Figure 4 It can be seen that when the grid uniform division scale is reduced to 6.25m, a large number of grids that are not passed by the observed rays and do not carry valid projection information begin to appear. Based on this, when multi-scale grid division is adopted, this embodiment uses 6.25m, 3.125m, and 1.5625m as the minimum grid division scale, performs multiple working surface reconstruction operations, and the corresponding grid division numbers are 446, 1130, and 1790, respectively, and the convergence results and reconstruction efficiency are obtained, as shown in Table 3.

[0057] Table 3 Results of multi-scale meshing and reconstruction of working surface

[0058]

[0059] It can be seen that when the minimum grid scale is 6.25m, 3.125m, and 1.5625m, the average convergence results are reduced from 0.5210dB / m, 0.9265dB / m, and 5.0449dB / m of uniform scale to 0.3479dB / m, 0.8286dB / m, and 1.1425dB / m, respectively; the reconstruction results change from 0.63 The reconstruction time was reduced from 221.4356s, 936.8193s and 4668.2367s to 194.0055s, 493.1762s and 833.2183s respectively. This shows that by effectively integrating the grids that are not passed by the observed rays and effectively pruning the reconstruction search space, the accuracy, stability and efficiency of the reconstruction results are significantly improved.

[0060] (2) Based on MPGA, an interval of 100 generations is set to evaluate all populations, and reconstruction operations are performed according to the multi-population dynamic adjustment strategy. The results are shown in Table 4.

[0061] Table 4 Results of working surface reconstruction based on DMPGA multi-scale meshing

[0062]

[0063] It can be seen that the average convergence value and change amplitude obtained based on the DMPGA multi-scale reconstruction model are not much different from those of the MPGA algorithm, but the calculation time is reduced from 194.0055s, 493.1762s and 833.2183s to 85.3674s, 219.434s and 351.7312s, respectively, and its reconstruction efficiency is further greatly improved.

[0064] Therefore, the present invention adopts the above-mentioned coal mine working face reconstruction method based on search space pruning and dynamic adjustment of multiple populations, which can finely match the exploration sizes of different working faces and the distribution characteristics of different density rays, realize multi-scale grid division, and improve the accuracy, stability and efficiency of working face reconstruction.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A coal mine working face reconstruction method based on search space pruning and multi-population dynamic adjustment, characterized in that: The following steps are involved: S1. Perform preliminary grid division on the exploration area of ​​the working face, and perform multi-scale grid division by search space pruning based on the sum of ray intercepts in each network; S2. Based on the divided multi-scale grids, a multi-scale reconstruction objective function is constructed, and the population is eliminated and dynamically adjusted through a dynamic multi-population genetic algorithm to obtain a super population. Then, a refined search is performed on the super population, and the multi-scale reconstruction objective function is iteratively solved to obtain the electromagnetic wave absorption coefficient corresponding to the multi-scale grid.

2. The method for coal mine working face reconstruction based on search space pruning and multi-population dynamic adjustment according to claim 1 is characterized in that: Step S1 includes: first, preliminarily dividing the exploration area of ​​the working face into square grids according to the size of the working face; second, dividing each preliminarily divided square grid into four parts; then, determining the grids that need to be divided again according to the ray intercepts in the grids after the four-division, and iterating in sequence until the minimum scale limit is reached.

3. The method for coal mine working face reconstruction based on search space pruning and multi-population dynamic adjustment according to claim 1 is characterized in that: In step S2, the multi-scale reconstruction objective function is expressed as follows: Where A is the total number of rays, B is the number of grids, X = (x1, x2, ···, x j ,···,x B ) is the absorption coefficient vector of B grids to be inverted, H i is the electromagnetic wave field strength data collected by the i-th ray, r i is the length of the i-th ray, H0 is the initial field strength of the emission, d ij is the intercept of the i-th ray at the j-th grid.

4. The method for coal mine working face reconstruction based on search space pruning and multi-population dynamic adjustment according to claim 1 is characterized in that: Through dynamic multi-population genetic algorithm, including: S21, initialize the population, each population independently performs selection, crossover and mutation operations, explores different areas of the solution space, and records the elite individuals of all populations; S22, regularly introduce the best individuals from various populations into the target population through the immigration operator, so that the populations can co-evolve; S23, set a fixed evaluation period, calculate the average fitness of each sub-population, screen the sub-populations according to the fitness ranking, and merge the screened sub-populations and their corresponding elite individuals to obtain a super population; S24. Use smaller crossover probability and mutation probability to conduct a detailed search on the super population until the preset maximum number of iterations is reached.

5. The method for coal mine working face reconstruction based on search space pruning and multi-population dynamic adjustment according to claim 4 is characterized in that: Different control parameters are used for each population.

Images