Mine blasting excavation lumpiness grading control method and device and computer equipment

Abstract

The invention discloses a mine blasting excavation lumpiness grading control method and device and computer equipment, and relates to the field of blasting control. The method comprises the following steps: acquiring rock mass parameters, blasting design parameters and blasting lumpiness data, and constructing a structured database; initializing a hyper-parameter of the XGBoost model, and iteratively optimizing the hyper-parameter by using a hybrid optimization algorithm based on an adaptive particle swarm and improved bacterial foraging to obtain a target parameter; and a mine blasting excavation lumpiness grading control model is constructed according to the target parameters, the mine blasting excavation lumpiness grading control model is trained through the structured database, and a blasting control decision under the target working condition is made through the trained mine blasting excavation lumpiness grading control model. The method is suitable for controlling the mine blasting excavation lumpiness under the influence of various complex conditions, and the problems that traditional mine blasting excavation excessively depends on experience of operators, so that the discreteness of lumpiness grading is large, and the construction efficiency is low are solved.

Description

Technical Field

[0001] This application relates to the field of blasting control technology, and in particular to a method, device and computer equipment for controlling the block size distribution of mine blasting excavation. Background Technology

[0002] Mining blasting is a crucial step in mining operations, and a well-designed blasting system can effectively improve mining efficiency and reduce costs. During blasting, controlling the size of the blasted rock blocks is paramount, as the size distribution of the blasted rock directly impacts subsequent crushing and transportation operations.

[0003] Traditional methods for predicting blasting block size cannot fully consider various factors affecting blasting block size, such as the physical and mechanical properties of the rock mass, blasting design parameters, and on-site geological conditions, resulting in insufficient prediction accuracy. Furthermore, existing intelligent prediction methods suffer from slow convergence speeds and a tendency to get trapped in local optima, making it difficult to meet the accuracy requirements of mine blasting under complex conditions.

[0004] Therefore, how to propose a method for controlling the block size distribution of mine blasting excavation that is applicable to complex geological conditions has become a problem that urgently needs to be solved in current mine blasting technology. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, device, and computer equipment for controlling the block size distribution of mine blasting excavation, in order to address the above-mentioned technical problems, thereby improving the blasting effect and mining efficiency by optimizing the blasting design parameters.

[0006] Firstly, this application provides a method for controlling the block size gradation of blasting excavation in mines. The method includes:

[0007] Obtain rock mass parameters, blasting design parameters, and blasting block size data to construct a structured database;

[0008] The hyperparameters of the XGBoost model are initialized, and the hyperparameters are iteratively optimized using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain the target parameters.

[0009] A block size distribution control model for mine blasting excavation is constructed based on the target parameters. The block size distribution control model for mine blasting excavation is trained using a structured database. The trained block size distribution control model for mine blasting excavation is then used to formulate blasting control decisions under the target working conditions.

[0010] In one embodiment, the rock mass parameters include the rock's elastic modulus, longitudinal wave velocity, shear strength, and unit weight.

[0011] Blasting design parameters include row spacing, hole spacing, and explosive consumption per unit.

[0012] The blasting block size data includes the size of the crushed stone and the proportion of gradation.

[0013] In one embodiment, the hyperparameters are iteratively optimized using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain the target parameters, including:

[0014] The fitness value of the current population particles is obtained according to the fitness function. Based on the fitness value, the individual extreme value and the global extreme value are determined, and the first particle position corresponding to the individual extreme value and the second particle position corresponding to the global extreme value are obtained; where the fitness function is the coefficient of determination.

[0015] Update the velocity and position of the current population particles based on the position of the first particle and the position of the second particle, update the fitness value of the current population particles based on the velocity and position of the current population particles, update the individual extreme value and global extreme value, as well as the corresponding position of the first particle and the position of the second particle based on the updated fitness value;

[0016] Compare the global extrema before and after the update. If they are the same, perform a migration operation based on the updated fitness value. If they are different, perform a tendency operation iteratively based on the updated second particle position, and perform a migration operation after the iteration stops.

[0017] The population particle after the migration operation is obtained as the current population particle, and the next iteration operation is performed.

[0018] Obtain the target parameters after the iteration stops.

[0019] In one embodiment, updating the velocity and position of the current population particles based on the positions of the first and second particles includes:

[0020] The inertia weight is obtained based on the fitness value of the current population particles, the average fitness value, and the minimum fitness value.

[0021] Update the velocity of the current population particles based on the inertia weight, the position of the first particle, and the position of the second particle;

[0022] Update the position of the current population particles based on their current velocity.

[0023] In one embodiment, the migration operation includes:

[0024] Obtain the migration probability, historical minimum fitness value, and historical maximum fitness value from the previous iteration, and combine them with the fitness value of the current population particles to obtain the current migration probability.

[0025] Based on the current migration probability, at least one particle is randomly selected from the current population particles to initialize its position, thereby updating the population particles.

[0026] In one embodiment, hyperparameters include the maximum depth of the tree, learning rate, random seed, minimum child node weight, number of trees, minimum split gain, subsample ratio, column sampling ratio per tree, and L1 regularization term.

[0027] Secondly, this application also provides a device for controlling the block size gradation of blasting excavation in mines. The device includes:

[0028] The data acquisition module is used to acquire rock mass parameters, blasting design parameters, and blasting block size data to build a structured database;

[0029] The parameter optimization module is used to initialize the hyperparameters of the XGBoost model. It iteratively optimizes the hyperparameters using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain the target parameters.

[0030] The model building module is used to construct a control model for the block size distribution of mine blasting excavation based on target parameters, train the control model for the block size distribution of mine blasting excavation using a structured database, and use the trained control model for the block size distribution of mine blasting excavation to make blasting control decisions under target working conditions.

[0031] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method for controlling the block size distribution of blasting excavation in mines.

[0032] Fourthly, this application also provides a computer-readable storage medium. This computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps in the above-described method for controlling the block size distribution of mine blasting excavation.

[0033] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps in the above-described method for controlling the block size distribution of blasting excavation in mines.

[0034] The aforementioned method, apparatus, and computer equipment for controlling the block size gradation of mine blasting excavation include the following steps: acquiring rock mass parameters, blasting design parameters, and blasting block size data to construct a structured database; initializing the hyperparameters of the XGBoost model, iteratively optimizing the hyperparameters using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain target parameters; constructing a mine blasting excavation block size gradation control model based on the target parameters; training the mine blasting excavation block size gradation control model using the structured database; and using the trained mine blasting excavation block size gradation control model to formulate blasting control decisions under target working conditions. This technical solution constructs a structured database based on on-site geological conditions, blasting block size gradation, and blasting design parameters. Combined with parameter optimization of the prediction model, a blasting block size gradation control model is obtained, which can determine the target values ​​of blasting design parameters for a specific project based on the average values ​​of the upper and lower envelopes of the blasting block size. This method integrates on-site geological conditions, blasting design parameters, and on-site blasting block size gradation results into a mine blasting excavation block size gradation control model. It boasts high accuracy, strong adaptability, and can directly guide engineering practice. Furthermore, the model parameters can be adjusted according to the actual geological conditions of the quarry and the blasting block size gradation control requirements, enabling it to accurately predict the blasting design parameters required for different types of quarries. The prediction model exhibits strong engineering applicability. Therefore, this method is highly targeted, widely adaptable, highly accurate, reasonable, feasible, and economically efficient. Attached Figure Description

[0035] Figure 1 This is a flowchart of a method for controlling the block size distribution in mine blasting excavation, as described in one embodiment.

[0036] Figure 2 Here is a flowchart of a hybrid optimization algorithm in one embodiment;

[0037] Figure 3 A flowchart for constructing a block size distribution control model for mine blasting excavation in one embodiment;

[0038] Figure 4 To optimize the blasting block size distribution result in one embodiment;

[0039] Figure 5 This is a structural block diagram of a mine blasting excavation block size gradation control device in one embodiment. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0041] This application provides a method for controlling the block size distribution in mine blasting excavation, such as... Figure 1 As shown, it includes the following steps:

[0042] Step 101: Obtain rock mass parameters, blasting design parameters, and blasting block size data to construct a structured database.

[0043] Through on-site blasting tests, data on on-site rock mass parameters, blasting design parameters, and blasting block size were collected.

[0044] Specifically, the explosive is first selected based on impedance matching. This involves conducting detonation velocity tests on the explosive at the engineering site to obtain the longitudinal wave velocity of the rock mass, and then selecting the explosive whose wave impedance is closest to that of the rock mass. The rock mass wave impedance is the product of the rock mass density and the longitudinal wave velocity.

[0045] Next, by consulting the geological exploration materials, the elastic modulus, longitudinal wave velocity, shear strength f', shear strength C', and unit weight of the rock at the site were determined and recorded to obtain the rock mass parameters.

[0046] Based on rock mass parameters, explosive properties, and the control requirements of blasting block size distribution curve, and referring to the results of field explosion tests, the blasting design parameters, including row spacing, hole spacing, and explosive consumption per unit, are determined and recorded based on experience.

[0047] In-situ rock blasting tests were conducted to screen the blasted block sizes. After the rock blasting tests were conducted, the gradation results of the blasted block size data for the project were obtained, which centrally covered the gradation proportions of block sizes less than 5mm, 10mm, 20mm, 50mm, 100mm, 200mm and 400mm after blasting. The blasted block size gradation curve for the project was then plotted.

[0048] To eliminate the influence of different dimensions among the obtained rock mass parameters, blasting design parameters, and blasting block size data, a structured database is formed. This embodiment employs a normalization method to process the rock mass parameters, blasting design parameters, and blasting block size data. The formula is expressed as:

[0049]

[0050] Among them, Z normalize For the normalized data, x i For the i-th data, s is the mean of the data and s is the standard deviation.

[0051] The normalized data is organized into a structured database, which can then be used as training samples for subsequent model training.

[0052] Step 102: Initialize the hyperparameters of the XGBoost model, and iteratively optimize the hyperparameters using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain the target parameters.

[0053] like Figure 2 As shown, the specific steps of the hybrid optimization algorithm include:

[0054] Step 1021: Construct a preliminary XGBoost model. Using a structured database, determine the feature variables (input parameters) and target variables (output parameters) of the XGBoost model. Set the initial hyperparameter values ​​and search range for the XGBoost model, forming a hyperparameter space as the search object for the improved bacterial foraging algorithm. The hyperparameters include: maximum tree depth, learning rate, random seed, minimum child node weight, number of trees, minimum split gain, subsample ratio, column sampling ratio for each tree, and L1 regularization term.

[0055] In this embodiment, the XGBoost model uses the XGBoost Regressor model (XGBRegressor) to solve the problem of predicting block size gradation.

[0056] Step 1022: Initialize particle swarm parameters.

[0057] Based on the optimization requirements of the hyperparameters, an initial particle swarm is generated. Each particle represents a potential combination of hyperparameters. The initial swarm consists of L particles, where the high-dimensional position parameter of each particle represents a set of hyperparameter combination information, and the velocity is represented by the particle's movement speed in the corresponding dimensional search space. The particle position x and velocity v are used to represent the current value and change of the hyperparameters, and the initial particle swarm is randomly initialized.

[0058] Step 1023, initialize algorithm parameters, specifically including:

[0059] 1. Set the algorithm parameters for IBFO (Improved Bacterial Foraging Algorithm) and APSO (Adaptive Particle Swarm Optimization Algorithm). The IBFO algorithm parameters include: maximum number of iterations N for the tendency operation. b Trend operation step size C step Migration probability P ed The APSO algorithm parameters include: maximum number of particle swarm iterations k. m The fitness threshold σ of the global optimal solution, the maximum inertia weight value, the minimum inertia weight value, and the learning factor.

[0060] 2. Set the fitness function. In this embodiment, the coefficient of determination R is... 2 As a fitness function, used to measure the error between the XGBoost model's predicted values ​​and the actual blast block size gradation data, the fitness function formula is expressed as:

[0061]

[0062] In the formula, y i This represents the actual value of the i-th training sample in the structured database. This represents the fitted value of the i-th training sample. This represents the average value of all training samples, where n is the number of samples.

[0063] The closer the fitness function value is to 1, the better the prediction performance of the XGBoost model under the corresponding parameter combination, thus guiding the hybrid optimization algorithm to search for the optimal parameters.

[0064] Step 1024: Find the individual and global extrema. Specifically, this includes:

[0065] 1. Calculate the fitness value of each particle in the current population, and select the fitness value and position of an individual particle as the individual extreme value F corresponding to that particle. pi and position B pi B pi Consider it as the position of the first particle.

[0066] 2. Find the global extremum F of the current population particles based on their fitness values. g and corresponding position B g B g Consider it as the position of the second particle.

[0067] 3. The particle with the fitness value closest to 1 is taken as the current optimal fitness solution.

[0068] Step 1025, Particle Update. Specifically, this includes:

[0069] 1. Update the velocity v and position x of the population particles according to the following formulas:

[0070] v i (t+1)=w·v i (t)+c1·rand1·(B pi -x i (t))+c2·rand2·(B g -x i (t))

[0071] x i (t+1)=x i (t)+v i (t+1)

[0072] Where c1 and c2 are learning factors, r1 and r2 are random numbers in the range of 0 to 1, and ω is the inertia weight. The expression for the inertia weight is as follows:

[0073]

[0074] Where, ω min and ω max These are the minimum and maximum inertia weight values, f, respectively. if is the fitness value of the current particle i. ave f is the average fitness value min This represents the minimum fitness value. During the optimization process, the adaptive particle swarm optimization algorithm automatically updates the inertia weight value based on the current fitness value to optimize the target search space capability.

[0075] 2. Calculate the fitness value of the current particle. If the new fitness value is better, update the individual extreme value F. pi and global extremum F g and particle position B pi and B g .

[0076] Step 1026: Compare the global extrema before and after the update. If the global extrema have changed, proceed to step 1027 to perform the trend operation; otherwise, proceed to step 1028 to perform the migration operation.

[0077] Step 1027: Perform a orientation operation on the particles, updating their positions by simulating bacterial foraging behavior. Randomly flip the particles according to the following formula:

[0078] x i (t+1)=x i (t)+C step ·sign(B g -x i (t))

[0079] Among them, B g C represents the position corresponding to the global extremum of all particles, i.e., the position of the second particle. step To increase the stride length of swimming.

[0080] Continue iterating the particle orientation operation until the maximum number of iterations N for the orientation operation is reached. b After the iteration stops, proceed to step 1028.

[0081] Step 1028: Sort all particles by their fitness values, select the particles with higher fitness values, and then determine their migration probability. Particles with higher fitness are randomly moved into the optimization space. By simulating the particle migration process, the algorithm avoids getting trapped in local optima, maintains population diversity, and enhances the algorithm's ability to escape local optima. The migration probability is calculated using the following formula:

[0082]

[0083] in, This represents the probability of migration at the current moment. f represents the migration probability at a previous time step. min f represents the lowest fitness value in history. maxf represents the highest fitness value in history. i This represents the current fitness value of a specific particle i.

[0084] Step 1029: Increase the number of iterations and return to step 1024 to continue the population particle update process. The parameters to be optimized in XGBoost are used as the high-dimensional position vectors of particles in the hybrid optimization algorithm, and the particle velocity reflects the adjustment trend of the parameter combination. In each iteration, the individual optimal and globally optimal particle positions are updated until the maximum number of iterations or the fitness value convergence threshold is reached. At this point, the parameter combination corresponding to the globally optimal particle is the optimal core parameter value of the XGBoost model.

[0085] When the set maximum number of particle swarm iterations k is reached m Once the fitness value convergence threshold is reached, the algorithm ends and selects the particle with the best fitness in the current particle swarm. The solution corresponding to this particle is the optimal parameter of the prediction model for the project, i.e., the target parameter.

[0086] Step 103: Construct a block size distribution control model for mine blasting excavation based on the target parameters, train the block size distribution control model for mine blasting excavation using a structured database, and use the trained block size distribution control model for mine blasting excavation to formulate blasting control decisions under the target working conditions.

[0087] Figure 3 This diagram illustrates the process of constructing a block size gradation control model for mine blasting excavation based on target parameters. The target parameters determined in step 102 are applied to the XGBoost model to construct the block size gradation control model for mine blasting excavation. The block size gradation control model is then used to determine the blasting design parameters for the new target working condition. Based on these parameters, blasting excavation is carried out on the rock mass under the new target working condition to ensure that the post-blast rock block size gradation curve is within the required range.

[0088] Specifically, the target working condition requires determining the target particle size range and excavation volume of the blasted boulders based on the specific mine excavation needs. For example, the control requirements for the blasting block size distribution curve for the new working condition are: the proportion of blocks smaller than 5mm, 10mm, 20mm, 50mm, 100mm, 200mm, and 400mm after this blast must be within a given range.

[0089] The input dataset for the mine blasting excavation block size gradation control model under the target working condition includes: the rock's elastic modulus, longitudinal wave velocity, shear strength, and unit weight under the new working condition, as well as the gradation ratio control ranges for block sizes less than 5mm, 10mm, 20mm, 50mm, 100mm, 200mm, and 400mm after blasting under the target working condition. The input dataset for the mine blasting excavation block size gradation control model under the target working condition consists of blasting design parameter values ​​that meet the gradation control requirements for this working condition, including row spacing, hole spacing, and explosive consumption per unit volume.

[0090] It is important to note that the blasting design parameters that need to be determined on-site include the charge structure and the blasting hole network parameters. Considering the on-site construction conditions, apart from the blasting design parameters calculated in step 103, the blasting charge structure, hole diameter, and explosive density are all taken as uniform parameters, and therefore their influence does not need to be considered.

[0091] For a specific mountain in a mine to be excavated, the excavation height is 220m, divided into 15 layers, each 15m high. The rock type is limestone. A uniform decoupled continuous charge will be used, with a uniform aperture of 115mm and a uniform explosive density of 1300kg / m³. In this scenario, if... Figure 3 As shown, the process flow for controlling the block size gradation of mine blasting excavation is as follows:

[0092] Step 1: Based on the basic excavation needs of the mine, determine the requirements for the upper and lower aggregates to be mined, as shown in Table 1.

[0093] Table 1 Requirements for Mining Block Size Gradation

[0094] Grading Boundaries 600 400 200 100 80 60 40 20 10 5 Upper limit of grade 100 100 82 65 60 53 45 32 23 15 Lower limit of gradation 100 80 57 42 40 35 30 23 17 10

[0095] Step 2: Select explosives and conduct on-site blasting tests

[0096] 2.1 Selection of Explosives Based on Impedance Matching

[0097] Based on the three types of explosives available from the local explosives company, with densities of 1050 kg / m³, 1150 kg / m³, and 1300 kg / m³, and detonation velocities of 3300 m / s, 3600 m / s, and 4200 m / s, respectively, and the longitudinal wave velocity of the rock mass being approximately 3900 m / s and its density approximately 2200 kg / m³, the wave impedance of the rock is approximately 2200 × 3900 = 8.58 × 10⁶. The wave impedances of the three explosives are 3.47 × 10⁶, 4.14 × 10⁶, and 5.46 × 10⁶, respectively. According to the impedance matching principle, the explosive closest to the wave impedance of the rock mass, i.e., the third explosive, is selected.

[0098] 2.2 Determine rock mass parameters using engineering geological exploration materials.

[0099] By reviewing the engineering geological exploration data of different blasting zones on site and conducting mechanical property tests on rock samples from different blasting zones, the elastic modulus, longitudinal wave velocity, shear strength f', shear strength C', and unit weight of the rocks in different blasting zones were determined, as detailed in Table 2.

[0100] 2.3 Determine the blasting design parameters.

[0101] Based on the rock mass parameters, explosive properties, and blasting block size control requirements, and referring to the field test blasting results, the blasting design parameters for different blasting zones are determined, including row spacing, hole spacing, and explosive consumption per unit area, as detailed in Table 2.

[0102] Table 2 Basic Information on the Demolition Zone

[0103]

[0104] 2.4 Field blasting tests were conducted using the selected explosives and determined blasting design parameters.

[0105] In-situ rock blasting tests were conducted to screen the blasted block sizes. After the rock blasting tests were conducted, the gradation results of 10 sets of blasted block size data for the project were obtained, which centrally covered the gradation ratio of block sizes less than 5mm, 10mm, 20mm, 50mm, 100mm, 200mm and 400mm after blasting. See Table 3 for details.

[0106] Table 3 Results of blasting material screening proportions

[0107]

[0108] Step 3: Establish a data processing model and build a structured database.

[0109] A normalization algorithm was used to eliminate the dimensional influence among rock mass parameters, blasting design parameters, and blasting block size parameters in Tables 2 and 3, forming a structured database for training the blasting block size control model. The input parameters of the control model are: rock elastic modulus, rock longitudinal wave velocity, rock shear strength f', rock shear strength C', unit weight, and the proportion of blasted block sizes less than 5mm, 10mm, 20mm, 50mm, 100mm, 200mm, and 400mm, respectively. The output parameters of the control model are: row spacing, hole spacing, and explosive consumption per unit volume.

[0110] Step 4: Construct a block size gradation control model for blasting excavation

[0111] Using the structured database in Table 4, an XGBoost prediction model was trained. The IBFO-APSO hybrid optimization algorithm was then used to train the XGBoost model. The optimal core parameters of the XGBoost prediction model under this working condition were calculated. A block size gradation control model for mine blasting excavation based on the IBFO-APSO-XGBoost model was constructed, specifically as follows:

[0112] 4.1 Based on the structured database in Table 4, determine the input and output parameters for training the XGBoost model, determine the XGBRegressor regression model as the XGBoost model type, and initialize the XGBoost model hyperparameters, where the maximum tree depth is 10, the learning rate is 0.02, the random seed is 0, the minimum child node weight is 1, the number of trees is 400, the minimum split gain is 0.05, the subsample ratio is 0.8, the column sampling ratio of each tree is 0.7, and the L1 regularization term is 1.

[0113] 4.2 Constructing the IBFO-APSO Hybrid Optimization Algorithm

[0114] 4.2.1 Determining the search range of hyperparameters for the XGBoost model

[0115] The maximum tree depth ranges from 3 to 15, the learning rate ranges from 0.01 to 0.1, the random seed ranges from 0 to 100, the minimum child node weight ranges from 1 to 10, the number of trees ranges from 100 to 1000, the minimum split gain ranges from 0.01 to 0.1, the subsample ratio ranges from 0.6 to 0.9, the column sampling ratio for each tree ranges from 0.5 to 0.9, and the L1 regularization term ranges from 0 to 10.

[0116] 4.2.2 Initialize Algorithm Parameters

[0117] In the hybrid optimization algorithm, the number of particle optimizations is 5. In each round of optimization, the fitness of the current particle is calculated and the particle's velocity and position are updated. The number of APSO particles L is set to 5. The high-dimensional position of each particle represents a combination of hyperparameters, and its initial parameters are shown in Table 4.

[0118] Table 4 Initial parameters of each particle

[0119]

[0120]

[0121] In the IBFO algorithm, the maximum number of iterations N b The value is 100, and the step size C is approaching the operation. step The migration probability P is 0.1. ed The maximum number of iterations N for the approach operation is 0.5; in the APSO algorithm, the maximum number of iterations N for the approach operation is 0.5. bThe fitness threshold σ for the global optimal solution is 0.95, and the maximum number of iterations k in the particle swarm optimization is 50. m ω is 100 max It is 0.9, ω min The initial particle velocity v is determined as a vector (0.003, 0.5, 3.33, 0.3, 30, 0.003, 0.01, 0.01, 0.33) based on the hyperparameter search range of the XGBoost model.

[0122] 4.2.3 Finding the current individual particle and the global extremum

[0123] The current fitness value of each particle is calculated according to the fitness function. The fitness value of particle 1 is 0.85, the fitness value of particle 2 is 0.87, the fitness value of particle 3 is 0.83, the fitness value of particle 4 is 0.89, and the fitness value of particle 5 is 0.86. The current initial global fitness optimum is recorded as the fitness value of particle 4, which is 0.89.

[0124] 4.2.4 Particle Motion Velocity and Position Update

[0125] Taking particle 1 as an example, to calculate the particle's velocity and position, firstly, according to the inertia weight calculation formula, the current inertia weight is ω1 = ω max =0.9, and update the particle's velocity:

[0126] v1(2)=w·v1(1)+c1·rand1·(B pi -x1(1))+c2·rand2·(B g -x1(1))

[0127] = 0.9·(0.003,...,0.33)+1.5·0.8·((0.05,...,0.8)-(0.02,...,1))+1.5·0.7·((0.05,...,0.8)-(0.02,...,1))

[0128] =(0.07,-4.05,53.99,0.27,139.5,-0.019,-0.216,0.122,0.847)

[0129] Update the particle positions again:

[0130] x i (2)=x1(1)+v1(2)=(0.02,...,1)+(0.07,...,0.847)

[0131] =(0.09,5.95,53.99,1.27,539,0.03,0.58,0.82,1.85)

[0132] At this point, the fitness value of particle 1 is updated to 0.87 according to the fitness function. Since the current fitness of particle 1 is lower than the current global extreme value of 0.89, the global extreme value is not updated.

[0133] 4.2.5 Perform particle orientation operation

[0134] Since the global extremum remains unchanged, there is no need to perform a tendency operation on particle 1, and we can proceed directly to step 4.2.6.

[0135] 4.2.6 Perform particle migration operation

[0136] Calculate the migration probability of particle 1:

[0137]

[0138] This indicates that particle 1 has a high fitness at this point. In order to avoid local optima, promote global search, and increase the diversity of solutions, a migration operation is performed on particle 1.

[0139] After updating the positions of all particles in the particle swarm according to the above operations, proceed to step 4.2.3 to continue the optimization process of the algorithm.

[0140] 4.2.7 After reaching the set number of particle iterations, bacterial chemotaxis, and migrations, the IBFO-APSO hybrid optimization algorithm ends and selects the particle with the highest fitness value among the current particles. Its position is (0.02, 11, 7, 1, 453, 0.06, 0.82, 0.65, 1.2), which is the optimal parameter corresponding to the prediction model of this project: learning rate of 0.02, maximum tree depth of 11, random seed of 7, minimum child node weight of 1, number of trees of 453, minimum split gain of 0.06, subsample ratio of 0.82, column sampling ratio of each tree of 0.65, and L1 regularization term of 1.2.

[0141] 4.3 Constructing the IBFO-APSO-XGBoost Prediction Model

[0142] An XGBoost prediction model was applied to construct a block size gradation control model for mine blasting excavation based on the IBFO-APSO hybrid algorithm to optimize the XGBoost model.

[0143] Step 5: Use the blasting block size control model to determine the blasting design parameters under the new working conditions.

[0144] 5.1 Based on the rock mass parameters and blasting block size control requirements under the new working conditions, the input data for the block size control model under this working condition is constructed, as shown in Table 5.

[0145] Table 5 Input data for the gradation curve control model under new operating conditions

[0146]

[0147] 5.2 Using the mine blasting excavation block size gradation control model based on the IBFO-APSO-XGBoost model in step 4, and combined with the input data of the gradation curve control model under the new working condition in Table 5, the blasting design parameters that meet the block size control requirements of this working condition are calculated: row spacing is 3.9, hole spacing is 4.1, and explosive consumption is 92.

[0148] Step 6: Using the blasting design parameters calculated in section 5.2, blast and excavate the rock mass under the new working conditions. The post-blast rock block size distribution curve is shown in [reference needed]. Figure 4 This enables pre-blast control of the block size distribution in mine blasting excavation.

[0149] Through the above steps, this application provides a method for controlling the block size distribution of mine blasting excavation based on the IBFO-APSO-XGBoost model. It deeply integrates technologies such as on-site monitoring and machine learning, realizes precise control of blasting block size, improves the efficiency and safety of mine excavation, and has important practical application value.

[0150] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0151] Based on the same inventive concept, this application also provides a mine blasting excavation block size gradation control device for implementing the above-mentioned mine blasting excavation block size gradation control method. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more mine blasting excavation block size gradation control device embodiments provided below can be found in the limitations of the mine blasting excavation block size gradation control method above, and will not be repeated here.

[0152] In one embodiment, such as Figure 5 As shown, a device for controlling the block size distribution in mine blasting excavation is provided, comprising:

[0153] The data acquisition module is used to acquire rock mass parameters, blasting design parameters, and blasting block size data to build a structured database;

[0154] The parameter optimization module is used to initialize the hyperparameters of the XGBoost model. It iteratively optimizes the hyperparameters using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain the target parameters.

[0155] The model building module is used to construct a control model for the block size distribution of mine blasting excavation based on target parameters, train the control model for the block size distribution of mine blasting excavation using a structured database, and use the trained control model for the block size distribution of mine blasting excavation to make blasting control decisions under target working conditions.

[0156] Each module in the aforementioned mine blasting and excavation block size gradation control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0157] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in all of the above method embodiments.

[0158] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in all of the above method embodiments.

[0159] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in all of the above method embodiments.

[0160] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0161] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0162] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0163] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for controlling the block size distribution in mine blasting excavation, characterized in that, The method includes: Obtain rock mass parameters, blasting design parameters, and blasting block size data to construct a structured database; The hyperparameters of the XGBoost model are initialized, and the hyperparameters are iteratively optimized using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain the target parameters. A block size distribution control model for mine blasting excavation is constructed based on the target parameters. The block size distribution control model for mine blasting excavation is trained using the structured database. The trained block size distribution control model for mine blasting excavation is then used to formulate blasting control decisions under the target working conditions.

2. The method according to claim 1, characterized in that: The rock mass parameters include the rock's elastic modulus, longitudinal wave velocity, shear strength, and unit weight. The blasting design parameters include row spacing, hole spacing, and explosive consumption per unit; The blasting block size data includes the size of the crushed stone blocks and the gradation ratio.

3. The method according to claim 1, characterized in that, The hyperparameters are iteratively optimized using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain the target parameters, including: The fitness value of the current population particles is obtained according to the fitness function. Based on the fitness value, the individual extreme value and the global extreme value are determined, and the first particle position corresponding to the individual extreme value and the second particle position corresponding to the global extreme value are obtained; wherein, the fitness function is the coefficient of determination. Update the velocity and position of the current population particles based on the first particle position and the second particle position, update the fitness value of the current population particles based on the velocity and position of the current population particles, update the individual extreme value and the global extreme value, and the corresponding first particle position and second particle position based on the updated fitness value; Compare the global extrema before and after the update. If they are the same, perform a migration operation based on the updated fitness value. If they are different, perform a tendency operation iteratively based on the updated second particle position, and perform a migration operation after the iteration stops. The population particle after the migration operation is obtained as the current population particle, and the next iteration operation is performed. The target parameters are obtained after the iteration stops.

4. The method according to claim 3, characterized in that, The step of updating the velocity and position of the current population particles based on the first particle position and the second particle position includes: The inertia weight is obtained based on the fitness value of the current population particles, the average fitness value, and the minimum fitness value. The velocity of the current population particles is updated based on the inertia weight, the position of the first particle, and the position of the second particle. Update the position of the current population particles based on their current velocity.

5. The method according to claim 3, characterized in that, The migration operation includes: Obtain the migration probability, historical minimum fitness value, and historical maximum fitness value from the previous iteration, and combine them with the fitness value of the current population particles to obtain the current migration probability. Based on the current migration probability, at least one particle is randomly selected from the current population particles to initialize its position, thereby updating the population particles.

6. The method according to claim 1, characterized in that: The hyperparameters include the maximum tree depth, learning rate, random seed, minimum child node weight, number of trees, minimum split gain, subsample ratio, column sampling ratio per tree, and L1 regularization term.

7. A device for controlling the block size distribution in mine blasting excavation, characterized in that, The device includes: The data acquisition module is used to acquire rock mass parameters, blasting design parameters, and blasting block size data to build a structured database; The parameter optimization module is used to initialize the hyperparameters of the XGBoost model and iteratively optimize the hyperparameters using a hybrid optimization algorithm based on adaptive particle swarm optimization and improved bacterial foraging to obtain the target parameters. The model building module is used to construct a control model for the block size distribution of mine blasting excavation based on the target parameters, train the control model for the block size distribution of mine blasting excavation using the structured database, and formulate blasting control decisions under the target working conditions using the trained control model for the block size distribution of mine blasting excavation.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images