Blasting vibration speed prediction method and system based on feature selection and GA-XGBoost model, computer equipment and storage medium

Abstract

A blasting vibration speed prediction method based on feature selection and a GA-XGBoost model comprises the following steps: collecting a plurality of basic features in a surface mine blasting operation process, and extracting a part of the basic features as target features through Pearson correlation analysis; an XGBoost model is initialized, optimization training is carried out on the XGBoost model through GA and a data set constructed based on target features, and a blasting vibration speed prediction model is obtained; and acquiring dynamic blasting parameters in real time based on the target characteristics, and inputting the dynamic blasting parameters into the blasting vibration speed prediction model for prediction to obtain a predicted speed. The invention provides a blasting vibration speed prediction method and system based on feature selection and a GA-XGBoost model, computer equipment and a storage medium, and can accurately predict the blasting vibration speed in the blasting process of a strip mine area, thereby guiding the blasting process, and realizing the effects of optimizing the blasting process and ensuring the blasting efficiency and field safety.

Description

Technical Field

[0001] The present invention relates to the technical field of blasting vibration velocity prediction, and in particular to a blasting vibration velocity prediction method, system, computer equipment and storage medium based on feature selection and GA-XGBoost model. Background Art

[0002] Blasting operations in open-pit mines inevitably cause ground vibration. Excessive vibration can cause equipment failure and threaten personnel safety. Therefore, blasting vibration peak velocity is a core indicator for assessing blasting safety, and its prediction accuracy directly impacts mining project safety and environmental impact control.

[0003] In the existing technology, the blasting vibration velocity is mainly predicted by the empirical formula method. Several common empirical formulas are shown in Table 1.

[0004] Table 1: Empirical formulas in the prior art

[0005]

[0006] In Table 1, V is the blasting vibration velocity; Q is the maximum single-stage explosive quantity; R is the distance from the blast center; K is the site coefficient; α is the attenuation coefficient; β is the empirical attenuation coefficient, H is the elevation difference; γ is the elevation magnification coefficient, a and b are dimensional analysis indices; and f is a geological parameter.

[0007] The essence of these traditional empirical formulas is to establish a simplified mathematical relationship between charge amount, distance from the explosion center, and vibration velocity through linear regression. This modeling method based on the idea of ​​statistical averaging is difficult to accurately depict the impact of nonlinear characteristics such as rock anisotropy and joint and fissure distribution on vibration wave propagation under complex geological conditions. Especially in special geological structural areas such as layered rock masses and fault fracture zones, there are often systematic deviations between measured data and theoretical predictions. Moreover, although numerical simulation methods can construct three-dimensional geological models through finite element and discrete element algorithms, their reliability is limited by the idealized treatment of rock mechanical parameters. Existing studies generally adopt the homogeneous medium assumption, simplifying key parameters such as Young's modulus and Poisson's ratio into isotropic constants, ignoring the dynamic response characteristics of actual rock masses under the coupling of multiple fields such as groundwater flow field and geostress field.

[0008] In recent years, machine learning models represented by algorithms such as support vector machines (SVM) and random forests (RF) have demonstrated advantages in nonlinear fitting in the prediction of blasting vibration velocity. Compared with traditional empirical formulas, they can comprehensively consider factors such as the geological conditions of the blasting site and have better accuracy. However, existing research still has obvious deficiencies in feature engineering and model optimization, mainly in two aspects. First, in terms of feature selection, most studies simply stack more than ten feature parameters such as hole depth, unit consumption, blast center distance and elevation difference, without considering the multicollinearity problem between features, resulting in the risk of overfitting of the model. Second, in terms of hyperparameter optimization, researchers mostly use traditional methods such as grid search or random search, which makes it difficult to balance the efficiency of solution space traversal and the quality of the solution. In particular, for complex models such as deep neural networks (DNNs), the combined optimization of key parameters such as learning rate and number of hidden layer nodes has not yet formed a systematic methodology. Summary of the Invention

[0009] In order to address the deficiencies in the prior art, the present invention provides a blasting vibration velocity prediction method, system, computer equipment and storage medium based on feature selection and GA-XGBoost model, which can accurately predict the blasting vibration velocity during the blasting process in open-pit mines, thereby guiding the blasting process, optimizing the blasting process, and ensuring blasting efficiency and on-site safety.

[0010] In order to achieve the above object, the specific scheme adopted by the present invention is:

[0011] The blasting vibration velocity prediction method based on feature selection and GA-XGBoost model includes the following steps:

[0012] During the blasting operation in an open-pit mine, multiple basic features are collected, and a portion of the basic features is extracted as the target feature through Pearson correlation analysis;

[0013] Initialize the XGBoost (EXtreme Gradient Boosting, distributed gradient boosting library) model, and optimize and train the XGBoost model using GA (Genetic Algorithm) and a dataset constructed based on target features to obtain a blasting vibration velocity prediction model;

[0014] Dynamic blasting parameters are collected in real time based on target characteristics and input into the blasting vibration velocity prediction model for prediction to obtain the predicted velocity.

[0015] Preferably, the basic characteristics include the distance from the explosion center, the total amount of explosives, the hole spacing, the row spacing, the hole depth, the maximum single-stage explosives and the filling length; the target characteristics include the distance from the explosion center, the total amount of explosives, the hole spacing, the row spacing, the hole depth and the maximum single-stage explosives.

[0016] Preferably, the method of extracting a portion of all basic features as target features through Pearson correlation analysis includes:

[0017] Calculate the Pearson correlation coefficient matrix of all basic features;

[0018] Determine the degree of correlation between any two basic features based on the Pearson correlation coefficient matrix;

[0019] If the correlation between a basic feature and other basic features is lower than the preset correlation threshold, the basic feature will be eliminated;

[0020] All remaining basic features are taken as target features.

[0021] Preferably, the objective function of the XGBoost model is:

[0022]

[0023] Among them, Obj is the optimization target, n is the total number of samples, is the loss function of a single sample, t is the number of iterations, f k is the regularization term;

[0024] The XGBoost model includes multiple decision trees. For the kth decision tree,

[0025]

[0026] Where T is the number of leaf nodes, w j is the leaf weight, γ and λ are hyperparameters.

[0027] Preferably, when optimizing the XGBoost model, a new decision tree is added in each iteration, and the optimal splitting point of the new decision tree is determined based on the objective function.

[0028] Preferably, the method for determining the optimal splitting point of the new decision tree based on the objective function includes:

[0029] Performing a second-order Taylor approximation expansion on the objective function yields:

[0030]

[0031] Among them, g i and h i are the first and second order derivatives of the loss function respectively;

[0032] Calculate the model gain when adding a new decision tree:

[0033]

[0034] Among them, G L ,H L and G R ,H R is the gradient and second-order derivative of the two child nodes of the alternative split point, λ and γ are regularization parameters;

[0035] The optimal split point is re-determined from the alternative split points by maximizing the model gain.

[0036] Preferably, when optimizing the training of the XGBoost model, the determination coefficient R is used. 2 , root mean square error RMSE and mean absolute error MAE determine the model performance.

[0037] A blasting vibration velocity prediction system based on feature selection and GA-XGBoost model is used to implement the above-mentioned blasting vibration velocity prediction method based on feature selection and GA-XGBoost model. The system includes:

[0038] Data acquisition module, used to collect multiple basic features during open-pit mine blasting operations and to collect dynamic blasting parameters in real time based on target features;

[0039] The data preprocessing module is used to extract a part of all basic features as target features through Pearson correlation analysis;

[0040] The model training module is used to initialize the XGBoost (EXtreme Gradient Boosting, distributed gradient boosting library) model and optimize the XGBoost model through GA (Genetic Algorithm) and a dataset constructed based on target features to obtain a blasting vibration velocity prediction model;

[0041] The prediction module is used to predict the dynamic blasting parameters using the blasting vibration velocity prediction model of the training number to obtain the predicted velocity.

[0042] Computer equipment, including:

[0043] memory for storing computer programs;

[0044] A processor is used to read and execute the computer program to implement the above-mentioned blasting vibration velocity prediction method based on feature selection and GA-XGBoost model.

[0045] The storage medium is used to store a computer program, which, when executed, implements the above-mentioned blasting vibration velocity prediction method based on feature selection and GA-XGBoost model.

[0046] The present invention first performs feature selection, selecting target features that significantly influence blasting vibration velocity from all available basic features, eliminating redundant information in the data input dimension. Furthermore, the XGBoost model is globally optimized in the parameter space dimension using Generative Adversarial Networks (GA), thereby training a blasting vibration velocity prediction model with higher accuracy and better performance. Based on this blasting vibration velocity prediction model, the present invention can accurately predict the blasting vibration velocity during open-pit mine blasting, thereby guiding the blasting process, optimizing the blasting process, and ensuring blasting efficiency and on-site safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0048] Figure 1 is a flow chart of the prediction method of the present invention;

[0049] Figure 2 This is a waveform diagram obtained during a blasting test in an open-pit mine.

[0050] Figure 3 It is a schematic diagram of the comparison results of the importance of target features;

[0051] Figure 4 is a heat map of the Pearson correlation coefficient matrix;

[0052] Figure 5 It is the fitness curve of GA with different initial population sizes;

[0053] Figure 6 Schematic diagram of performance comparison of three models in a specific implementation method. DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] like Figure 1As shown, the present invention first provides a blasting vibration velocity prediction method based on feature selection and GA-XGBoost model, which is suitable for blasting vibration velocity prediction in open-pit mines. The method includes S1 to S3.

[0056] S1. During the blasting operation in the open-pit mine, multiple basic features are collected, and a portion of all the basic features is extracted as target features through Pearson correlation analysis. More specifically, the blasting operation adopts a deep-hole step blasting process. During the blasting process, on the one hand, a blasting vibrometer is used to synchronously record the vibration velocity in the radial, tangential, and vertical directions along the deep hole. On the other hand, based on pre-determined monitoring points, a highly sensitive sensor and a digital signal conversion module are fixed to the rock mass using gypsum powder, so that the sensor can form a rigid connection with the rock mass, thereby using the sensor to measure the impact of the blasting process on the rock mass. The data obtained by the blasting vibrometer and all sensors are combined to obtain rich sample data, and then the basic features are determined based on the sample data.

[0057] In one embodiment of the present invention, six monitoring points were selected within the blast zone, covering different distances from the detonation center. This allows sensors to fully capture the propagation and attenuation patterns of the vibration waves generated by the blast. Based on this, basic characteristics include distance from the detonation center, total explosive charge, hole spacing, row spacing, hole depth, maximum single-stage explosive charge, and packing length.

[0058] Considering the complex environment of a blasting site, the sample data acquired may contain outliers. Therefore, after acquiring the sample data, the sample data is first preprocessed to eliminate the impact of unexpected factors such as outliers on the accuracy of the sample data. In one embodiment of the present invention, the acquired sample data can be divided into multiple categories. During data preprocessing, the mean of the sample data for each category is first calculated. The difference between the sample data for that category and the mean is then calculated. The difference is then evaluated. If the difference exceeds a preset deviation threshold, the sample data is treated as an outlier and eliminated. If the difference does not exceed the deviation threshold, the sample data is retained. This method can effectively eliminate outliers from the sample data, ensuring the accuracy of the sample data and, therefore, the accuracy of the entire blasting vibration velocity prediction method. In another embodiment of the present invention, for sample data of the same category, the difference between any two sample data can be calculated. The sample data is then grouped based on the difference. Sample data with a difference less than the preset difference threshold are grouped together. Sample data that cannot be assigned to any group is treated as an outlier and eliminated.

[0059] After removing outliers from the sample data, the sample data can be normalized to facilitate subsequent processing. This maps the sample data to a numerical range of [0, 1]. After normalization, the sample data have the same scale, which can significantly reduce the dimensional differences between different sample data. More specifically, the specific method of normalization is as follows:

[0060]

[0061] where x new is the value of the sample data after normalization, x is the sample data to be normalized, x min is the minimum value of the sample data of the same type, x max It is the maximum value among the sample data of the same type. After normalization, the type of sample data can be used as the basic feature.

[0062] Furthermore, because the sample data obtained is diverse, not all of it significantly affects the blasting vibration velocity. To screen out the basic features that truly significantly affect the blasting vibration velocity from all basic features as target features and reduce interference from other basic features, the present invention performs a Pearson correlation analysis on the basic features. Specifically, the method of extracting a portion of the basic features as target features from all basic features through Pearson correlation analysis includes steps S11 to S14.

[0063] S11. Calculate the Pearson correlation coefficient matrix of all basic features. The Pearson correlation coefficient matrix consists of multiple Pearson correlation coefficients. The Pearson correlation coefficient is used to measure the linear correlation between variables, and its value range is [-1, 1]. The calculation method of the Pearson correlation coefficient is an existing technology and will not be repeated here. The heat map of the Pearson correlation coefficient matrix is ​​as follows: Figure 4 shown.

[0064] S12. Determine the correlation between any two basic features based on the Pearson correlation coefficient matrix.

[0065] S13. If the correlation between a basic feature and other basic features is lower than a preset correlation threshold, the basic feature is eliminated. Specifically, if the correlation between a basic feature and other basic features is very low, it indicates that the basic feature does not affect the other basic features, nor is it affected by the other basic features. Accordingly, the basic feature does not significantly affect the blasting vibration velocity and can be eliminated.

[0066] S14. All remaining basic features are used as target features.

[0067] Among the above basic features, after selecting the basic features through S11 to S14, the selected target features include the distance from the explosion center, the total explosive quantity, the hole spacing, the row spacing, the hole depth and the maximum single-stage explosive quantity. The comparison results of the importance of each target feature are as follows: Figure 3 shown.

[0068] After selecting the target features, the sample data corresponding to the target features are combined into a dataset, and the dataset is divided into a training set and a test set in an 8:2 ratio for subsequent use. Furthermore, when dividing the training set and the test set, the division is based on a random seed to ensure that the results of each division are consistent, thus avoiding anomalies in the training set and test set caused by random division.

[0069] S2. Initialize the XGBoost (EXtreme Gradient Boosting, distributed gradient boosting library) model, and optimize and train the XGBoost model through GA (Genetic Algorithm) and a data set constructed based on target features to obtain a blasting vibration velocity prediction model.

[0070] In the present invention, the objective function of the XGBoost model consists of a loss function and a regularization term, and the objective function is specifically set as:

[0071]

[0072] Among them, Obj is the optimization target, n is the total number of samples, is the loss function of a single sample, t is the number of iterations, f k is the regularization term. In addition, the XGBoost model includes multiple decision trees. For the k-th decision tree,

[0073]

[0074] Where T is the number of leaf nodes, w j is the leaf weight, γ and λ are hyperparameters.

[0075] Furthermore, when optimizing the XGBoost model, a new decision tree is added in each iteration, and the optimal splitting point of the new decision tree is determined based on the objective function. The method of determining the optimal splitting point of the new decision tree based on the objective function includes S211 to S213.

[0076] S211. Perform a second-order Taylor approximation expansion on the objective function to obtain:

[0077]

[0078] Among them, g i and h iare the first and second order derivatives of the loss function, respectively.

[0079] S212. Calculate the model gain when adding a new decision tree:

[0080]

[0081] Among them, G L ,H L and G R ,H R are the gradient and second-order derivative of the two child nodes of the alternative split point, respectively, and λ and γ are regularization parameters.

[0082] S213 : Re-determine the optimal splitting point from the candidate splitting points by maximizing the model gain.

[0083] To ensure the accuracy of the trained blasting vibration velocity prediction model and to facilitate subsequent use of the blasting vibration velocity prediction model to accurately predict the vibration velocity during open-pit mine blasting, the present invention optimizes the training process of the XGBoost model using GA. The specific training process includes S221 to S225.

[0084] S221. Optimize the hyperparameters of the XGBoost model using a general algorithm. A set of hyperparameters is randomly generated using the general algorithm as the initial population of the XGBoost model. Key hyperparameters of the XGBoost model are encoded as chromosomes, with each chromosome representing a possible hyperparameter configuration. Key hyperparameters include learning rate, tree depth, subsample ratio, and regularization coefficient. The definitions and functions of the various hyperparameters in the XGBoost model are known from the prior art and are not further described herein.

[0085] S222. Evaluate the performance of each individual in the initial population through the fitness function, where each individual corresponds to a hyperparameter randomly generated by GA.

[0086] S223. Execute GA, and carry out the selection phase, crossover phase, and mutation phase in sequence. More specifically, in one embodiment of the present invention, the number of individuals in the initial population is set to 50, and the number of evolutionary generations is set to 150. In the selection phase, the roulette wheel selection method is adopted, and individuals with higher fitness have a higher probability of being selected to enter the next generation; in the crossover phase, the crossover probability is set to 0.4, and some genes of two parent individuals are randomly selected for exchange; in the mutation phase, certain genes are randomly perturbed with a mutation probability of 0.2, so as to avoid the population from falling into a local optimum. The fitness curves of the iterative process of GA with different initial population sizes are shown as follows: Figure 5 shown.

[0087] S224. Substituting the hyperparameters optimized by the GA into the XGBoost model to obtain a GA-XGBoost model, and training the GA-XGBoost model. During the training process, the GA-XGBoost model gradually fits the data using a forward stepwise additive model, and each decision tree in the GA-XGBoost model is constructed based on the residuals.

[0088] After GA optimized the XGBoost model's hyperparameters, the resulting GA-XGBoost model achieved higher accuracy. Consequently, the accuracy of the final blasting vibration velocity prediction model was also significantly improved, enabling it to accurately predict blasting vibration velocity during open-pit mine blasting. To verify the accuracy and stability of the GA-optimized XGBoost model, an evaluation of the optimized model was required.

[0089] S225, using the coefficient of determination R 2 , root mean square error RMSE and mean absolute error MAE determine the model performance.

[0090] S3. Dynamic blasting parameters are collected in real time based on target characteristics, and input into a blasting vibration velocity prediction model for prediction to obtain a predicted velocity.

[0091] In a specific embodiment of the present invention, the hyperparameters of the XGBoost model optimized by GA include learning rate, maximum depth of a single decision tree, minimum loss reduction, total number of generated decision trees, and regularization coefficient, wherein the learning rate is set to 0.703, the maximum depth of a single decision tree is set to 3, the minimum loss reduction is set to 0.031, and the total number of generated decision trees is set to 200.

[0092] Furthermore, the ordinary XGBoost model, the XGBoost model based on S1 feature selection, and the GA-XGBoost model based on S1 feature selection were used as three models for comparison, and training and prediction experiments were conducted respectively to evaluate the performance of the three models. The performance comparison results of the three models are shown in the figure below. Figure 6 As shown in the figure, the GA-XGBoost model based on feature selection of S1 has the best performance, and the coefficient of determination R 2 The result is 0.944, the root mean square error RMSE is 4.785, and the mean absolute error MAE is 3.684. 2A high value indicates that the model has a high degree of fit to the data and can well explain the changes in the data; low RMSE and MAE values ​​indicate that the model prediction is accurate, the deviation between the predicted value and the true value is small, and the prediction stability of the outliers is significantly improved.

[0093] In summary, the present invention performs feature selection in S1, and can select target features that significantly affect the blasting vibration velocity from all the basic features that can be collected, eliminating redundant information in the dimension of data input; further, in S2, the XGBoost model is globally tuned in the dimension of the parameter space through GA, so that a blasting vibration velocity prediction model with higher accuracy and better performance can be trained. Based on this blasting vibration velocity prediction model, the present invention can accurately predict the blasting vibration velocity during the blasting process in the open-pit mine area, thereby guiding the blasting process, optimizing the blasting process, and ensuring blasting efficiency and on-site safety.

[0094] During a blasting process in an open-pit mine, the present invention was used to predict the blasting vibration velocity and monitor it through a preset sensor network. The monitoring points were arranged at six locations with different distances from the blast center around the blasting area. The direction of the sensor was calibrated using a light level to ensure orthogonal alignment in the three axes of X, Y, and Z, where the X direction is the longitudinal direction, the Y direction is the tangential direction, and the Z direction is the vertical direction. After data processing, the waveforms in the three directions are as follows: Figure 2 shown.

[0095] The present invention also provides a blasting vibration velocity prediction system based on feature selection and GA-XGBoost model, which is used to implement the above-mentioned blasting vibration velocity prediction method based on feature selection and GA-XGBoost model. The system includes a data acquisition module, a data preprocessing module, a model training module and a prediction module.

[0096] The data acquisition module is used to collect multiple basic features during open-pit mine blasting operations and to collect dynamic blasting parameters in real time based on target features.

[0097] The data preprocessing module is used to extract a part of all basic features as target features through Pearson correlation analysis.

[0098] The model training module is used to initialize the XGBoost (EXtreme Gradient Boosting, distributed gradient boosting library) model and optimize the XGBoost model through GA (Genetic Algorithm) and a data set constructed based on target features to obtain a blasting vibration velocity prediction model.

[0099] The prediction module is used to predict the dynamic blasting parameters using the blasting vibration velocity prediction model of the training number to obtain the predicted velocity.

[0100] The present invention also provides a computer device including a memory and a processor.

[0101] Memory for storing computer programs.

[0102] A processor is used to read and execute the computer program to implement the above-mentioned blasting vibration velocity prediction method based on feature selection and GA-XGBoost model.

[0103] Finally, the present invention provides a storage medium for storing a computer program, which, when executed, implements the above-mentioned blasting vibration velocity prediction method based on feature selection and GA-XGBoost model.

[0104] The above embodiments can be implemented in whole or in part through software, hardware (such as circuits), firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired method (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage system such as a server or data center that contains a collection of one or more available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, or tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0105] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0106] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0107] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0108] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0109] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0110] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units 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 interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0111] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0112] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0113] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer system (which can be a personal computer, a server, or a network system, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0114] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0115] There are a few points to note:

[0116] (1) The drawings of the embodiments of the present invention only relate to the structures related to the embodiments of the present invention. Other structures may refer to conventional designs.

[0117] (2) For the sake of clarity, the thickness of layers or regions in the drawings used to describe the embodiments of the present invention are exaggerated or reduced, that is, these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region, or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or intervening elements may be present.

[0118] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to form new embodiments.

[0119] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

[0120] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A blasting vibration velocity prediction method based on feature selection and GA-XGBoost model is characterized by: The steps include: During the blasting operation in an open-pit mine, multiple basic features are collected, and a portion of the basic features is extracted as the target feature through Pearson correlation analysis; Initialize the XGBoost (EXtreme Gradient Boosting, distributed gradient boosting library) model, and optimize and train the XGBoost model using GA (Genetic Algorithm) and a dataset constructed based on target features to obtain a blasting vibration velocity prediction model; Dynamic blasting parameters are collected in real time based on target characteristics and input into the blasting vibration velocity prediction model for prediction to obtain the predicted velocity.

2. The blasting vibration velocity prediction method based on feature selection and GA-XGBoost model according to claim 1 is characterized in that: The basic characteristics include the distance from the blast center, the total amount of explosives, the hole spacing, the row spacing, the hole depth, the maximum single-stage explosives and the filling length; the target characteristics include the distance from the blast center, the total amount of explosives, the hole spacing, the row spacing, the hole depth and the maximum single-stage explosives.

3. The blasting vibration velocity prediction method based on feature selection and GA-XGBoost model according to claim 2, characterized in that: Methods for extracting a portion of all basic features as target features through Pearson correlation analysis include: Calculate the Pearson correlation coefficient matrix of all basic features; Determine the degree of correlation between any two basic features based on the Pearson correlation coefficient matrix; If the correlation between a basic feature and other basic features is lower than the preset correlation threshold, the basic feature will be eliminated; All remaining basic features are taken as target features.

4. The blasting vibration velocity prediction method based on feature selection and GA-XGBoost model according to claim 1, characterized in that: The objective function of the XGBoost model is: Among them, Obj is the optimization target, n is the total number of samples, is the loss function of a single sample, t is the number of iterations, f k is the regularization term; The XGBoost model includes multiple decision trees. For the kth decision tree, Where T is the number of leaf nodes, w j is the leaf weight, γ and λ are hyperparameters.

5. The blasting vibration velocity prediction method based on feature selection and GA-XGBoost model according to claim 4 is characterized in that: When optimizing the XGBoost model, a new decision tree is added in each iteration, and the optimal split point of the new decision tree is determined based on the objective function.

6. The blasting vibration velocity prediction method based on feature selection and GA-XGBoost model according to claim 5, characterized in that: Methods for determining the optimal splitting point of a new decision tree based on the objective function include: Performing a second-order Taylor approximation expansion on the objective function yields: Among them, g i and h i are the first and second order derivatives of the loss function respectively; Calculate the model gain when adding a new decision tree: Among them, G L ,H L and G R ,H R is the gradient and second-order derivative of the two child nodes of the alternative split point, λ and γ are regularization parameters; The optimal split point is re-determined from the alternative split points by maximizing the model gain.

7. The blasting vibration velocity prediction method based on feature selection and GA-XGBoost model according to claim 1, characterized in that: When optimizing the training of the XGBoost model, the determination coefficient R is used. 2 , root mean square error RMSE and mean absolute error MAE determine the model performance.

8. Blasting vibration velocity prediction system based on feature selection and GA-XGBoost model, characterized by: A method for predicting blasting vibration velocity based on feature selection and GA-XGBoost model according to any one of claims 1 to 7 is implemented, wherein the system comprises: Data acquisition module, used to collect multiple basic features during open-pit mine blasting operations and to collect dynamic blasting parameters in real time based on target features; The data preprocessing module is used to extract a part of all basic features as target features through Pearson correlation analysis; The model training module is used to initialize the XGBoost (EXtreme Gradient Boosting, distributed gradient boosting library) model and optimize the XGBoost model through GA (Genetic Algorithm) and a dataset constructed based on target features to obtain a blasting vibration velocity prediction model; The prediction module is used to predict the dynamic blasting parameters using the blasting vibration velocity prediction model of the training number to obtain the predicted velocity.

9. Computer device, characterized in that include: memory for storing computer programs; A processor is configured to read and execute the computer program to implement the blasting vibration velocity prediction method based on feature selection and GA-XGBoost model as described in any one of claims 1 to 7.

10. A storage medium, characterized in that Used to store a computer program, which, when executed, implements the blasting vibration velocity prediction method based on feature selection and GA-XGBoost model as described in any one of claims 1 to 7.

Images