Intelligent design method for tunnel blasting scheme

Abstract

The invention discloses an intelligent design method for a tunnel blasting scheme, which comprises the following steps: S1, data acquisition and preprocessing: acquiring geological data of an area where a tunnel is located through geological survey and field survey, and preprocessing the data to generate a geological model; s2, intelligent analysis and parameter matching: intelligently matching optimal blasting parameters by using a big data analysis technology according to the geological model and historical blasting data; s3, blasting scheme generation: generating a specific blasting design scheme based on the matched blasting parameters; s4, simulation and optimization: performing simulation analysis on the blasting scheme by using a numerical simulation technology, evaluating the blasting effect and safety, and optimizing the blasting scheme according to a simulation result; and S5, scheme verification and implementation: carrying out field test and verification on the optimized blasting scheme. Through an intelligent algorithm and a software tool, automatic and rapid blasting scheme design is realized, the design period is greatly shortened, and the efficiency is improved.

Description

Technical Field

[0001] The invention relates to the technical field of tunnel engineering blasting design, and in particular to an intelligent design method for a tunnel blasting scheme. Background Art

[0002] In tunnel engineering projects, blasting design is a key link in tunnel excavation. Traditional blasting design methods mainly rely on manual experience and manual calculations, and have problems such as low design efficiency and poor accuracy. These problems are particularly prominent in tunnel projects with complex geological conditions, which may lead to construction delays, increased costs and safety hazards. Therefore, improving the intelligence and automation level of blasting design is the key to improving construction efficiency and safety.

[0003] The intelligent blasting scheme design method aims to improve the efficiency and accuracy of blasting design, ensure construction safety and reduce construction costs by integrating multi-source data and intelligent algorithms. The method includes collecting geological survey data, using construction equipment sensor data, applying intelligent algorithms for design, and optimizing the scheme according to the on-site conditions.

[0004] Although the intelligent blasting scheme design method has significant advantages in tunnel construction projects, its application still faces a series of challenges. For example, complex geological conditions and construction environment may affect the accuracy and transmission efficiency of data, and the stability and accuracy of construction equipment sensors are crucial to the design results. In addition, changes in on-site conditions require the scheme to be highly flexible and adjustable. These challenges need to be effectively addressed in the intelligent blasting scheme design process to ensure its application effect in actual engineering.

[0005] Through the intelligent blasting plan design method, real-time optimization and adjustment of blasting design in tunnel construction can be achieved to ensure the scientificity and rationality of the plan, thereby improving the overall efficiency and safety of tunnel construction.

[0006] The existing technology has the following major shortcomings in tunnel blasting design, which limit its effectiveness and efficiency in practical applications:

[0007] 1. Low design efficiency: Traditional methods rely on experience and manual calculations, with long design cycles and low efficiency.

[0008] 2. Inaccurate blasting parameters: The geological conditions are complex and changeable, and it is difficult for traditional methods to accurately match blasting parameters, resulting in unsatisfactory blasting results.

[0009] 3. High safety risk: The uncertainty in traditional design methods increases the safety risk during the blasting process.

[0010] 4. Lack of optimization methods: Traditional blasting design lacks scientific optimization methods, and it is difficult to find the best solution under complex constraints.

[0011] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present invention, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0012] The purpose of the present invention is to provide an intelligent design method for a tunnel blasting scheme to solve the technical problems existing in the prior art.

[0013] In order to achieve the above object, the present invention adopts the following technical solutions:

[0014] The present invention provides a tunnel blasting scheme intelligent design method, the design method comprising the following steps:

[0015] S1. Data collection and preprocessing: Through geological surveys and on-site investigations, geological data of the tunnel area, including rock types, geological structures, and groundwater conditions, are collected, and the data is preprocessed to generate a geological model;

[0016] S2. Intelligent analysis and parameter matching: Using big data analysis technology, according to geological models and historical blasting data, intelligently match the optimal blasting parameters, including explosive type, charge amount, hole spacing and row spacing;

[0017] S3, blasting plan generation: based on the matched blasting parameters, generate a specific blasting design plan, including blasting drawings, blasting sequence and safety protection measures;

[0018] S4, simulation and optimization: Use numerical simulation technology to simulate and analyze the blasting plan, evaluate the blasting effect and safety, and optimize the blasting plan based on the simulation results;

[0019] S5. Scheme verification and implementation: Conduct field tests and verification on the optimized blasting scheme to ensure the feasibility and effectiveness of the design scheme, and finally implement the blasting operation.

[0020] Furthermore, the data acquisition and preprocessing include geological data acquisition equipment, data processing software and geological model generation tools; wherein, the geological data acquisition equipment includes geological radar and core drilling machine, which obtain detailed geological information of the tunnel area through on-site survey and geological investigation; the data processing software includes data cleaning tools and data format conversion tools, which ensure the accuracy and consistency of the data by cleaning, outlier processing and format conversion of the collected data; the geological model generation tool includes three-dimensional modeling software and geological analysis software, which generates a three-dimensional geological model by modeling the processed data, providing a basis for the subsequent blasting plan design.

[0021] Furthermore, the intelligent analysis and parameter matching include a big data analysis platform and a machine learning algorithm; wherein the big data analysis platform integrates data mining, statistical analysis and pattern recognition technologies, and is capable of conducting in-depth analysis of a large amount of historical blasting data to extract key features and rules; the machine learning algorithm is used to intelligently match the optimal blasting parameters, including the type of explosives, charge, hole spacing and row spacing, through training and verification.

[0022] Furthermore, the machine learning algorithm uses an integrated time series analysis and prediction model LSTM to analyze geological conditions and historical blasting data and predict future geological changes;

[0023] Model input: geological data sequence X = {x1, x2, ..., x t}, each x i Including the characteristics of geological conditions; blasting history data sequence Y = {y1, y2, ..., y t}, each y i Including blasting parameters and results;

[0024] Model output: predictions of future geological conditions

[0025] Recommended values ​​of blasting parameters corresponding to time

[0026] The algorithm implementation process is as follows:

[0027] 1) LSTM network architecture:

[0028] Input layer: accepts geological data and blasting history data; hidden layer: LSTM units, capturing long-term dependencies in time series; output layer: predicts future geological conditions and blasting parameters;

[0029] 2) LSTM unit formula:

[0030] Input gate: i t =σ(W i ·[h t-1 ,x t ]+b i )

[0031] Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f )

[0032] Output gate: o t =σ(W o ·[h t-1 ,x t ]+bo )

[0033] Candidate memory cells:

[0034] Update memory unit:

[0035] The final output is h t =o t tanh(C t )

[0036] Input Gate i t : Determines how much new input information will be stored in the cell state; where,

[0037] σ is the sigmoid function, which compresses the input to between 0 and 1 and is used to indicate the amount of information retained; W i is the weight matrix of the input gate; [h t-1 ,x t ] is the concatenation of the hidden state of the previous time step and the input of the current time step; b i is the bias of the input gate;

[0038] Forget Gate t : Determines how many previous cell states will be retained;

[0039] W f is the weight matrix of the forget gate; b f is the bias of the forget gate;

[0040] Output gate o t : Determines how many current unit states will be output;

[0041] W o is the weight matrix of the output gate; b o is the bias of the output gate;

[0042] Candidate memory cells The candidate value at time t, this value will be combined with the output of the input gate to update the state of the current unit;

[0043] tanh is the hyperbolic tangent function, which compresses the input to between -1 and 1; W ( is the weight matrix of the candidate memory unit; b ( is the bias of the candidate memory cell;

[0044] Update memory unit C t .: is the new state of the current time step, which is determined by the combination of the previous state and the current candidate state; where,

[0045] f tControls the amount of previous state retained; C t-1 is the cell state at the previous time step; i t Control how much new information is added to the state; is the current candidate status;

[0046] The final output is h t : is the value of the current unit output, which depends on the current unit state and the output gate; where,

[0047] o t is the output of the output gate, controlling part of the state information of the output; tanh(C t ) is a nonlinear transformation of the current state;

[0048] 3) Training process:

[0049] Loss function: Mean Squared Error (MSE): Among them, y i is the actual value, is the predicted value, n is the number of samples;

[0050] Optimization algorithm: Adam optimizer;

[0051] 4) Prediction process: Use the trained LSTM model to input current geological data and blasting history data to predict future geological conditions and blasting parameters.

[0052] Furthermore, the blasting plan generation includes blasting design software and automated design tools; the blasting design software is used to automatically generate blasting drawings, including blasthole layout diagrams, charging schematic diagrams and detonation network diagrams; the automated design tool is used to automatically generate blasting sequences and safety protection measures based on intelligently matched blasting parameters.

[0053] Furthermore, the simulation and optimization include numerical simulation software and optimization algorithms; the numerical simulation software has high-precision computing capabilities and is used to simulate stress wave propagation, rock fracture and kinetic energy release phenomena during blasting, and to evaluate blasting effects and safety; the optimization algorithm is used to optimize the blasting plan based on the simulation results, including adjusting the charge, hole spacing and row spacing parameters to improve blasting effects and safety.

[0054] Furthermore, the optimization algorithm adopts a Bayesian optimization model, the model input: blasting parameter search space χ; objective function f(x), representing the evaluation index of blasting effect; model output: optimal blasting parameter x * ;

[0055] The algorithm implementation process is as follows:

[0056] 1) Gaussian process model of objective function:

[0057] Establish the Gaussian process prior distribution of the target function f(x) where μ(x) is the mean function, k(x,x 3 ) is the kernel function;

[0058] 2) Select the acquisition function:

[0059] The acquisition function EI guides the next step in selecting the optimal parameter point;

[0060] Expected Improvement Function

[0061] Among them, f(x + ) is the current best observed value;

[0062] 3) Update the Gaussian process model:

[0063] Based on new observational data Update the mean and covariance matrix of the Gaussian process;

[0064] 4) Bayesian optimization process:

[0065] Initialization: Randomly select several initial blast parameter points {x1,x2,...,x n}, calculate the objective function value {f(x1),f(x2),...,f(x n )};

[0066] Loop execution:

[0067] According to the current Gaussian process model, select the next parameter point x n+1 = arg maxEI(x);

[0068] Calculate f(x n+1 ), and (x n+1 ,f(x n+1 ))Add observation data;

[0069] Update the Gaussian process model;

[0070] Until the stopping condition is met;

[0071] 5) Output optimal parameters:

[0072] According to the final Gaussian process model, select the blasting parameter x with the optimal objective function value * .

[0073] Furthermore, the scheme verification and implementation is used to transmit the preliminary blasting scheme back to the trolley construction personnel, correct the deviation according to the on-site conditions, and generate the final blasting scheme implementation archive; the scheme verification and implementation includes field test equipment and implementation tools; the field test equipment includes drilling machines, charge loaders, detonators and protective equipment, which are used to conduct on-site tests and verification of the optimized blasting scheme; the implementation tools include construction management software and safety monitoring systems, which ensure the safety and effectiveness of blasting operations by recording and analyzing data and phenomena during the blasting process.

[0074] By adopting the above technical solution, the present invention has the following beneficial effects:

[0075] Improve design efficiency: Through intelligent algorithms and software tools, automatic and rapid blasting plan design can be achieved, which greatly shortens the design cycle and improves efficiency.

[0076] Accurately match blasting parameters: Utilize big data analysis and machine learning technology to accurately match geological conditions and blasting parameters to ensure the accuracy and consistency of blasting effects.

[0077] Reduce safety risks: Through intelligent analysis and simulation, potential safety risks can be identified and prevented in advance, improving the safety of blasting operations.

[0078] Scientific optimization: Using intelligent optimization algorithms, we can find the best blasting plan under various constraints to improve economic and safety benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0080] Figure 1 A flowchart for implementing the intelligent design method for tunnel blasting scheme provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0081] The technical solution of the present invention will be described clearly and completely below 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 creative work are within the scope of protection of the present invention.

[0082] The specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present invention, and is not used to limit the present invention.

[0083] Combination Figure 1 As shown, this embodiment provides a method for intelligent design of a tunnel blasting scheme, comprising the following steps:

[0084] Data collection and preprocessing: Through geological surveys and on-site investigations, geological data of the tunnel area are collected, including rock types, geological structures, groundwater conditions, etc. The data are preprocessed to generate a geological model.

[0085] Intelligent analysis and parameter matching: Using big data analysis technology, according to geological models and historical blasting data, intelligently match the optimal blasting parameters, including type of explosives, charge, hole spacing and row spacing.

[0086] Blasting plan generation: Based on the matching blasting parameters, a specific blasting design plan is generated, including blasting drawings, blasting sequence and safety protection measures.

[0087] Simulation and optimization: Use numerical simulation technology to simulate and analyze blasting plans, evaluate blasting effects and safety, and optimize blasting plans based on simulation results.

[0088] Scheme verification and implementation: Conduct field tests and verification on the optimized blasting scheme to ensure the feasibility and effectiveness of the design scheme, and finally implement the blasting operation.

[0089] Data acquisition and preprocessing module: This module includes geological data acquisition equipment, data processing software and geological model generation tools. Geological data acquisition equipment includes geological radar, core drilling machine, etc., which obtain detailed geological information of the tunnel area through on-site survey and geological investigation. Data processing software includes data cleaning tools and data format conversion tools, which ensure the accuracy and consistency of data by cleaning, outlier processing and format conversion of the collected data. Geological model generation tools include 3D modeling software and geological analysis software, which generate 3D geological models by modeling the processed data, providing a basis for the subsequent blasting plan design.

[0090] The following table describes the collected data:

[0091]

[0092]

[0093] Intelligent analysis and parameter matching module: This module includes a big data analysis platform and a machine learning algorithm. The big data analysis platform integrates technologies such as data mining, statistical analysis, and pattern recognition, and can conduct in-depth analysis of a large amount of historical blasting data to extract key features and patterns. Through training and verification, the machine learning algorithm can intelligently match the optimal blasting parameters, including the type of explosives, charge, hole spacing, and row spacing. The intelligent analysis and parameter matching module can effectively improve blasting efficiency and safety, and reduce construction costs and environmental impact.

[0094] The machine learning algorithm uses an integrated time series analysis and prediction model LSTM (Long Short-Term Memory Network) to analyze geological conditions and historical blasting data and predict future geological changes.

[0095] Input: geological data sequence X = {x1, x2, ..., x t}, each x i Including the characteristics of geological conditions (such as rock type, geological structure, etc.); blasting history data sequence Y = {y1, y2, ..., y t}, each y i Including blasting parameters and results (such as type of explosives, charge, blasting effect, etc.).

[0096] Output: Predicted values ​​of future geological conditions

[0097] Recommended values ​​of blasting parameters corresponding to time

[0098] Algorithm implementation:

[0099] 1.LSTM network architecture:

[0100] Input layer: accepts geological data and blasting history data. Hidden layer: LSTM units, capturing long-term dependencies in time series. Output layer: predicts future geological conditions and blasting parameters.

[0101] 2.LSTM unit formula:

[0102] Input gate: i t =σ(W i ·[h t-1 ,x t ]+b i )

[0103] Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f )

[0104] Output gate: o t=σ(W o ·[h t-1 ,x t ]+b o )

[0105] Candidate memory cells:

[0106] Update memory unit:

[0107] The final output is h t =o t tanh(C t )

[0108] 3. Training process:

[0109] Loss function: Mean Squared Error (MSE):

[0110] Optimization algorithm: Adam optimizer.

[0111] 4. Prediction process: Use the trained LSTM model to input current geological data and blasting history data to predict future geological conditions and blasting parameters.

[0112] Blasting plan generation module: This module includes blasting design software and automated design tools. The blasting design software has powerful graphics processing and calculation functions, and can automatically generate blasting drawings, including blasthole layout diagrams, charge schematic diagrams, and detonation network diagrams. The automated design tool can automatically generate blasting sequences and safety protection measures based on intelligently matched blasting parameters. The blasting plan generation module can effectively improve the efficiency and accuracy of blasting plan design and ensure the safety and effectiveness of blasting operations.

[0113] Simulation and optimization module: This module includes numerical simulation software and optimization algorithms. The numerical simulation software has high-precision computing capabilities and can simulate phenomena such as stress wave propagation, rock fracture and kinetic energy release during blasting, and evaluate blasting effects and safety. The optimization algorithm can optimize the blasting scheme based on the simulation results, such as adjusting parameters such as charge amount, hole spacing and row spacing to improve blasting effects and safety. The simulation and optimization module can effectively improve the quality and reliability of blasting schemes and reduce construction costs and environmental impacts.

[0114] Blasting parameter optimization: Bayesian optimization

[0115] Input: blast parameter search space (such as type of explosives, charge, hole spacing, row spacing); objective function f(x), which represents the evaluation index of blasting effect (such as uniformity and safety of blasting effect, etc.).

[0116] Output: optimal blasting parameter x* .

[0117] Algorithm implementation:

[0118] 1. Gaussian process model of objective function:

[0119] Establish the Gaussian process prior distribution of the target function f(x) where μ(x) is the mean function, k(x,x 3 ) is the kernel function.

[0120] 2. Select the acquisition function:

[0121] The acquisition function (such as expected improvement function EI and upper confidence limit LCB) guides the next step of selecting the optimal parameter point.

[0122]

[0123] Among them, f(x + ) is the current best observed value.

[0124] 3. Update the Gaussian process model:

[0125] Based on new observational data Update the mean and covariance matrix of the Gaussian process.

[0126] 4. Bayesian optimization process:

[0127] Initialization: Randomly select several initial blast parameter points {x1,x2,...,x n}, calculate the objective function value {f(x1),f(x2),...,f(x n )}.

[0128] Loop execution:

[0129] According to the current Gaussian process model, select the next parameter point x n+1 =argmaxEI(x).

[0130] Calculate f(x n+1 ), and (x n+1 ,f(x n+1 ))Add observation data.

[0131] Update the Gaussian process model.

[0132] Until a stopping condition (such as a maximum number of iterations or a convergence criterion) is met.

[0133] 5. Output optimal parameters:

[0134] According to the final Gaussian process model, select the blasting parameter x with the optimal objective function value * .

[0135] By combining LSTM for time series analysis and Bayesian optimization for parameter selection, the tunnel blasting plan can be intelligentized and automated, improving design efficiency and accuracy.

[0136] Scheme verification and implementation module: The preliminary blasting scheme is sent back to the trolley construction personnel, and the deviation is corrected according to the on-site conditions, and the final blasting scheme is generated for implementation and archiving. This module includes field test equipment and implementation tools. The field test equipment includes drilling machines, charge loaders, detonators and protective equipment, etc., which can conduct on-site tests and verification of the optimized blasting scheme. Implementation tools include construction management software and safety monitoring systems, which ensure the safety and effectiveness of blasting operations by recording and analyzing data and phenomena during the blasting process. The scheme verification and implementation module can effectively ensure the feasibility and effectiveness of the blasting scheme, guide actual construction, and ensure the safety, efficiency and environmental protection of blasting operations.

[0137] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent design of tunnel blasting scheme, characterized in that: The design method comprises the following steps: S1. Data collection and preprocessing: Through geological surveys and on-site investigations, geological data of the tunnel area, including rock types, geological structures, and groundwater conditions, are collected, and the data is preprocessed to generate a geological model; S2. Intelligent analysis and parameter matching: Using big data analysis technology, according to geological models and historical blasting data, intelligently match the optimal blasting parameters, including explosive type, charge amount, hole spacing and row spacing; S3, blasting plan generation: based on the matched blasting parameters, generate a specific blasting design plan, including blasting drawings, blasting sequence and safety protection measures; S4, simulation and optimization: use numerical simulation technology to simulate and analyze the blasting plan, evaluate the blasting effect and safety, and optimize the blasting plan based on the simulation results; S5. Scheme verification and implementation: Conduct field tests and verification on the optimized blasting scheme to ensure the feasibility and effectiveness of the design scheme, and finally implement the blasting operation.

2. The intelligent design method for tunnel blasting scheme according to claim 1, characterized in that: The data acquisition and preprocessing includes geological data acquisition equipment, data processing software and geological model generation tools; wherein, the geological data acquisition equipment includes geological radar and core drilling machine, which obtain detailed geological information of the tunnel area through on-site survey and geological investigation; the data processing software includes data cleaning tools and data format conversion tools, which ensure the accuracy and consistency of the data by cleaning, outlier processing and format conversion of the collected data; the geological model generation tool includes three-dimensional modeling software and geological analysis software, which generates a three-dimensional geological model by modeling the processed data, providing a basis for the subsequent blasting plan design.

3. The intelligent design method for tunnel blasting scheme according to claim 1, characterized in that: The intelligent analysis and parameter matching includes a big data analysis platform and a machine learning algorithm; wherein the big data analysis platform integrates data mining, statistical analysis and pattern recognition technology, and can conduct in-depth analysis of a large amount of historical blasting data to extract key features and laws; the machine learning algorithm is used to intelligently match the optimal blasting parameters, including the type of explosives, charge amount, hole spacing and row spacing, through training and verification.

4. The intelligent design method for tunnel blasting scheme according to claim 3 is characterized in that: The machine learning algorithm uses an integrated time series analysis and prediction model LSTM to analyze geological conditions and historical blasting data and predict future geological changes; Model input: geological data sequence X = {x1, x2, ..., x t }, each x i Including the characteristics of geological conditions; blasting history data sequence Y = {y1, y2, ..., y t }, each y i Including blasting parameters and results; Model output: predictions of future geological conditions Recommended values ​​of blasting parameters corresponding to time The algorithm implementation process is as follows: 1) LSTM network architecture: Input layer: accepts geological data and blasting history data; hidden layer: LSTM units, capturing long-term dependencies in time series; output layer: predicts future geological conditions and blasting parameters; 2) LSTM unit formula: Input gate: i t =σ(W i ·[h t-1 ,x t ]+b i ) Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f ) Output gate: o t =σ(W o ·[h t-1 ,x t ]+b o ) Candidate memory cells: Update memory unit: The final output is h t =o t tanh(C t ) Input Gate i t : Determines how much new input information will be stored in the cell state; where, σ is the sigmoid function, which compresses the input to between 0 and 1 and is used to indicate the amount of information retained; W i is the weight matrix of the input gate; [h t-1 ,x t ] is the concatenation of the hidden state of the previous time step and the input of the current time step; b i is the bias of the input gate; Forget Gate t : Determines how much of the previous unit state will be retained; W f is the weight matrix of the forget gate; b f is the bias of the forget gate; Output gate o t : Determines how many current unit states will be output; W o is the weight matrix of the output gate; b o is the bias of the output gate; Candidate memory cells The candidate value at time t, this value will be combined with the output of the input gate to update the state of the current unit; tanh is the hyperbolic tangent function, which compresses the input to between -1 and 1; W ( is the weight matrix of the candidate memory unit; b ( is the bias of the candidate memory cell; Update memory unit C t .: is the new state of the current time step, which is determined by the combination of the previous state and the current candidate state; where, f t Controls the amount of previous state retained; C t-1 is the cell state at the previous time step; i t Control how much new information is added to the state; is the current candidate status; The final output is h t : is the value of the current unit output, which depends on the current unit state and the output gate; where, o t is the output of the output gate, controlling part of the state information of the output; tanh(C t ) is a nonlinear transformation of the current state; 3) Training process: Loss function: Mean Squared Error (MSE): Among them, y i is the actual value, is the predicted value, n is the number of samples; Optimization algorithm: Adam optimizer; 4) Prediction process: Use the trained LSTM model to input current geological data and blasting history data to predict future geological conditions and blasting parameters.

5. The intelligent design method for tunnel blasting scheme according to claim 1, characterized in that: The blasting plan generation includes blasting design software and automated design tools; the blasting design software is used to automatically generate blasting drawings, including blasthole layout diagrams, charge schematic diagrams and detonation network diagrams; the automated design tools are used to automatically generate blasting sequences and safety protection measures based on intelligently matched blasting parameters.

6. The intelligent design method for tunnel blasting scheme according to claim 1, characterized in that: The simulation and optimization include numerical simulation software and optimization algorithms; the numerical simulation software has high-precision computing capabilities and is used to simulate stress wave propagation, rock fracture and kinetic energy release phenomena during blasting, and to evaluate blasting effects and safety; the optimization algorithm is used to optimize blasting plans based on simulation results, including adjusting charge amount, hole spacing and row spacing parameters to improve blasting effects and safety.

7. The intelligent design method for tunnel blasting scheme according to claim 6, characterized in that: The optimization algorithm adopts the Bayesian optimization model. The model input is: blasting parameter search space x; objective function f(x), which represents the evaluation index of blasting effect; model output is: optimal blasting parameter x * ; The algorithm implementation process is as follows: 1) Gaussian process model of objective function: Establish the Gaussian process prior distribution of the target function f(x) where μ(x) is the mean function, k(x,x 3 ) is the kernel function; 2) Select the acquisition function: The acquisition function EI guides the next step in selecting the optimal parameter point; Expected Improvement Function Among them, f(x + ) is the current best observed value; 3) Update the Gaussian process model: Based on new observational data Update the mean and covariance matrix of the Gaussian process; 4) Bayesian optimization process: Initialization: Randomly select several initial blast parameter points {x1,x2,...,x n }, calculate the objective function value {f(x1),f(x2),...,f(x n )}; Loop execution: According to the current Gaussian process model, select the next parameter point x n+1 = arg maxEI(x); Calculate f(x n+1 ), and (x n+1 ,f(x n+1 ))Add observation data; Update the Gaussian process model; Until the stopping condition is met; 5) Output optimal parameters: According to the final Gaussian process model, select the blasting parameter x with the optimal objective function value * .

8. The intelligent design method for tunnel blasting scheme according to claim 1, characterized in that: The scheme verification and implementation is used to transmit the preliminary blasting scheme back to the trolley construction personnel, correct the deviation according to the on-site conditions, and generate the final blasting scheme implementation archive; the scheme verification and implementation includes on-site test equipment and implementation tools; the on-site test equipment includes drilling machines, charge loaders, detonators and protective equipment, which are used to conduct on-site tests and verification of the optimized blasting scheme; the implementation tools include construction management software and safety monitoring systems, which ensure the safety and effectiveness of the blasting operation by recording and analyzing the data and phenomena during the blasting process.

Images