A tunnel blasting scheme intelligent design method

Abstract

The application provides a tunnel blasting scheme intelligent design method, and relates to the technical field of tunnel engineering. The application obtains blasting design parameters based on a tunnel engineering geological exploration report and a forepoling geological prediction; obtains a drill-and-blast method tunnel blasting scheme 01 based on the blasting design parameters (4) by using a tunnel engineering blasting theory and a computer deep learning technology; establishes an existing drill-and-blast method tunnel blasting scheme sample library of built or under-construction tunnels based on the blasting design parameters (10) by using the computer deep learning technology, and obtains a to-be-built drill-and-blast method tunnel blasting scheme 02 by inputting the blasting design parameters and tunnel engineering requirements; and obtains a final drill-and-blast method tunnel blasting scheme according to the economic, efficient, safe and simple principle by using an artificial selection method based on the obtained drill-and-blast method tunnel blasting scheme group. The method avoids the influence of subjective factors, combines a traditional experience design method, big data processing and artificial intelligence technology, and provides an efficient scheme for design and construction personnel.

Description

Technical Field

[0001] This invention relates to the field of tunnel engineering technology, and in particular to an intelligent design method for tunnel blasting schemes. Background Technology

[0002] Tunnels, as engineering structures on transportation routes, have significant social and economic benefits. Tunnel construction is an important part of the development process of various countries, and the design of tunnel blasting schemes plays a decisive role in tunnel construction.

[0003] Currently, drill-and-blast method remains the primary construction method in tunnel excavation. Blasting schemes and parameter designs are mainly based on the experience of blasting technicians and engineering analogies. This leads to problems such as significant subjective influence and low design efficiency in blasting design. With increasingly stringent requirements for efficiency, quality, safety, and environmental protection, manual experience and engineering analogies are no longer sufficient to meet the evolving demands of blasting technology. Therefore, there is an urgent need to develop an intelligent design method for drill-and-blast tunnel blasting schemes and parameters. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent design method for tunnel blasting schemes.

[0005] A method for intelligent design of tunnel blasting schemes, specifically including the following steps:

[0006] Step 1: Based on the tunnel's engineering geological survey report and the advanced geological prediction of the tunnel face, obtain the blasting design parameters ①;

[0007] The blasting design parameters ① are the blasting design parameters; including the burial depth, chainage range, surrounding rock grade, surrounding rock parameters, stratum parameters, geological structure, geostress distribution, and hydrogeological characteristics of the tunnel blasting scheme design section.

[0008] The engineering geological survey report and the advanced geological prediction of the tunnel face are provided by the design institute before tunnel excavation and by the tunnel construction unit during tunnel excavation. Before excavation, the blasting design parameters ② obtained from the engineering geological survey report are used to assist the tunnel engineering design unit in the design process. During excavation, the blasting design parameters ③ obtained from the advanced geological prediction of the tunnel face are used to optimize the blasting plan for the tunnel construction unit based on the actual conditions of the tunnel face, improving blasting efficiency and safety. If blasting design parameters ② and ③ are the same, the parameters remain unchanged; that is, blasting design parameter ① is the blasting design parameter ② obtained from the tunnel's engineering geological survey report. If the two parameters are different, then blasting design parameter ① is the blasting design parameter ③ obtained by the tunnel construction unit using the actual advanced geological prediction of the tunnel face.

[0009] Step 2: Based on the blasting design parameters ④, the tunnel blasting scheme 01 of the drill-and-blast method is obtained by using the tunnel engineering blasting theory and computer deep learning technology.

[0010] The blasting design parameters ④ include the blasting design parameters ①, the tunnel engineering requirements, and the mechanical equipment capacity.

[0011] The tunnel blasting scheme 01 of the drill-and-blast method includes a construction section view, a blasthole layout view, a slotting hole layout view, and a blasting parameter table.

[0012] Step 2.1: According to the tunnel engineering requirements, the tunnel excavation section, i.e., the shape and size of the tunnel face, is obtained.

[0013] Step 2.2: According to the blasting design parameters ④, a tunnel excavation method selection model is constructed based on deep learning technology.

[0014] Step 2.2.1: A sample library is constructed, which includes the blasting design parameters ④ and the corresponding tunnel excavation method.

[0015] Step 2.2.2: A tunnel excavation method selection model is constructed based on a neural network, and the sample library is substituted into the tunnel excavation method selection model for training.

[0016] The deep neural network of the tunnel excavation method selection model is set to have 5 layers, i.e., an input layer, 3 hidden layers, and an output layer. The input layer corresponds to the parameters in the blasting design parameters ④, the output layer corresponds to the tunnel excavation method, and the hidden layers use the tangent function or the logarithmic function as the transfer function, i.e., the activation function of the hidden layer.

[0017] After the tunnel excavation method selection model is established, the sample library is input to train it. When the squared error between the target value and the actual value is less than the expected value, the trained tunnel excavation method selection model is obtained.

[0018] Step 2.2.3: The corresponding tunnel excavation method is obtained by inputting the corresponding parameters of the blasting design parameters ④ into the trained tunnel excavation method selection model.

[0019] Step 2.3: Based on the blasting design parameters ⑤ and according to the tunnel engineering blasting theory, the blasthole arrangement method and the tunnel blasting parameters are obtained by using deep learning technology and computer modeling technology.

[0020] The blasting design parameters ⑤ include the blasting design parameters ①, the shape and size of the excavation section, the tunnel excavation method, the tunnel engineering requirements, and the mechanical equipment capacity.

[0021] Step 2.3.1: A blasthole arrangement method selection model is constructed based on deep learning technology according to the blasting design parameters ⑤.

[0022] wherein 3 types of blast holes are arranged in the working face: cut hole, auxiliary hole, and peripheral hole;

[0023] Step 2.3.1.1: Constructing a sample library including the blast design parameters (5) and the corresponding different types of blast hole arrangement modes;

[0024] Step 2.3.1.2: Based on the neural network, constructing a blast hole arrangement mode selection model, and substituting the sample library into the blast hole arrangement mode selection model for training.

[0025] The deep neural network of the blast hole arrangement mode selection model is set to have 5 layers, i.e. an input layer, 3 hidden layers, and an output layer; wherein the input layer corresponds to the parameters in the blast design parameters (5), the output layer corresponds to different types of blast hole arrangement modes, and the hidden layers use tangent function or logarithmic function as the transfer function, i.e. the activation function of the hidden layer;

[0026] After the establishment of the blast hole arrangement mode selection model, the sample library is inputted for training, and when the squared error between the target value and the actual value is less than the expectation, the trained blast hole arrangement mode selection model is obtained;

[0027] Step 2.3.1.3: Inputting the corresponding parameters of the blast design parameters (5) into the trained blast hole arrangement mode selection model to obtain the corresponding different types of blast hole arrangement modes;

[0028] Step 2.3.2: Based on the blast design parameters (6), according to the tunnel engineering blasting theory, using computer modeling technology to obtain the tunnel blasting parameters.

[0029] The blast design parameters (6) include blast design parameters, excavation section shape and size, and tunnel engineering requirements;

[0030] The tunnel blasting parameters include blast hole depth, explosive unit consumption, single cycle charge quantity, blast hole spacing, and single hole charge quantity. In the tunnel engineering blasting theory, the formula is used for modeling calculation according to the blast design parameters (6), and Python is used for modeling operation, and the specific process is as follows:

[0031] 1) The data of the blast design parameters (6) is stored in txt;

[0032] 2) Functions are written, including blast hole depth function, explosive unit consumption function, single cycle charge quantity function, blast hole spacing function, and single hole charge quantity function, which are derived from the tunnel engineering blasting theory and passed in parameters;

[0033] 3) Load data and calculate, return the calculation result, which is the tunnel blasting parameters;

[0034] Step 2.4: Initiation sequence and initiation method;

[0035] The initiation sequence and initiation method depend on the selected initiation equipment and the tunnel construction requirements;

[0036] Step 2.5: Based on the tunnel excavation section, the tunnel excavation method, the blast hole arrangement mode, the initiation sequence and initiation method, and the tunnel blasting parameters, an AI drawing software is used to obtain a drill-and-blast tunnel blasting scheme 01.

[0037] Step 3: Based on the blasting design parameters ⑦, a computer deep learning technology is used to construct a drill-and-blast tunnel blasting scheme sample library, and the blasting design parameters ⑦ are input to obtain a corresponding to-be-built drill-and-blast tunnel blasting scheme 02.

[0038] The blasting design parameters ⑦ include the blasting design parameters ① and the tunnel engineering requirements.

[0039] Step 3.1: A sample library is constructed, which includes existing drill-and-blast tunnel blasting schemes that have been built or are under construction and corresponding characteristics;

[0040] Step 3.2: A drill-and-blast tunnel blasting scheme selection model is constructed based on a neural network, and the sample library is substituted into the drill-and-blast tunnel blasting scheme selection model for training.

[0041] The deep neural network of the drill-and-blast tunnel blasting scheme selection model is set to have 5 layers, namely an input layer, 3 hidden layers, and an output layer; the input layer corresponds to the parameters in the blasting design parameters ⑦, the output layer corresponds to the drill-and-blast tunnel blasting scheme, and the hidden layers use tangent functions or logarithmic functions as transfer functions, i.e., the activation functions of the hidden layers;

[0042] After the drill-and-blast tunnel blasting scheme selection model is established, the sample library is input to train it, and when the squared error between the target value and the actual value is less than the expected value, the trained drill-and-blast tunnel blasting scheme selection model is obtained.

[0043] Step 3.3: The blasting design parameters ⑦ of the to-be-built tunnel are input into the trained drill-and-blast tunnel blasting scheme selection model to obtain the corresponding drill-and-blast tunnel blasting scheme 02;

[0044] Step 4: Based on the obtained drill-and-blast tunnel blasting scheme group, i.e., the drill-and-blast tunnel blasting scheme 01 and the drill-and-blast tunnel blasting scheme 02, an artificial selection method is used to select the final drill-and-blast tunnel blasting scheme.

[0045] The artificial selection selects the drill-and-blast tunnel blasting scheme according to the principles of economy, efficiency, safety, and simplicity: the drill-and-blast tunnel blasting scheme group is listed with economic indicators, cyclic footage, safety evaluation indicators, and construction demand indicators to determine the comparison number indicators.

[0046] The comparison index specifically includes ① economic index: total cost required by tunneling unit size; ② cycle footage: how many meters are excavated in each tunneling cycle, and then the construction period of the entire tunnel is obtained ③ safety evaluation index: probability of inducing danger in one tunneling cycle; and ④ construction demand: whether the construction is simple and convenient.

[0047] If the drilling and blasting method tunnel blasting scheme 01 is consistent with the drilling and blasting method tunnel blasting scheme 02, the final drilling and blasting method tunnel blasting scheme is directly output; if not, the final drilling and blasting method tunnel blasting scheme is selected by the artificial selection method according to the principles of high economic benefit, fast tunneling speed, good safety performance and simple construction.

[0048] The beneficial effects produced by the above technical scheme are as follows:

[0049] The present application provides a kind of tunnel blasting scheme intelligent design method. The tunnel blasting scheme intelligent design method obtains blasting design parameters based on tunnel engineering geological exploration report and forepoling geological prediction;Based on blasting design parameters ④, using tunnel engineering blasting theory and computer deep learning technology, drilling and blasting method tunnel blasting scheme 01 is obtained;Based on blasting design parameters ⑦, using computer deep learning technology, establish the existing drilling and blasting method tunnel blasting scheme sample library of built or under construction, input blasting design parameters, tunnel engineering requirements to obtain the corresponding to be built drilling and blasting method tunnel blasting scheme 02;Based on the obtained drilling and blasting method tunnel blasting scheme group, the final drilling and blasting method tunnel blasting scheme is selected by the artificial selection method according to the principles of economy, efficiency, safety and simplicity. The method avoids the influence of subjective factors, combines traditional experience design method, big data processing and artificial intelligence technology, and proposes an intelligent optimization design scheme for tunnel blasting, which can provide an efficient scheme for designers and construction personnel. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 It is the flow chart of the tunnel blasting scheme intelligent design method provided by the embodiment of the present application;

[0051] Figure 2 It is the flow chart of the blasting scheme designed by using tunnel engineering blasting theory provided by the embodiment of the present application;

[0052] Figure 3 It is the tunnel excavation method selection model schematic diagram provided by the embodiment of the present application based on tunnel engineering blasting theory by using deep neural network technology;

[0053] Figure 4 It is the tunnel excavation method selection flow chart provided by the embodiment of the present application;

[0054] Figure 5is a borehole arrangement mode selection model schematic diagram provided by the embodiment of the present application, which utilizes deep neural network technology and is based on a tunnel engineering blasting theory;

[0055] Figure 6 is a borehole arrangement mode selection flowchart provided by the embodiment of the present application;

[0056] Figure 7 is a tunnel blasting parameter flowchart provided by the embodiment of the present application;

[0057] Figure 8 is a tunnel blasting scheme selection model schematic diagram provided by the embodiment of the present application, which utilizes deep neural network technology and is based on an existing drill-and-blast method tunnel blasting scheme library;

[0058] Figure 9 is a drill-and-blast method tunnel blasting scheme selection model schematic diagram provided by the embodiment of the present application;

[0059] Figure 10 is a drill-and-blast method tunnel blasting scheme group selection flowchart provided by the embodiment of the present application;

[0060] Figure 11 is a bench method tunnel blasting scheme schematic diagram provided by the embodiment of the present application. DETAILED DESCRIPTION

[0061] The specific implementation of the present application will be further described in detail below in combination with the drawings and examples. The following examples are used to illustrate the present application, but are not used to limit the scope of the present application.

[0062] An intelligent tunnel blasting scheme design method, Figure 1 is a flowchart schematic diagram of the intelligent tunnel blasting scheme design method of the embodiment. As shown in Figure 1 , the following steps are included:

[0063] Step 1: Based on the engineering geological exploration report of the tunnel and the advanced geological prediction of the tunnel face, the blasting design parameters ① are obtained;

[0064] The blasting design parameters ① are blasting design parameters; which include the burial depth of the tunnel blasting scheme design section, the pile number range, the surrounding rock grade, the surrounding rock parameters, the stratum parameters, the geological structure, the ground stress distribution, and the hydrogeological characteristics. In the subsequent steps, new different blasting design parameters ②, ③, etc. will be formed due to the tunnel engineering requirements, the mechanical equipment capacity, the tunnel excavation section, the support conditions, the construction period requirements, etc. These parameters are all prepared for designing the drill-and-blast method tunnel blasting scheme, and therefore can be summarized as blasting design parameters.

[0065] The engineering geological exploration report and the tunnel face advance geological prediction, the engineering geological exploration report is provided by the design institute when the tunnel is not excavated, and the tunnel face advance geological prediction is provided by the tunnel construction unit when the tunnel is excavated; when the tunnel is not excavated, the blasting design parameter ② obtained from the engineering geological exploration report is used to assist the design of the tunnel engineering design unit; when the tunnel is excavated, the blasting design parameter ③ obtained from the tunnel face advance geological prediction is used to optimize the blasting scheme for the tunnel face of the tunnel engineering construction unit, improve the blasting efficiency and safety, if the parameters of the blasting design parameter ② and the blasting design parameter ③ are the same, the parameter does not change, that is, the blasting design parameter ① is the blasting design parameter ② obtained from the engineering geological exploration report of the tunnel. The parameters are different, and the blasting design parameter ① is the blasting design parameter ③ obtained from the tunnel engineering construction unit by using the tunnel face advance geological prediction.

[0066] The blasting design parameter ② obtained from the engineering geological exploration report is formed in the early design stage when the tunnel is not excavated, and may be inconsistent with the blasting design parameter ③ obtained from the tunnel engineering construction unit by using the tunnel face advance geological prediction, because the geological conditions encountered in the tunnel excavation process are complex, the engineering geological exploration report is formed by using remote sensing interpretation, geological mapping, drilling, in-situ testing, geophysical prospecting (ground geophysical prospecting, aerial geophysical prospecting) and indoor test, and the blasting design parameter ③ obtained from the tunnel engineering construction unit by using the tunnel face advance geological prediction is obtained by using geological investigation method, TSP (Tunnel Seismic Prediction ahead) seismic wave method, geological radar method or advanced drilling and deepening blast hole method, etc. When the construction unit adopts full computer three-arm rock drilling jumbo for advanced drilling and deepening blast hole method, the surrounding rock conditions are obtained. The former is comprehensive and general, and the latter is specific and real-time.

[0067] Step 2: Based on the blasting design parameter ④, the tunnel engineering blasting theory and computer deep learning technology are used to obtain the tunnel blasting scheme 01.

[0068] First, the principle of neural network is introduced. Neural network is a new discipline after the wide application of computer. It is a complex calculation method inspired by the human brain neural system and simulating the structure of human brain neurons and neural connections. Neural network technology mainly takes the working process of human nerve as the design model and calculates by using the method of human nerve processing related content. Neural network does not need to determine the mathematical equation of the mapping relationship between input and output in advance, but learns a certain rule through its own training to get the result closest to the expected output value when given input value.

[0069] The neural network is usually composed of an input layer, a hidden layer and an output layer, and is fully interconnected between layers, and the nodes in each layer are not connected, the hidden layer can have multiple, which through constantly self-repeated inference remodeling neural network, and then get the final result.

[0070] The blasting design parameters (IV) include the blasting design parameters (I), the tunnel engineering requirements and the mechanical equipment capacity.

[0071] The drill-and-blast method tunnel blasting scheme 01 includes a construction section view, a blast hole layout view, a slotting hole layout view and a blasting parameter table; as shown in Figure 11

[0072] Step 2.1: According to the tunnel engineering requirements, the tunnel excavation section, i.e. the shape and size of the tunnel face, is obtained; for example, the highway tunnel excavation section includes the highway construction limit and the space required for the auxiliary equipment such as ventilation pipes, lighting equipment, disaster prevention equipment, monitoring equipment, operation and management equipment, as well as the allowance amount, and the construction allowable error.

[0073] Step 2.2: According to the blasting design parameters (IV), a tunnel excavation method selection model is constructed based on deep learning technology;

[0074] Step 2.2.1: A sample library is constructed, which includes the blasting design parameters (IV) and the corresponding tunnel excavation method;

[0075] In a specific embodiment, the blasting design parameters (IV) include the surrounding rock grade, the shape and size of the tunnel excavation section, the support condition, the construction period requirement, the work area length, the mechanical equipment capacity and the economy. The tunnel excavation method mainly includes the full-face excavation method, the bench method and the partial excavation method. The specific tunnel excavation method is in one-to-one correspondence with the set of blasting design parameters (IV).

[0076] Step 2.2.2: A tunnel excavation method selection model is constructed based on a neural network, and the sample library is substituted into the tunnel excavation method selection model for training;

[0077] The deep neural network of the tunnel excavation method selection model is set to have 5 layers, i.e. an input layer, 3 hidden layers and an output layer; wherein the input layer corresponds to 7 parameters in the blasting design parameters (IV), i.e. the surrounding rock grade, the shape and size of the tunnel section, the support condition, the construction period requirement, the work area length, the mechanical equipment capacity and the economy, and the characteristics thereof are selected as parameters, such as the surrounding rock grade, which is divided into 6 types: I, II, III, IV, V and VI grade surrounding rock, so there are 7 nodes, and the output layer corresponds to the tunnel excavation method, so there is 1 node, and the number of hidden layer nodes is determined through research, and the tangent function or the logarithmic function is selected as the transfer function, i.e. the activation function of the hidden layer, as shown in Figure 3

[0078] ​​After the tunnel excavation method selection model is established, the sample library is input to train it, and when the square error of the target value and the actual value is less than the expectation, the trained tunnel excavation method selection model is obtained.

[0079] Through a large amount of sample library data training, and the tunnel excavation method selection model has a self-learning function, with the increase of the sample library data generated in the construction process, the accuracy is continuously improved, and has the characteristics of high accuracy and high intelligence. In specific implementation, the data of the sample library is not less than 300 groups. In principle, the sample quantity of each tunnel excavation method should not be less than 100. For example: full-face excavation method, bench method, and partial excavation method, each 100 groups. It should be noted that if the tunnel engineering demand, the above three tunnel excavation methods need to be further divided, and the sample library data needs to be increased accordingly, such as when the bench method is subdivided into long bench method, short bench method and micro bench method, 200 groups of sample library data need to be added accordingly.

[0080] Step 2.2.3: Input the blasting design parameter ④ corresponding parameter into the trained tunnel excavation method selection model, and obtain the corresponding tunnel excavation method in this embodiment, as shown in Figure 4 .

[0081] The case of the present application takes the bench excavation method as an example.

[0082] Step 2.3: Based on the blasting design parameter ⑤, according to the tunnel engineering blasting theory, the depth learning technology and computer modeling technology are used to obtain the blast hole arrangement mode and the tunnel blasting parameter.

[0083] The blasting design parameter ⑤ includes the blasting design parameter ①, the excavation section shape and size, the tunnel excavation method, the tunnel engineering requirement, and the mechanical equipment capacity;

[0084] According to the previously obtained blasting design parameter ①, the excavation section shape and size, the tunnel excavation method, the excavation cycle footage, the drilling instrument and the blasting equipment, the tunnel blasting design is made to determine the blast hole arrangement mode and the tunnel blasting parameter, wherein the tunnel blasting parameter includes the number of blast holes, the depth and angle, the charge quantity and the charge structure, and the initiation sequence;

[0085] Step 2.3.1: According to the blasting design parameter ⑤, a blast hole arrangement mode selection model is constructed based on the depth learning technology;

[0086] Three types of blast holes are arranged on the working face: cutting hole, auxiliary hole and peripheral hole; wherein the auxiliary hole and the peripheral hole are sometimes subdivided according to the blasting demand, such as the peripheral hole can be divided into floor hole and peripheral hole, as shown in Figure 11 .

[0087] Step 2.3.1.1: Constructing a sample library including the blasting design parameters (5) and corresponding different kinds of hole arrangement patterns;

[0088] In specific embodiments, the blasting design parameters (5) include the surrounding rock grade, the tunnel excavation section shape and size, the support condition, and the tunnel excavation method. Different kinds of hole arrangement patterns mainly include ① the hole inclination of the cut hole, the straight cut hole, the compound cut hole, and the hole inclination angle; ② the straight line type, the polygonal type, and the arc type of the auxiliary hole; and ③ the smooth blasting and the pre-splitting blasting of the peripheral hole and the hole inclination angle. The specific hole arrangement pattern is in one-to-one correspondence with the set of blasting design parameters (5).

[0089] Step 2.3.1.2: Based on the neural network, a hole arrangement pattern selection model is constructed, and the sample library is substituted into the hole arrangement pattern selection model for training.

[0090] The deep neural network of the hole arrangement pattern selection model is set to have 5 layers, i.e., an input layer, 3 hidden layers, and an output layer. The input layer corresponds to 4 parameters in the blasting design parameters (5), i.e., the surrounding rock grade, the tunnel excavation section shape and size, the support condition, and the tunnel excavation method, and the features thereof are selected as parameters. For example, the surrounding rock grade is divided into 6 grades: I, II, III, IV, V, and VI grade surrounding rock, so there are 4 nodes. The output layer corresponds to different kinds of hole arrangement patterns, so there are 3 nodes. The number of hidden layer nodes is determined through research, and the tangent function or the logarithmic function is selected as the transfer function, i.e., the activation function of the hidden layer, as shown in the following formula: Figure 5 .

[0091] After the hole arrangement pattern selection model is established, the sample library is input to train it, and when the squared error between the target value and the actual value is less than the expectation, the trained hole arrangement pattern selection model is obtained.

[0092] Through a large amount of sample library data training, the hole arrangement pattern selection model has a self-learning function, and as the sample library data generated during the construction process increases, the accuracy continuously improves, and the model has the characteristics of high accuracy and high intelligence. In specific implementation, the data of the sample library should not be less than 300 groups. In principle, the number of samples selected for each hole arrangement pattern should not be less than 100. For example: the arrangement patterns of the cut hole, the auxiliary hole, and the peripheral hole are each 100 groups. It should be noted that when different kinds of hole arrangement patterns need to be further subdivided into different kinds, the corresponding number of sample library data should also be increased accordingly.

[0093] Step 2.3.1.3: The corresponding parameters of the blasting design parameters (5), i.e., the above-mentioned 4 parameters, are input to the trained hole arrangement pattern selection model to obtain the corresponding different kinds of hole arrangement patterns, as shown in the following table: Figure 6 .

[0094] Step 2.3.2: Based on the blasting design parameters ⑥, according to the tunnel engineering blasting theory, the tunnel blasting parameters are obtained by using computer modeling technology.

[0095] The blasting design parameters ⑥ include blasting design parameters, excavation section shape and size, and tunnel engineering requirements.

[0096] The tunnel blasting parameters include hole depth, explosive unit consumption, single cycle charge quantity, hole spacing, and single hole charge quantity. In the tunnel engineering blasting theory, the formula is used for modeling calculation according to the blasting design parameters ⑥. Python, Excel, Matlab, and other computer tools can be used for parameter modeling calculation. In this embodiment, Python is used for modeling operation, and the specific process is as follows:

[0097] 1) The data of the blasting design parameters ⑥ are stored in txt.

[0098] 2) Functions are written, including hole depth function, explosive unit consumption function, single cycle charge quantity function, hole spacing function, and single hole charge quantity function. The functions are derived from the tunnel engineering blasting theory and the parameters are passed in.

[0099] 3) Load the data and calculate, return the calculation result, and the result is the tunnel blasting parameter.

[0100] For example, the explosive unit consumption is calculated according to the following formula.

[0101]

[0102] In the formula, k is the explosive unit consumption, kg / m 3 ; f is the rock coefficient, f = R / 10, R is the uniaxial compressive strength of rock, MPa; S is the excavation section area, m 2 .

[0103] In some specific embodiments, the blasting design parameters ⑥ requirements include tunnel excavation section width, excavation section area, rock uniaxial compressive strength, and surrounding rock classification.

[0104] The tunnel blasting parameters include hole depth, explosive unit consumption, single cycle charge quantity, hole spacing, and single hole charge quantity. It should be noted that the blasting parameter calculation methods of the cutting hole, auxiliary hole, and peripheral hole are not completely consistent, and attention should be paid to the difference when writing the formula.

[0105] The corresponding tunnel blasting parameters are obtained by inputting the blasting design parameters ⑥. As shown in the following table. Figure 7

[0106] Step 2.4: Initiation sequence and initiation method.

[0107] ​The initiation sequence and initiation method depend on the selected initiation equipment and the tunnel construction requirements. In this case, a non-electrically guided detonation tube millisecond delay initiation system is used, and the controlled blasting method of the peripheral holes will be selected according to the tunnel construction requirements. Generally, smooth blasting is used, and the peripheral holes are initiated last. In this case, the controlled blasting method of the peripheral holes uses smooth blasting. The holes are initiated in the order of the cut hole, the auxiliary hole, the floor hole, and the peripheral hole. Figure 11 The values marked around the blast hole in the figure are the segment numbers of the detonators. If pre-splitting blasting is used, the peripheral holes are initiated first.

[0108] Step 2.5: Based on the tunnel excavation section, tunnel excavation method, blast hole arrangement, initiation sequence and initiation method, and tunnel blasting parameters, use AI drawing software to obtain the drill-and-blast tunnel blasting scheme 01. As shown in Figure 2 .

[0109] Step 3: Based on the blasting design parameters ⑦, use computer deep learning technology to build a drill-and-blast tunnel blasting scheme sample library, and input the blasting design parameters ⑦ to obtain the corresponding to-be-built drill-and-blast tunnel blasting scheme 02.

[0110] The blasting design parameters ⑦ include the blasting design parameters ① and the tunnel engineering requirements.

[0111] Step 3.1: Build a sample library, which includes existing drill-and-blast tunnel blasting schemes that have been built or are under construction and corresponding characteristics.

[0112] In specific embodiments, the existing drill-and-blast tunnel blasting schemes that have been built or are under construction include construction section drawings, blast hole arrangement drawings, cut hole arrangement three-view drawings, and blasting parameter tables. The corresponding characteristics mainly include the surrounding rock grade of the drill-and-blast tunnel using this blasting scheme, the shape and size of the tunnel excavation section, the support conditions, the construction period requirements, the work area length, the mechanical equipment capacity, and the economy. The specific drill-and-blast tunnel blasting scheme sample library and the surrounding rock grade of the drill-and-blast tunnel using this blasting scheme, the shape and size of the tunnel excavation section, the support conditions, the construction period requirements, the work area length, the mechanical equipment capacity, and the economy are in one-to-one correspondence.

[0113] Step 3.2: Based on the neural network, build a drill-and-blast tunnel blasting scheme selection model, and input the sample library into the drill-and-blast tunnel blasting scheme selection model for training.

[0114] The depth neural network of the tunnel blasting scheme selection model of the drill-and-blast method is set to have 5 layers, i.e., an input layer, 3 hidden layers and an output layer; the input layer corresponds to the 7 parameters in the blasting design parameter ⑦, i.e., the surrounding rock grade, the tunnel cross-section shape and size, the support condition, the construction period requirement, the work area length, the mechanical equipment capacity and the economy, and the features thereof are selected as the parameters, for example, the surrounding rock grade is divided into 6 grades: I, II, III, IV, V and VI grade surrounding rock, so there are 7 nodes, and the output layer corresponds to the tunnel blasting scheme of the drill-and-blast method, so there is 1 node, and the number of nodes in the hidden layer is determined through research. The tangent function or the logarithmic function is selected as the transfer function, i.e., the activation function of the hidden layer, as shown in Figure 8 .

[0115] After the tunnel blasting scheme selection model of the drill-and-blast method is established, the sample library is input to train the model, and when the squared error of the target value and the actual value is less than the expectation, the trained tunnel blasting scheme selection model of the drill-and-blast method is obtained.

[0116] Through a large amount of sample library data training, the tunnel blasting scheme selection model of the drill-and-blast method has a self-learning function, and as the sample library data generated in the construction process increases, the accuracy is continuously improved, and has the characteristics of high accuracy and high intelligence. In specific implementation, the data of the sample library is not less than 100 groups.

[0117] Step 3.3: input the above-mentioned 7 parameters in the blasting design parameter ⑦ of the tunnel to be built into the trained tunnel blasting scheme selection model of the drill-and-blast method to obtain the corresponding tunnel blasting scheme 02 of the drill-and-blast method. As shown in Figure 9 .

[0118] Step 4: based on the obtained tunnel blasting scheme group of the drill-and-blast method, i.e., the tunnel blasting scheme 01 of the drill-and-blast method and the tunnel blasting scheme 02 of the drill-and-blast method, the final tunnel blasting scheme of the drill-and-blast method is selected by using the artificial selection method.

[0119] In some specific implementation cases, the tunnel blasting scheme of the drill-and-blast method is selected according to the principles of economy, efficiency, safety and simplicity.

[0120] The tunnel blasting scheme group of the drill-and-blast method is listed with economic indicators, cycle footage, safety evaluation indicators and construction demand indicators respectively to determine the comparison number indicators.

[0121] The comparison number indicators specifically include ① economic indicators: total cost required for tunnel excavation per unit size; ② cycle footage: how many meters are excavated in each excavation cycle, and then the construction period of the entire tunnel excavation is obtained ③ safety evaluation indicators: the probability of inducing danger in one excavation cycle, such as the risk of inducing rock burst ④ construction demand: whether the construction is simple and convenient.

[0122] If the drilling and blasting tunnel blasting scheme 01 is consistent with the drilling and blasting tunnel blasting scheme 02, the final drilling and blasting tunnel blasting scheme is directly output; if not, the final drilling and blasting tunnel blasting scheme is selected by manual selection according to the principles of high economic benefit, fast tunneling speed, good safety performance and simple construction. Figure 10 , Figure 11 as shown in FIGS.

[0123] The above description is only the preferred embodiments of the present disclosure and the explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or equivalent features without departing from the above inventive concept. For example, the above features are replaced with the technical features disclosed in the embodiments of the present disclosure (but not limited to) having similar functions to form technical solutions.

Claims

1. A method for intelligent design of a tunnel blasting scheme, characterized in that, The method comprises the following steps: Step 1: obtaining the blasting design parameter ① based on the tunnel-based engineering geological exploration report and the tunnel face advanced geological prediction; Step 2: obtaining the drill-and-blast method tunnel blasting scheme 01 based on the blasting design parameter ④ by using the tunnel engineering blasting theory and the computer deep learning technology; Step 3: constructing the drill-and-blast method tunnel blasting scheme sample library by using the computer deep learning technology based on the blasting design parameter ⑦, and inputting the blasting design parameter ⑦ to obtain the corresponding to-be-built drill-and-blast method tunnel blasting scheme 02; Step 4: obtaining the final drill-and-blast method tunnel blasting scheme by using the artificial selection method based on the obtained drill-and-blast method tunnel blasting scheme group, i.e. the drill-and-blast method tunnel blasting scheme 01 and the drill-and-blast method tunnel blasting scheme 02; The blasting design parameter ① in step 1 is the blasting design parameter; wherein it includes the burial depth of the tunnel blasting scheme design section, the pile number range, the surrounding rock grade, the surrounding rock parameter, the stratum parameter, the geological structure, the ground stress distribution, and the hydrogeological characteristics; The engineering geological exploration report and the tunnel face advanced geological prediction are provided by the design institute when the tunnel is not excavated, and are provided by the tunnel construction unit when the tunnel is excavated; When the tunnel is not excavated, the blasting design parameter ② obtained by using the engineering geological exploration report is used for auxiliary design for the tunnel engineering design unit; when the tunnel is excavated, the blasting design parameter ③ obtained by using the tunnel face advanced geological prediction is used for optimization of the blasting scheme for the tunnel engineering construction unit according to the actual situation of the tunnel tunnel face, so as to improve the blasting efficiency and safety; if the parameters of the blasting design parameter ② and the blasting design parameter ③ are the same, the parameter is not changed, i.e. the blasting design parameter ① is the blasting design parameter ② obtained by the engineering geological exploration report of the tunnel; if the parameters are different, the blasting design parameter ① is the blasting design parameter ③ obtained by the tunnel engineering construction unit by using the tunnel face advanced geological prediction; The blasting design parameter ④ in step 2 includes the blasting design parameter ①, the tunnel engineering requirement, and the mechanical equipment capacity; The drill-and-blast method tunnel blasting scheme 01 includes the construction section view, the blast hole arrangement view, the slot hole arrangement view, and the blasting parameter table; Step 2 specifically comprises the following steps: Step 2.1: obtaining the tunnel excavation section, i.e. the tunnel face shape and size, according to the tunnel engineering requirement; Step 2.2: constructing the tunnel tunneling excavation method selection model based on the deep learning technology according to the blasting design parameter ④; Step 2.2.1: constructing the sample library, which includes the blasting design parameter ④ and the corresponding tunnel tunneling excavation method; Step 2.2.2: constructing the tunnel tunneling excavation method selection model based on the neural network, and inputting the sample library into the tunnel tunneling excavation method selection model for training; The deep neural network of the tunnel tunneling excavation method selection model is set to have 5 layers, i.e. an input layer, 3 hidden layers, and an output layer; wherein the input layer corresponds to the parameters in the blasting design parameter ④, the output layer corresponds to the tunnel tunneling excavation method, and the hidden layer uses the tangent function or the logarithmic function as the transfer function, i.e. the activation function of the hidden layer; After the tunnel excavation method selection model is established, the sample library is input to train it, and when the squared error of the target value and the actual value is less than the expectation, the trained tunnel excavation method selection model is obtained; Step 2.2.3: Input the corresponding parameters of the blasting design parameters (IV) into the trained tunnel excavation method selection model to obtain the corresponding tunnel excavation method; Step 2.3: Based on the blasting design parameters (V), according to the tunnel engineering blasting theory, the deep learning technology and computer modeling technology are used to obtain the blast hole arrangement mode and the tunnel blasting parameters; The blasting design parameters (V) include the blasting design parameters (I), the excavation section shape and size, the tunnel excavation method, the tunnel engineering requirements, and the mechanical equipment capacity; Step 2.3.1: According to the blasting design parameters (V), a blast hole arrangement mode selection model is constructed based on deep learning technology; Among them, 3 types of blast holes are arranged on the working face: cut hole, auxiliary hole, and peripheral hole; Step 2.3.1.1: A sample library is constructed, which includes the blasting design parameters (V) and the corresponding different types of blast hole arrangement modes; Step 2.3.1.2: A blast hole arrangement mode selection model is constructed based on neural network, and the sample library is substituted into the blast hole arrangement mode selection model for training; The deep neural network of the blast hole arrangement mode selection model is set to have 5 layers, namely an input layer, 3 hidden layers, and an output layer; wherein the input layer corresponds to the parameters in the blasting design parameters (V), and the output layer corresponds to different types of blast hole arrangement modes, and the hidden layer uses tangent function or logarithmic function as the transfer function, i.e. the activation function of the hidden layer; After the blast hole arrangement mode selection model is established, the sample library is input to train it, and when the squared error of the target value and the actual value is less than the expectation, the trained blast hole arrangement mode selection model is obtained; Step 2.3.1.3: Input the corresponding parameters of the blasting design parameters (V) into the trained blast hole arrangement mode selection model to obtain the corresponding different types of blast hole arrangement modes; Step 2.3.2: Based on the blasting design parameters (V), according to the tunnel engineering blasting theory, computer modeling technology is used to obtain the tunnel blasting parameters; The blasting design parameters (V) include the blasting design parameters (I), the excavation section shape and size, and the tunnel engineering requirements; The tunnel blasting parameters include blast hole depth, explosive unit consumption, single cycle charge quantity, blast hole spacing, and single hole charge quantity; in the tunnel engineering blasting theory, the formula is used for modeling calculation according to the blasting design parameters (V), and Python is used for modeling operation, and the specific process is as follows: 1) The data of the blasting design parameters (V) is stored in txt; 2) Functions are written, including blast hole depth function, explosive unit consumption function, single cycle charge quantity function, blast hole spacing function, and single hole charge quantity function, which are derived from the tunnel engineering blasting theory and the input parameters; 3) Load the data and calculate, return the calculation result, the result is the tunnel blasting parameters; Step 2.4: Initiation sequence and initiation method; The initiation sequence and initiation method are based on the selected initiation equipment and tunnel construction requirements; Step 2.5: Based on the tunnel excavation section, tunnel excavation method, blast hole arrangement, blasting sequence and method, and tunnel blasting parameters, an AI drawing software is used to obtain a drilling and blasting method tunnel blasting scheme 01; The blasting design parameters ⑦ in step 3 include the blasting design parameters ① and the tunnel engineering requirements; Step 3 specifically includes the following steps: Step 3.1: Construct a sample library including existing drilling and blasting method tunnel blasting schemes and corresponding characteristics of the built or under-construction tunnels; Step 3.2: Based on a neural network, a drilling and blasting method tunnel blasting scheme selection model is constructed, and the sample library is substituted into the drilling and blasting method tunnel blasting scheme selection model for training; The deep neural network of the drilling and blasting method tunnel blasting scheme selection model is set to have 5 layers, i.e., an input layer, 3 hidden layers, and an output layer; the input layer corresponds to the parameters in the blasting design parameters ⑦, the output layer corresponds to the drilling and blasting method tunnel blasting scheme, and the hidden layers use tangent function or logarithmic function as the transfer function, i.e., the activation function of the hidden layer; After the drilling and blasting method tunnel blasting scheme selection model is established, the sample library is inputted for training, and when the squared error between the target value and the actual value is less than the expected value, the trained drilling and blasting method tunnel blasting scheme selection model is obtained; Step 3.3: The drilling and blasting method tunnel blasting scheme 02 corresponding to the drilling and blasting method tunnel blasting scheme selection model is obtained by inputting the parameters in the blasting design parameters ⑦ of the tunnel to be built into the trained drilling and blasting method tunnel blasting scheme selection model; In step 4, the artificial selection of the drilling and blasting method tunnel blasting scheme is based on the principles of economy, efficiency, safety, and simplicity: the drilling and blasting method tunnel blasting scheme group is listed with economic indicators, cycle footage, safety evaluation indicators, and construction demand indicators to determine the comparison number indicators; The comparison number indicators specifically include ① economic indicators: total cost required for tunnel excavation per unit size; ② cycle footage: how many meters are excavated per excavation cycle, and then the construction period of the entire tunnel is obtained; ③ safety evaluation indicators: the probability of inducing danger during an excavation cycle; and ④ construction demand: whether the construction is simple and convenient; If the drilling and blasting method tunnel blasting scheme 01 and the drilling and blasting method tunnel blasting scheme 02 are consistent, the final drilling and blasting method tunnel blasting scheme is directly outputted; if they are not consistent, the final drilling and blasting method tunnel blasting scheme is selected by comparing the indicators based on the principles of high economic efficiency, fast excavation speed, good safety performance, and simple construction.

Images