Optimization method of tunnel blasting parameters based on the concave-convex morphological distribution characteristics of surrounding rock after blasting

Abstract

The present invention belongs to the field of underground engineering technology and relates to a method for optimizing tunnel blasting parameters based on the distribution characteristics of the concave-convex morphology of the surrounding rock after blasting. The method comprises: S1 collecting images of the concave-convex morphology of the surrounding rock after tunnel blasting construction; S2 inputting the collected images into a computer neural network, outputting the types of rock depressions and protrusions corresponding to the images, and classifying the images according to the types; using the YOLO algorithm to identify the classified image data, obtain coordinate parameters of the concave-convex morphology in the image, and determine their distribution positions in the tunnel; S3 judging the type of concave-convex morphological distribution characteristics based on the classification and location of the rock depressions and protrusions; S4 adjusting blasting parameters based on the type of concave-convex morphological distribution characteristics. The method of the present invention is conducive to quickly determining the main reasons for poor tunnel blasting results and adjusting the corresponding blasting parameters in real time, effectively improving the efficiency and accuracy of tunnel blasting parameter optimization.

Description

Technical Field

[0001] The present application relates to the field of underground engineering technology, and in particular to a method for optimizing tunnel blasting parameters based on the concave-convex morphological distribution characteristics of surrounding rocks after blasting. Background Art

[0002] During tunnel construction, face blasting and excavation have a significant impact on construction quality and cost. Current tunnel construction technical specifications generally recommend manually observing the number of blasthole scars and calculating the blasthole scar rate after each blast to assess the blasting performance of that cycle. This is supplemented by on-site measurements of tunnel profile over- and underbreak data to adjust blasting parameters for the next cycle. However, relying solely on the blasthole scar rate to assess tunnel blasting performance is a one-sided metric, making it difficult to accurately guide blasting parameter optimization. Furthermore, tunnel profile over- and underbreak data also poses a problem in that they cannot clearly identify the cause of poor tunnel blasting performance. Therefore, in current construction, assessment of tunnel blasting performance and parameter optimization still rely primarily on manual experience, a labor-intensive and low-level approach. In engineering practice, it is difficult to assess tunnel blasting performance for each cycle and to adjust smooth face blasting parameters in real time. This often leads to poor smooth face blasting performance, severe over- and underbreak, and increased tunnel construction costs in projects with complex and variable surrounding rock conditions or limited on-site technical resources. Summary of the Invention

[0003] In order to overcome the above-mentioned defects, the present application optimizes the smooth surface blasting parameters based on the concave-convex morphological distribution characteristics of the tunnel face after tunnel blasting, and provides a tunnel blasting parameter optimization method based on the concave-convex morphological distribution characteristics of the surrounding rock after blasting, which is conducive to quickly determining the main reasons leading to poor tunnel blasting effects, thereby improving the efficiency and accuracy of tunnel blasting parameter optimization, achieving better results in smooth surface blasting, and promoting the intelligent development of tunnel blasting design and construction.

[0004] The method for optimizing smooth blasting parameters based on the concave-convex morphological distribution characteristics of the surrounding rock provided in the embodiment of the present application includes:

[0005] S1, collect images of the surrounding rock concave-convex morphology after tunnel blasting construction;

[0006] S2. Input the collected images into a computer neural network, output the types of rock depressions and protrusions corresponding to the images, and classify the images according to the types. Use the YOLO (You Only Look Once) algorithm to identify the classified images, obtain the coordinate parameters of the concave and convex shapes in the images, and determine the distribution of the images in the tunnel.

[0007] S3. Determine the type of concave-convex morphological distribution characteristics of the image at the tunnel distribution location based on the types of rock depressions and convexities corresponding to the image;

[0008] S4. Adjust blasting parameters according to the distribution characteristics of the concave and convex morphology.

[0009] Furthermore, the computer neural network in step S2 includes a VGG16 convolutional neural network.

[0010] Furthermore, the types of rock depressions and protrusions in step S2 include: flat rock, rock depression and rock protrusion.

[0011] Furthermore, before using the YOLO algorithm to identify the classified image data, the image data of the rock flat type is deleted.

[0012] Furthermore, the distribution position of the image in the tunnel in step S2 is represented by a bounding box with coordinate information.

[0013] Furthermore, the steps further include: performing blasting processing according to the adjusted blasting parameters, and returning to execute steps S1 to S4.

[0014] Furthermore, the concave-convex morphological distribution feature types in step S3 include:

[0015] Feature 1: The location is the tunnel face, and the surrounding rock shape is concave or convex;

[0016] Feature 2: The location is on the side wall, and the surrounding rock shape is concave;

[0017] Feature 3: The location is on the side wall, and the surrounding rock shape is convex;

[0018] Feature 4: The appearance position is at the contour line, and the surrounding rock shape is convex.

[0019] Furthermore, adjusting the blasting parameters according to the concave-convex morphological distribution characteristic type includes:

[0020] Feature 1: The parameter that needs to be adjusted is the drilling depth, and the adjustment method is to control the drilling inclination angle;

[0021] Feature 2: The parameter that needs to be adjusted is the thickness of the light explosion layer, and the adjustment method is to increase the thickness of the light explosion layer;

[0022] Feature 3: The parameter that needs to be adjusted is the peripheral hole spacing, and the adjustment method is to reduce the peripheral hole spacing;

[0023] Feature 4: The parameter that needs to be adjusted is the charge amount, and the adjustment method is to increase the charge amount in the slot hole.

[0024] Furthermore, the adjustment range is:

[0025] Feature 1: Keep the drilling inclination angle unchanged;

[0026] Feature 2: Increase the thickness of the light explosion layer by 0.1×15×d;

[0027] Feature 3: Reduce the peripheral hole spacing to 0.1×18×d;

[0028] Feature 4: Increase the amount of charge for the slot hole by 0.1kg / m;

[0029] Where d is the diameter of the blasthole in mm.

[0030] Furthermore, the resolution of the image in step S1 is greater than or equal to 20 million pixels.

[0031] Compared with the existing technology, the beneficial effects of this application are: it overcomes the problems in the existing technology that the optimization path of blasting parameters is unclear, the optimization design is more dependent on manual labor, and the degree of intelligence is low. It is conducive to quickly determining the main reasons for poor tunnel blasting effects and adjusting the corresponding blasting parameters, effectively improving the efficiency and accuracy of tunnel blasting parameter optimization, and achieving better results in tunnel blasting design and construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a schematic diagram of the steps of the smooth blasting parameter optimization method based on the concave-convex morphological distribution characteristics of the surrounding rock in Example 1 of the present application;

[0033] Figure 2 Schematic diagram of the method for collecting images of surrounding rock after tunnel blasting in Example 1 of the present application;

[0034] Figure 3 This is a schematic diagram of the VGG16 convolutional neural network model architecture in Example 1 of this application;

[0035] Figure 4 Schematic diagram of the VGG16 feature extraction process in Example 1 of the present application;

[0036] Figure 5 This is a schematic diagram of the specific network structure of the YOLO algorithm in Example 1 of the present application;

[0037] Figure 6 This is the original detection image in Example 1 of the present application;

[0038] Figure 7 For Example 1 of this application Figure 6 The detection results corresponding to the original detection image in . DETAILED DESCRIPTION

[0039] The present application is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as limiting the scope of the above-mentioned subject matter of the present application to the following embodiments. All technologies implemented based on the content of the present application fall within the scope of protection of the present application.

[0040] Unless otherwise specified, in the description of the specific embodiments of this application, the terms indicating the orientation or position relationship such as "up", "down", "left", "right", "center", "inside", "outside", and "side" are based on the expression of the orientation or position relationship shown in the accompanying drawings, or the orientation or position relationship in which the product / device / apparatus is placed when it is usually used. These terms of orientation or position relationship are only for the convenience of describing the scheme of this application or simplifying the description in the specific embodiments to facilitate the technicians to quickly understand the scheme, and do not indicate or imply that a specific device / component / element must have a specific orientation, or be constructed and operated in a specific position relationship, and therefore should not be understood as limiting this application.

[0041] In the description of the embodiments of this application, the technical terms "first," "second," etc., merely distinguish one entity or operation from another and are not to be understood as indicating or implying relative importance or implicitly specifying the quantity, specific order, or primary-secondary relationship of the indicated technical features. In the description of the embodiments of this application, "plurality" means two or more, unless otherwise specifically defined.

[0042] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0043] Please see Figure 1 , Figure 1 Schematic diagram of the steps of the tunnel blasting parameter optimization method based on the uneven morphological distribution characteristics of the surrounding rock after blasting provided in an embodiment of the present application. The tunnel blasting parameter optimization method based on the uneven morphological distribution characteristics of the surrounding rock after blasting may include:

[0044] S1. Collect images of the concave and convex morphology of the surrounding rock after tunnel blasting construction.

[0045] After blasting in the tunnel, the tunnel face is first deslagging and ventilated, and a certain intensity of light is applied to the tunnel face to ensure the brightness of the camera area. Then, a high-resolution camera (resolution ≥ 20 million pixels) placed at a fixed position on the central axis of the tunnel is used to photograph the tunnel face and tunnel sidewalls. Focus is placed on the image acquisition of the locations with concave and convex morphological features of the surrounding rock, and ensure that occlusion and partial loss are eliminated during the image acquisition process to fully cover the tunnel face and the surrounding area. The schematic diagram of the surrounding rock image acquisition method after tunnel blasting is shown in the figure below. Figure 2 shown.

[0046] S2. Input the collected images into a computer neural network and output the classification of rock depressions and protrusions corresponding to the images; use the YOLO algorithm to identify the classified images and obtain the corresponding positions of the classified images.

[0047] Step S2 mainly includes two parts: classification of rock depressions and protrusions, and identification and detection of rock depression and protrusion positions. The following describes the processes in detail.

[0048] Classification of rock depressions and protrusions

[0049] Based on the heterogeneous nature of the tunnel face structure and the complexity of the geological structures in the study area, a geologically representative image database was constructed, covering three types of tunnel face morphologies (surrounding rock morphologies): flat rock, rock depression, and rock protrusion. Images from the image database and the three types of surrounding rock morphologies were pre-trained into a computer neural network to train the neural network for rock depression and protrusion classification. When the measured image data were then fed back into the neural network, the corresponding surrounding rock morphology could be output. During the training phase, the image dataset was divided into training, validation, and test sets in a ratio of 7:2:1. The specific sample distribution is shown in Table 1.

[0050] Table 1 Dataset sample distribution table

[0051]

[0052] As a preferred solution, the computer neural network used to classify rock depressions and protrusions based on image output is the VGG16 convolutional neural network. According to the sample distribution statistics in Table 1, the constructed data set belongs to a typical small-scale training sample scenario. In order to improve the monitoring efficiency and accuracy of surrounding rock morphology, the VGG16 convolutional neural network in deep learning is used to classify the concave and convex morphological characteristics of rocks after tunnel blasting. The VGG16 convolutional neural network consists of a series of convolutional layers to avoid the model being too complex and affecting the generation effect of the generator. The VGG16 convolutional neural network model architecture is as follows: Figure 3The VGG16 convolutional neural network constructs a feature extraction module, which captures the low-level semantic features (edge response, local texture primitives, etc.) of the tunnel surrounding rock image through its first layer of convolution kernel. The schematic diagram of the VGG16 feature extraction process is shown in Figure 4 Its feature extraction has the following advantages: 1. Preservation of spatial resolution, avoiding loss of spatial information of high-level features; 2. Balance between feature discriminability and generation feasibility, including key patterns for face morphology classification while avoiding excessive abstraction of high-level semantics.

[0053] Identification and detection of rock depressions and protrusions

[0054] A method for automatically detecting rock protrusions and depressions after tunnel blasting, based on the YOLO algorithm, was established. By acquiring, processing, and modeling images, YOLO's real-time detection capabilities were leveraged to rapidly identify and locate both rock protrusions and depressions. The YOLO algorithm directly predicted the bounding box positions of both rock depressions and rock protrusions by inputting image data of both types of rock depressions and rock protrusions. During the prediction process, the model simultaneously outputs the classification and coordinates of all bounding boxes, enabling parallel processing of target location and recognition. This mechanism ensures the real-time detection of the YOLO algorithm and improves its performance when detecting small or densely packed objects. Bounding box classification refers to the classification of bounding boxes into two types: rock depressions and rock protrusions. Therefore, the YOLO algorithm detection result displays both the coordinates and the type of the bounding box.

[0055] The YOLO algorithm also requires training before use. This training is performed using images of two types of tunnel face morphologies, including rock depressions and rock protrusions, from a pre-built image database. The image dataset is divided into training, validation, and test sets in a 7:2:1 ratio. The specific sample distribution is detailed in Table 2. Through training, a YOLO algorithm model is developed that outputs image results containing bounding box types and positioning.

[0056] Table 2 Dataset sample distribution table

[0057]

[0058] The YOLO algorithm extracts feature information from the image by stacking convolutional layers. Subsequently, the feature map output by the convolutional layer is processed using a fully connected layer to predict the object category and bounding box coordinates. This network structure design enables the YOLO algorithm to effectively extract key information from the image and achieve fast and accurate object detection. The specific network structure diagram of the YOLO algorithm is shown below. Figure 5As shown, it includes the stacking of multiple convolutional layers and fully connected layers, as well as the use of corresponding activation functions and pooling layers.

[0059] The YOLO algorithm is used to automatically detect rock protrusions and depressions on the tunnel face. The original image is as follows: Figure 6 The corresponding test results are shown in Figure 7 As shown. After the YOLO algorithm automatically detects the rock protrusions or depressions on the tunnel face, it will also output the coordinates of the center point of the bounding box ( x , y ) value.

[0060] S3. According to the types of rock depressions and protrusions corresponding to the image, the distribution characteristics of the concave-convex shape of the image at the tunnel distribution position are determined.

[0061] According to the detected concave and convex shapes and their coordinates, their distribution positions in the tunnel are determined, and the classification of the morphological features corresponding to the image can be obtained according to Table 3.

[0062] Table 3 Classification of distribution characteristics of surrounding rock concave-convex morphology after tunnel blasting

[0063]

[0064] The classification method for the distribution characteristics of the surrounding rock concave-convex morphology after tunnel blasting in Table 3 is as follows:

[0065] Ideally, after smooth blasting, the surrounding rock in the tunnel face area is smooth and flat. However, in actual construction, the surrounding rock will inevitably exhibit uneven morphological characteristics due to the influence of surrounding rock conditions, technical conditions, and construction management conditions. Based on a comprehensive analysis of on-site investigations, theoretical analysis, and expert experience, the distribution of uneven morphology and the causes of the surrounding rock at different locations after smooth blasting on the tunnel face are divided into the following four categories:

[0066] Feature 1: Concavity or convexity of the surrounding rock at the tunnel face;

[0067] It refers to the overall or partial depression or protrusion of the rock on the tunnel face after blasting. It is usually caused by the bottom of the blast hole not being in the same cross-sectional position during construction. The specific reason is generally that during drilling construction, the drill rod is not drilled according to the designed inclination angle, or due to the influence of surrounding rock and footage, the drill rod drift occurs.

[0068] Feature 2: Depression in the surrounding rock at the blasthole scar on the side wall;

[0069] It refers to the groove-like shape of the rock between the blast hole marks on the side wall of the tunnel after blasting. When rock depressions appear between multiple consecutive blast hole marks, the overall contour of the tunnel wall will appear wavy. This is generally caused by the small thickness of the light blast layer.

[0070] Feature 3: The bulge of the surrounding rock at the blasthole scar on the side wall;

[0071] It refers to the convex shape of the rock between the blast hole marks on the side wall of the tunnel after blasting. When rock bulges appear between multiple consecutive blast hole marks, the overall contour of the tunnel wall will have a tile-shaped feature; it is generally caused by the large distance between the surrounding holes.

[0072] Feature 4: bulge of surrounding rock at the contour line;

[0073] It refers to the junction of the tunnel wall and the heading face after blasting, that is, the surrounding rock at the tunnel contour line has local or large-scale bulges, which manifests as insufficient excavation progress and is generally caused by insufficient charging of the slot hole.

[0074] S4. Adjust blasting parameters according to the distribution characteristics of the concave and convex morphology.

[0075] Based on the different distribution characteristics of the surrounding rock concave and convex shapes, the corresponding blasting parameters that need to be adjusted are determined and adjusted. The adjustment parameters and adjustment methods corresponding to the different distribution characteristics of the surrounding rock concave and convex shapes are shown in Table 4, where d is the blasthole diameter.

[0076] Table 4 Blasting parameter adjustment method for different surrounding rock concave-convex morphological distribution characteristics

[0077]

[0078] Preferably, step S5 is further included, in which blasting is performed after adjusting the blasting parameters, and then steps S1 to S4 are executed to obtain a new round of adjusted blasting parameters, and the blasting parameters are repeatedly adjusted according to the concave-convex morphological distribution characteristics of the surrounding rock after the tunnel blasting in the next cycle, until no obvious concave-convex morphological distribution of the surrounding rock is detected, which indicates that the smooth blasting has achieved a better effect.

[0079] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application 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 make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A tunnel blasting parameter optimization method based on the concave-convex morphological distribution characteristics of the surrounding rock after blasting, characterized by: include: S1, collect images of the surrounding rock concave-convex morphology after tunnel blasting construction; S2. Input the collected images into the computer neural network, output the types of rock depressions and protrusions corresponding to the images, and Classify the images according to the type; use the YOLO algorithm to identify the classified images, obtain the coordinate parameters of the concave and convex shapes in the images, and determine the distribution location of the images in the tunnel; The distribution position of the image in the tunnel is represented by a bounding box with coordinate information; The types of rock depressions and protrusions include: flat rock, rock depression and rock protrusion; S3. Determine the concave and convex shape of the rock according to the type of rock concavity and convexity corresponding to the image and the distribution position of the image in the tunnel. Distribution feature type; S4. Adjust blasting parameters according to the distribution characteristics of concave and convex shapes; The concave-convex morphological distribution feature types in step S3 include: Feature 1: The location is the tunnel face, and the surrounding rock shape is concave or convex; Feature 2: The location is on the side wall, and the surrounding rock shape is concave; Feature 3: The location is on the side wall, and the surrounding rock shape is convex; Feature 4: The location of the occurrence is at the contour line, and the surrounding rock shape is convex; The adjusting of blasting parameters according to the concave-convex morphological distribution characteristic type includes: Feature 1: The parameter that needs to be adjusted is the drilling depth, and the adjustment method is to control the drilling inclination angle; Feature 2: The parameter that needs to be adjusted is the thickness of the light explosion layer, and the adjustment method is to increase the thickness of the light explosion layer; Feature 3: The parameter that needs to be adjusted is the peripheral hole spacing, and the adjustment method is to reduce the peripheral hole spacing; Feature 4: The parameter that needs to be adjusted is the charge amount, and the adjustment method is to increase the charge amount in the slot hole.

2. The tunnel blasting parameter optimization method based on the concave-convex morphological distribution characteristics of the surrounding rock after blasting according to claim 1, wherein In, The computer neural network in step S2 includes a VGG16 convolutional neural network.

3. The tunnel blasting parameter optimization method based on the concave-convex morphological distribution characteristics of the surrounding rock after blasting according to claim 1, wherein In, Before using the YOLO algorithm to identify the classified image data, the image data of the rock flat type is deleted.

4. The method for optimizing tunnel blasting parameters based on the concave-convex morphological distribution characteristics of the surrounding rock after blasting according to claim 1, wherein In, The steps further include: performing blasting according to the adjusted blasting parameters, and returning to execute steps S1 to S4.

5. The method for optimizing tunnel blasting parameters based on the concave-convex morphological distribution characteristics of surrounding rock after blasting according to claim 1, characterized in that: The adjustment range is: Feature 1: Keep the drilling inclination angle unchanged; Feature 2: Increase the thickness of the light explosion layer by 0.1×15×d; Feature 3: Reduce the peripheral hole spacing to 0.1×18×d; Feature 4: Increase the amount of charge for the slot hole by 0.1kg / m; Where d is the diameter of the blasthole in mm.

6. The method for optimizing tunnel blasting parameters based on the concave-convex morphological distribution characteristics of surrounding rock after blasting according to claim 1, characterized in that: The resolution of the image in step S1 is greater than or equal to 20 million pixels.

Images