A tunnel primary support construction method based on a concrete wet spraying manipulator

Abstract

The application relates to the technical field of tunnel construction, and discloses a tunnel primary support construction method based on a concrete wet spraying manipulator, which comprises the following steps: path planning is performed on a tunnel to be supported to generate a basic spraying path; feature extraction is performed on real-time image data to obtain texture feature parameters and glossiness feature parameters, and a real-time thickness deviation value of a concrete layer surface is calculated; the feature parameters and the real-time thickness deviation value are input into a state judgment model to obtain a layer combination opportunity coefficient; when the layer combination opportunity coefficient meets a lower spraying starting condition, the real-time thickness deviation value is coupled with a spraying thickness to perform coupling calculation, a dynamic compensation coefficient is obtained, the basic spraying path is corrected in real time, and a secondary spraying path is obtained; next layer spraying work is performed on a sprayed surface, and steps Z2 to Z6 are repeatedly executed until multi-layer spraying support is completed; and the application can improve the automation level and quality control precision of tunnel primary support construction.

Description

Technical Field

[0001] This invention relates to the field of tunnel construction technology, and in particular to a method for initial support construction of tunnels based on a concrete wet spraying robot. Background Technology

[0002] Initial support in tunnel construction is a crucial process for ensuring tunnel safety and structural stability. It typically involves using a wet-spraying robot to spray concrete onto the rock face to form a support layer. The wet-spraying robot mainly consists of a boom system, nozzles, and a control system. Its operational quality directly affects the load-bearing capacity and durability of the support layer. In tunnel construction, initial support often requires multi-layer spraying. The bonding quality between each layer and the uniformity of the spray thickness are core factors determining the support effect, thus placing high demands on the precise control of the spraying process.

[0003] In existing technologies, tunnel initial support construction methods based on concrete wet spraying robots typically rely on manual control through operator visual observation and experience. Operators visually assess the surface condition of the sprayed concrete layer and determine the timing of the next layer's spraying based on experience, while manually adjusting the robot arm's trajectory and nozzle parameters. This manually-driven control method struggles to objectively and quantitatively perceive subtle changes in the concrete surface texture and gloss, leading to a lack of scientific basis for judging the timing of interlayer bonding and making it prone to insufficient bonding strength due to improper interlayer spacing. Secondly, regarding thickness control, it is impossible to obtain real-time information on the thickness distribution of the sprayed layers, making it difficult to accurately compensate for local thickness deviations, resulting in uneven sprayed layer thickness and affecting the overall load-bearing performance of the support structure. Finally, adjustments to the spraying path are usually global parameter modifications, lacking spatial adaptability for local areas, making it difficult to achieve refined construction. Therefore, how to achieve accurate perception of the concrete surface condition, scientific determination of interlayer spraying timing, and adaptive correction of the spraying path has become a pressing technical problem to be solved in this field. Summary of the Invention

[0004] This invention provides a method for tunnel initial support construction based on a concrete wet spraying robot to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for tunnel initial support construction based on a concrete wet spraying robot, comprising: Z1: Obtain the design contour data of the tunnel section to be supported, and perform path planning on the tunnel section to be supported to generate the foundation spraying path; Z2: Real-time acquisition of real-time distance data between the nozzle and the surface of the sprayed concrete layer, as well as real-time image data of the surface of the sprayed concrete layer; Z3: Extract texture features and gloss features from the real-time image data to obtain texture feature parameters and gloss feature parameters, and calculate the real-time thickness deviation value of the sprayed concrete layer surface based on the real-time distance data and the preset spray thickness at the corresponding position in the basic spraying path. Z4: Input the texture feature parameters, gloss feature parameters and the real-time thickness deviation value into the pre-constructed interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface; Z5: When the interlayer bonding timing coefficient meets the lower layer spraying start condition of the tunnel section to be supported, the real-time thickness deviation value is coupled with the preset spraying thickness to obtain the dynamic compensation coefficient of the nozzle, and the unexecuted part in the basic spraying path is corrected in real time based on the dynamic compensation coefficient to obtain the secondary spraying path of the tunnel section to be supported. Z6: Based on the secondary spraying path, perform the next layer of spraying operation on the surface of the sprayed concrete layer, repeating steps Z2 to Z6 until the multi-layer spraying support of the tunnel section to be supported is completed.

[0006] In a preferred embodiment, the step of acquiring the design contour data of the tunnel cross-section to be supported and performing path planning on the tunnel cross-section to generate the foundation spraying path includes: Obtain the design contour data of the tunnel section to be supported, and extract the contour line of the tunnel section to be supported; According to the preset spray width and overlap width, the nozzle movement trajectory of the tunnel section to be supported is constructed, and the nozzle movement trajectory is used as the basic spray path.

[0007] In a preferred embodiment, the step of extracting texture features and gloss features from the real-time image data to obtain texture feature parameters and gloss feature parameters, and calculating the real-time thickness deviation value of the sprayed concrete layer surface based on the real-time distance data and the preset spray thickness at the corresponding position in the base spraying path, includes: The real-time image data is preprocessed, including median filtering for noise reduction and grayscale processing, to obtain a grayscale image of the sprayed concrete layer surface. Texture analysis is performed on the grayscale image to obtain the grayscale co-occurrence matrix of the grayscale image, and the contrast and energy of the grayscale co-occurrence matrix are integrated into the texture feature parameters of the grayscale image; The grayscale image is subjected to highlight region detection. Pixels in the grayscale image whose grayscale value is higher than a preset grayscale threshold are identified as highlight pixels. The total percentage of the highlight pixels in the grayscale image is used as a glossiness feature parameter. Based on the real-time distance data and the preset spraying thickness at the corresponding position in the basic spraying path, the real-time thickness deviation value of the sprayed concrete layer surface is calculated.

[0008] In a preferred embodiment, calculating the real-time thickness deviation value of the sprayed concrete layer surface based on the real-time distance data and the preset spray thickness at the corresponding position in the basic spraying path includes: Based on the normal direction on the nozzle movement trajectory, determine the nozzle attitude angle sequence in which the nozzle axis is consistent with the normal direction; Based on the total spray thickness of the design contour data and the preset number of layered sprays, the preset spray thickness on the nozzle movement trajectory is determined; For the sampling points on the nozzle movement trajectory, the difference between the real-time distance data and the preset injection thickness is used as the point thickness deviation value of the sampling point; The thickness deviation values ​​at the sampling points are integrated according to their spatial locations to generate a thickness deviation distribution map of a local area on the surface of the sprayed concrete layer.

[0009] In a preferred embodiment, the step of inputting the texture feature parameters, gloss feature parameters, and real-time thickness deviation value into a pre-constructed interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface includes: Obtain historical construction data of the tunnel section to be supported. The historical construction data includes historical texture feature parameters, historical gloss feature parameters, historical thickness deviation distribution map, and interlayer bonding strength value. A multivariate nonlinear regression analysis was performed on the historical texture feature parameters, historical gloss feature parameters, historical thickness deviation distribution map, and interlayer bonding strength value to obtain model coefficients; Based on the model coefficients, a model for determining the interlayer bonding state of the tunnel section to be supported is constructed. The texture feature parameters, the gloss feature parameters, and the real-time thickness deviation value are input into the interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface.

[0010] In a preferred embodiment, the formula for calculating the interlayer bonding timing coefficient is: in, This refers to the interlayer bonding timing coefficient. The contrast of the gray-level co-occurrence matrix. The preset contrast reference value, , , , These are the weighting coefficients of the interlayer bonding state determination model. The energy of the gray-level co-occurrence matrix. The preset energy baseline value, The gloss characteristic parameter at the current moment, The preset gloss reference value, The average of the point thickness deviation values ​​for all sampling points. This is the preset thickness deviation tolerance. It is an exponential function.

[0011] In a preferred embodiment, when the interlayer bonding timing coefficient satisfies the lower-layer spraying initiation condition of the tunnel section to be supported, the real-time thickness deviation value is coupled with the preset spraying thickness to obtain the dynamic compensation coefficient of the nozzle, and the unexecuted portion of the basic spraying path is corrected in real time based on the dynamic compensation coefficient to obtain the secondary spraying path of the tunnel section to be supported, including: When the interlayer bonding timing coefficient is greater than or equal to the preset target lower threshold, the lower layer spraying start condition of the tunnel section to be supported is triggered. The dynamic compensation coefficient of the nozzle is obtained by coupling the real-time thickness deviation value with the preset injection thickness. Based on the dynamic compensation coefficient, the unexecuted portion of the basic spraying path is corrected in real time to obtain the secondary spraying path of the tunnel section to be supported.

[0012] In a preferred embodiment, the formula for calculating the dynamic compensation coefficient is: in, For any target location point on the basic injection path, For any point in the thickness deviation distribution map, For point The dynamic compensation coefficient at that location. The preset compensation strength coefficient, For the weight function, For any point in the thickness deviation distribution map The point thickness deviation value, For the point Centered on, with a preset radius The circular neighborhood, For point The preset spray thickness at the location.

[0013] In a preferred embodiment, the step of real-time correction of the unexecuted portion of the basic spraying path based on the dynamic compensation coefficient to obtain the secondary spraying path of the tunnel section to be supported includes: The preset moving speed of each point on the basic injection path is mathematically multiplied by the reciprocal of the dynamic compensation coefficient, and the result of the mathematical product is used as the corrected moving speed of the corresponding point on the secondary injection path. The product of the preset injection flow rate at each point on the basic injection path and the dynamic compensation coefficient is used as the corrected injection flow rate at the corresponding point on the secondary injection path. Keeping the nozzle attitude angle sequence of each point on the basic injection path unchanged, and making real-time corrections to the unexecuted parts of the basic injection path based on the corrected moving speed and the corrected injection flow rate, a secondary injection path for the tunnel section to be supported is obtained.

[0014] In a preferred embodiment, the step of performing the next layer of spraying operation on the surface of the already sprayed concrete layer based on the secondary spraying path, repeating steps Z2 to Z6, until the multi-layer sprayed support of the tunnel section to be supported is completed, includes: After the current spraying operation of the tunnel section to be supported is completed, the end time of the spraying of the current layer in the tunnel section to be supported is taken as the start time of the spraying operation of the next layer, and the process returns to step Z2. The newly completed shotcrete layer in the tunnel section to be supported is used as the new shotcrete layer, and the lower shotcrete operation of the tunnel section to be supported is carried out. When the cumulative spray thickness reaches the total spray thickness required in the design profile data, the repetition stops, and the support of the tunnel section to be supported is completed. Compared with the prior art, the present invention has the following beneficial effects: 1. This invention achieves multi-dimensional quantitative perception of the surface condition of the concrete layer by real-time acquisition of image data and distance data of the sprayed concrete layer, and extracts texture and gloss features from the real-time image data. Simultaneously, it calculates the thickness deviation value by combining the real-time distance data. Based on this, the texture feature parameters, gloss feature parameters, and real-time thickness deviation value are input into a pre-constructed interlayer bonding state determination model. An interlayer bonding timing coefficient is calculated using a nonlinear formula containing an exponential decay term. This transforms the determination of interlayer bonding timing from traditional manual experience-based judgment to scientific quantitative decision-making, avoiding insufficient bonding strength caused by improper interlayer spacing, and significantly improving the interlayer bonding quality and overall structural stability of multi-layer sprayed support.

[0015] 2. This invention generates a thickness deviation distribution map and couples the thickness deviation values ​​at points in the map with the preset spraying thickness in the base spraying path to obtain a dynamic compensation coefficient. Based on this dynamic compensation coefficient, the moving speed and spraying flow rate of the unexecuted portion of the base spraying path are decoupled and corrected to generate a secondary spraying path that matches the actual state of the currently sprayed concrete layer. This spatially adaptive path correction method transforms traditional global parameter adjustment and optimization into localized fine-grained compensation, enabling precise supplementary spraying in areas with insufficient thickness. This avoids the problem of uneven sprayed layer thickness and improves the uniformity and load-bearing capacity of the support layer. Simultaneously, by repeatedly executing state monitoring and adaptive correction until multi-layer sprayed support is completed, the method achieves closed-loop intelligent control of the entire initial tunnel support construction process, providing reliable technical support for digital construction and quality traceability in tunnel engineering. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a method for initial tunnel support construction based on a concrete wet spraying robot, as provided in an embodiment of the present invention. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0018] This application provides a method for tunnel initial support construction based on a concrete wet spraying robot. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for tunnel initial support construction based on a concrete wet spraying robot can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0019] Reference Figure 1The diagram shown is a flowchart illustrating a tunnel initial support construction method based on a wet concrete spraying robot according to an embodiment of the present invention. In this embodiment, the tunnel initial support construction method based on a wet concrete spraying robot includes: Z1: Obtain the design contour data of the tunnel section to be supported, and perform path planning on the tunnel section to be supported to generate the foundation spraying path; In this embodiment of the invention, the step of obtaining the design contour data of the tunnel section to be supported and performing path planning on the tunnel section to be supported to generate the basic spraying path includes: Obtain the design contour data of the tunnel section to be supported, and extract the contour line of the tunnel section to be supported; According to the preset spray width and overlap width, the nozzle movement trajectory of the tunnel section to be supported is constructed, and the nozzle movement trajectory is used as the basic spray path.

[0020] During the construction preparation phase, the control system first acquires the design contour data of the tunnel cross-section to be supported. This design contour data comes from the tunnel construction design drawings and includes the geometric shape and dimensional information of the tunnel cross-section. The control system analyzes the acquired design contour data to extract the contour line of the tunnel cross-section to be supported, which serves as the basis for subsequent path planning. Subsequently, based on the pre-set spray width and the overlap width between adjacent spray zones, the control system constructs the nozzle movement trajectory within the area defined by the contour line.

[0021] The spray width refers to the width of the nozzle covering the sprayed surface during a single spraying operation. The overlap width is the overlapping portion reserved to ensure that there are no gaps between adjacent spray bands.

[0022] The nozzle movement trajectory is usually in the form of a spiral or an S-shaped reciprocating line to ensure that the nozzle can evenly cover the entire tunnel cross section.

[0023] The control system uses the completed nozzle movement trajectory as the basic spray path. This basic spray path includes not only the spatial position sequence of the nozzle, but also the attitude angle of the nozzle relative to the sprayed surface at each position and the preset spray thickness, providing a basis for subsequent automated spraying operations.

[0024] Furthermore, as the wet spraying robot begins its initial concrete spraying operation along the foundation spraying path, the status monitoring unit continues to operate. Distance sensors measure the distance between the nozzle and the sprayed concrete surface in real time, generating real-time distance data; the image acquisition unit simultaneously captures images of the sprayed concrete surface, generating real-time image data. Both types of data are transmitted to the control system in real time. The real-time distance data reflects the actual distance between the nozzle and the sprayed surface, while the real-time image data records the visual characteristics of the concrete surface.

[0025] Furthermore, the control system processes the received real-time image data. First, the real-time image data undergoes preprocessing, including applying a mean square filtering algorithm to remove noise introduced during image acquisition, and converting the color image to a grayscale image to obtain a grayscale image of the sprayed concrete layer surface. Then, the control system performs texture analysis on the grayscale image, specifically by constructing a gray-level co-occurrence matrix to quantify the image's texture features.

[0026] The gray-level co-occurrence matrix (GLCM) is a matrix that statistically represents the frequency of occurrence of specific gray-level combinations in different directions in an image. The control system extracts two parameters from it: contrast and energy. Contrast reflects the sharpness of the image and the depth of texture grooves, while energy reflects the uniformity and roughness of the texture. These two parameters together constitute the texture feature parameters.

[0027] Simultaneously, the control system performs highlight region detection on the grayscale image, counts the number of pixels with grayscale values ​​higher than a preset grayscale threshold, calculates the proportion of these highlight pixels to the total number of pixels in the image, and uses this proportion as a gloss characteristic parameter. This parameter is directly related to the moisture content and cement paste enrichment of the concrete surface. On the other hand, the control system calculates the real-time thickness deviation of the sprayed concrete layer surface based on real-time distance data and the preset spraying thickness at the corresponding position in the base spraying path.

[0028] Specifically, the control system compares the real-time distance data measured by the nozzle at each sampling point with the preset spray thickness at the same location in the base spraying path, calculates the difference between the two, and obtains the point thickness deviation value for that sampling point. The control system integrates the point thickness deviation values ​​of all sampling points according to their spatial location to generate a thickness deviation distribution map covering a local area of ​​the sprayed concrete layer.

[0029] Furthermore, the control system inputs the calculated texture feature parameters, gloss feature parameters, and thickness deviation distribution map into a pre-constructed interlayer bonding state determination model. This model is established based on historical construction data through multivariate nonlinear regression analysis, and it internally stores model coefficients as well as contrast benchmark values, energy benchmark values, gloss benchmark values, and thickness deviation tolerance. The model calculates an interlayer bonding timing coefficient based on the input parameters. This coefficient is a continuous variable between a first threshold and a second threshold, used to quantify whether the currently sprayed concrete layer is suitable for receiving the next layer of spraying.

[0030] The calculation of the interlayer bonding timing coefficient comprehensively considers the degree of deviation of the current concrete surface texture characteristics from the ideal state, the degree of deviation of the gloss characteristics from the ideal state, and the magnitude of the average thickness deviation. The influence of the thickness deviation is reflected in the form of exponential decay. When the average thickness deviation approaches or exceeds the tolerance, the coefficient will decrease significantly.

[0031] Furthermore, the control system monitors the value of the interlayer bonding timing coefficient in real time. When this coefficient falls within a preset target range, the control system determines that the current moment meets the conditions for the start of the next layer's injection and then initiates preparations for the next layer's injection. The control system acquires the current thickness deviation distribution map and performs coupled calculations with the preset injection thickness in the basic injection path.

[0032] The coupled calculation process is as follows: For each target location point on the basic injection path, the control system delineates a circular neighborhood centered on that point, extracts the point thickness deviation values ​​of all points in the thickness deviation distribution map within that neighborhood, and performs spatial weighted integration on these deviation values ​​to obtain the comprehensive influence value of the thickness deviation at that point.

[0033] The weighting function's value depends on the distance from each point in the neighborhood to the center point; the closer the point, the greater the weight, and the farther the point, the smaller the weight. This weighting method reflects the continuity of the thickness deviation's influence. Dividing the comprehensive influence value of the thickness deviation by the preset injection thickness at the center point, multiplying by the preset compensation intensity coefficient, and finally adding one yields the dynamic compensation coefficient for that point. The dynamic compensation coefficient reflects the adjustment range required based on the preset injection parameters to compensate for local thickness deviations.

[0034] Furthermore, the control system uses the calculated dynamic compensation coefficient to correct the unexecuted portions of the basic injection path in real time, generating a secondary injection path. The correction process includes: dividing the preset moving speed of each point on the basic injection path by the corresponding dynamic compensation coefficient to obtain the corrected moving speed of the corresponding point on the secondary injection path; multiplying the preset injection flow rate of each point on the basic injection path by the corresponding dynamic compensation coefficient to obtain the corrected injection flow rate of the corresponding point on the secondary injection path; while the nozzle's attitude angle sequence remains unchanged, maintaining a direction perpendicular to the sprayed surface. After this correction, in areas with large thickness deviations, the nozzle moving speed will slow down and the injection flow rate will increase, thus achieving compensatory injection in that area; in areas with small thickness deviations, the injection volume will be reduced accordingly to ensure overall thickness uniformity.

[0035] Furthermore, the control system controls the boom system to drive the nozzles, performing the spraying of the next layer of concrete on the currently sprayed concrete layer according to the moving speed, attitude angle, and spraying flow rate defined by the secondary spraying path. After spraying is completed, the newly formed concrete layer becomes the new sprayed concrete layer. Subsequently, the control system automatically returns to the step of real-time acquisition of distance and image data, and, using the new layer as the target, performs status monitoring, timing determination, path correction, and spraying operations again, repeating this cycle. When the cumulative sprayed thickness reaches the total sprayed thickness required by the design profile data, the control system stops repeating, completing the multi-layer shotcrete support for the entire tunnel cross-section.

[0036] In summary, by acquiring design contour data and planning the basic spraying path, an initial basis was provided for subsequent spraying operations; by collecting distance and image data in real time, continuous perception of the construction process was achieved; by extracting texture and gloss features from image data and calculating thickness deviations in conjunction with distance data, the concrete surface state was transformed into quantifiable parameters; by inputting the feature parameters into the interlayer bonding state determination model, the interlayer bonding timing coefficient was obtained, enabling scientific determination of the timing of lower layer spraying; by calculating dynamic compensation coefficients and correcting the path when conditions are met, a secondary spraying path adapted to actual thickness deviations was generated; by executing secondary spraying and repeating the cycle until multi-layer support was completed, the construction quality of the initial support of the entire tunnel section was ensured.

[0037] Z2: Real-time acquisition of real-time distance data between the nozzle and the surface of the sprayed concrete layer, as well as real-time image data of the surface of the sprayed concrete layer; In this embodiment of the invention, during the initial concrete spraying operation along the foundation spraying path by the wet spraying robot, the status monitoring unit operates continuously. A distance sensor measures the distance between the nozzle and the surface of the sprayed concrete layer in real time. The distance sensor can be a laser rangefinder, installed near the nozzle and moving with it. The laser rangefinder emits a laser beam towards the surface of the sprayed concrete layer. After the laser beam contacts the concrete surface, it reflects back to the sensor. The sensor calculates the distance from the nozzle to the concrete surface based on the time difference between laser emission and reception, generating real-time distance data. This data is transmitted to the control system in real-time as a digital signal for subsequent calculation of thickness deviation.

[0038] The image acquisition unit operates synchronously, capturing real-time images of the sprayed concrete surface. This unit utilizes an industrial camera, also mounted near the nozzle with its lens facing the concrete surface, and is equipped with supplementary lighting to ensure clear images even in low-light conditions within the tunnel. The industrial camera continuously acquires images at a preset frame rate, generating real-time image data. This real-time image data includes information such as the texture, gloss, and color of the concrete surface, and is transmitted to the control system in digital image format. The control system synchronously acquires real-time distance data and real-time image data at a fixed frequency, ensuring temporal correspondence between the two sets of data and providing accurate input for subsequent feature extraction and thickness calculation.

[0039] Z3: Extract texture features and gloss features from the real-time image data to obtain texture feature parameters and gloss feature parameters, and calculate the real-time thickness deviation value of the sprayed concrete layer surface based on the real-time distance data and the preset spray thickness at the corresponding position in the basic spraying path. In this embodiment of the invention, the step of extracting texture features and gloss features from the real-time image data to obtain texture feature parameters and gloss feature parameters, and calculating the real-time thickness deviation value of the sprayed concrete layer surface based on the real-time distance data and the preset spray thickness at the corresponding position in the base spraying path, includes: The real-time image data is preprocessed, including median filtering for noise reduction and grayscale processing, to obtain a grayscale image of the sprayed concrete layer surface. Texture analysis is performed on the grayscale image to obtain the grayscale co-occurrence matrix of the grayscale image, and the contrast and energy of the grayscale co-occurrence matrix are integrated into the texture feature parameters of the grayscale image; The grayscale image is subjected to highlight region detection. Pixels in the grayscale image whose grayscale value is higher than a preset grayscale threshold are identified as highlight pixels. The total percentage of the highlight pixels in the grayscale image is used as a glossiness feature parameter. Based on the real-time distance data and the preset spraying thickness at the corresponding position in the basic spraying path, the real-time thickness deviation value of the sprayed concrete layer surface is calculated.

[0040] The step of calculating the real-time thickness deviation value of the sprayed concrete layer surface based on the real-time distance data and the preset spraying thickness at the corresponding position in the basic spraying path includes: Based on the normal direction on the nozzle movement trajectory, determine the nozzle attitude angle sequence in which the nozzle axis is consistent with the normal direction; Based on the total spray thickness of the design contour data and the preset number of layered sprays, the preset spray thickness on the nozzle movement trajectory is determined; For the sampling points on the nozzle movement trajectory, the difference between the real-time distance data and the preset injection thickness is used as the point thickness deviation value of the sampling point; The thickness deviation values ​​at the sampling points are integrated according to their spatial locations to generate a thickness deviation distribution map of a local area on the surface of the sprayed concrete layer.

[0041] The control system first preprocesses the real-time image data. Preprocessing includes two steps: median filtering for denoising and grayscale conversion. Median filtering replaces the grayscale value of each pixel in the real-time image data with the median of the grayscale values ​​of all pixels in its neighborhood. This effectively removes isolated noise introduced during image acquisition due to factors such as dust and uneven lighting, while maintaining the clarity of image edges. Grayscale conversion transforms the denoised color image into a grayscale image. Specifically, it calculates a grayscale value for each pixel's red, green, and blue color components according to preset weighting coefficients, resulting in a grayscale image containing only brightness information, thus reducing the computational complexity of subsequent processing. After preprocessing, the control system obtains a grayscale image of the sprayed concrete surface.

[0042] Furthermore, the control system performs texture analysis on the grayscale image. The texture analysis employs the gray-level co-occurrence matrix (GLCM) method. The GLCM is a matrix that statistically analyzes the frequency of gray-level combinations between two pixels in a grayscale image that are within a specific direction and distance. The control system first quantizes the pixel gray levels of the grayscale image, then selects a direction and a step size, and counts the number of gray-level combinations between two pixels within that step size along that direction, thus constructing the GLCM.

[0043] Then, the control system calculates two statistics, contrast and energy, from the gray-level co-occurrence matrix. Contrast is a measure of how much the element values ​​in the gray-level co-occurrence matrix deviate from the diagonal, reflecting the sharpness of the image and the depth of the texture grooves; a higher contrast indicates deeper texture grooves and a sharper image. Energy is the sum of the squares of all elements in the gray-level co-occurrence matrix, reflecting the uniformity and roughness of the texture; a higher energy indicates a more uniform and rougher texture. The control system integrates the calculated contrast and energy into a texture feature parameter of the gray-level image, which characterizes the density and flow state of the concrete layer surface.

[0044] Furthermore, the control system performs highlight region detection on the grayscale image. Highlight region detection is a threshold segmentation method based on pixel grayscale values. The control system acquires the grayscale values ​​of all pixels in the grayscale image and compares the grayscale value of each pixel with a preset grayscale threshold.

[0045] The preset grayscale threshold is a value pre-set based on the ambient lighting conditions and the characteristics of the concrete surface, used to distinguish between normal areas and highlight areas. When the grayscale value of a pixel exceeds the preset grayscale threshold, the control system classifies that pixel as a highlight pixel, indicating that the area has strong reflectivity due to surface wetness or cement paste accumulation. The control system counts the number of all highlight pixels in the entire grayscale image and calculates the proportion of highlight pixels to the total number of pixels in the image; this proportion is used as a gloss characteristic parameter. The gloss characteristic parameter reflects the degree of wetness of the concrete surface and the distribution of cement paste.

[0046] Furthermore, the control system calculates the real-time thickness deviation of the sprayed concrete layer surface based on the real-time distance data and the preset spray thickness at the corresponding position in the basic spraying path. Before the first spraying begins, the control system has completed path planning and generated the basic spraying path, which includes not only the movement trajectory of the nozzle but also the preset spray thickness at each position along the trajectory.

[0047] The preset spray thickness is determined based on the total spray thickness required by the design contour data and the preset number of spray layers. It represents the thickness value that each layer should achieve at that location. During spraying, a distance sensor collects the distance between the nozzle and the surface of the sprayed concrete layer in real time, generating real-time distance data. The control system compares the real-time distance data with the preset spray thickness at the same location in the foundation spraying path, and calculates the difference between the two. This difference is the real-time thickness deviation value at that location; a positive value indicates that the actual thickness is greater than the preset thickness, and a negative value indicates that the actual thickness is less than the preset thickness.

[0048] Furthermore, when calculating the real-time thickness deviation, the control system first determines the nozzle attitude angle sequence based on the normal direction on the nozzle movement trajectory. The nozzle movement trajectory is a curve composed of a series of spatial points. At each point, the control system calculates the normal direction of the trajectory at that point, which is the direction perpendicular to the tangent of the trajectory and pointing inward to the sprayed surface.

[0049] The control system adjusts the nozzle axis to align with the normal direction, forming a nozzle posture angle sequence to ensure that the nozzle is always perpendicular to the sprayed surface, thus guaranteeing the adhesion and thickness uniformity of the concrete. Then, based on the total spray thickness required in the design profile data and the preset number of layer spraying times, the control system determines the preset spray thickness at each position on the nozzle movement trajectory.

[0050] The total shotcrete thickness refers to the total thickness of the concrete layers that the initial support of the tunnel needs to achieve. The number of spraying layers refers to the number of layers to be sprayed in order to achieve this total thickness. Dividing the total shotcrete thickness by the number of spraying layers gives the average shotcrete thickness of each layer, which is the preset shotcrete thickness.

[0051] Next, the control system compares the real-time distance data with the preset spray thickness at each sampling point on the nozzle's movement trajectory to calculate the point thickness deviation value. Sampling points are locations selected at certain spatial intervals on the nozzle's movement trajectory; each sampling point has corresponding real-time distance data and a preset spray thickness.

[0052] Finally, the control system integrates the thickness deviation values ​​of all sampling points according to their spatial location to generate a thickness deviation distribution map of a local area on the surface of the sprayed concrete layer.

[0053] In summary, the above steps, through median filtering and grayscale processing of real-time image data, eliminated the interference of dust and uneven lighting within the tunnel, resulting in clear and accurate grayscale images and laying a reliable foundation for subsequent feature extraction. By performing texture analysis on the grayscale images and constructing a grayscale co-occurrence matrix, contrast and energy were used as texture feature parameters to quantitatively characterize the surface density and flow state of the concrete layer. Highlight region detection was performed on the grayscale images, and the proportion of highlight pixels was used as a gloss feature parameter, enabling quantitative characterization of the surface wettability and cement paste distribution of the concrete layer. By calculating the real-time thickness deviation value based on the real-time distance data and the preset spray thickness at the corresponding position in the foundation spraying path, the visual and geometric features of the concrete layer surface were transformed into a quantifiable and processable parameter system. This provided accurate and reliable input for the scientific determination of the timing of interlayer bonding and the adaptive correction of the spraying path, thereby ensuring the construction quality of multi-layer shotcrete support.

[0054] Z4: Input the texture feature parameters, gloss feature parameters and the real-time thickness deviation value into the pre-constructed interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface; In this embodiment of the invention, the step of inputting the texture feature parameters, gloss feature parameters, and real-time thickness deviation value into a pre-constructed interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface includes: Obtain historical construction data of the tunnel section to be supported. The historical construction data includes historical texture feature parameters, historical gloss feature parameters, historical thickness deviation distribution map, and interlayer bonding strength value. A multivariate nonlinear regression analysis was performed on the historical texture feature parameters, historical gloss feature parameters, historical thickness deviation distribution map, and interlayer bonding strength value to obtain model coefficients; Based on the model coefficients, a model for determining the interlayer bonding state of the tunnel section to be supported is constructed. The texture feature parameters, the gloss feature parameters, and the real-time thickness deviation value are input into the interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface.

[0055] The formula for calculating the interlayer bonding timing coefficient is as follows: in, This refers to the interlayer bonding timing coefficient. The contrast of the gray-level co-occurrence matrix. The preset contrast reference value, , , , These are the weighting coefficients of the interlayer bonding state determination model. The energy of the gray-level co-occurrence matrix. The preset energy baseline value, The gloss characteristic parameter at the current moment, The preset gloss reference value, The average of the point thickness deviation values ​​for all sampling points. This is the preset thickness deviation tolerance. It is an exponential function.

[0056] First, the control system acquires historical construction data of the tunnel section to be supported. This historical construction data originates from construction records of similar geological sections already completed within the same tunnel project, or from data collected during the construction of previous test sections. Each set of historical construction data includes four components: historical texture feature parameters, i.e., the contrast and energy values ​​of the gray-level co-occurrence matrix recorded in past construction; historical gloss feature parameters, i.e., the proportion of highlight pixels recorded in past construction; historical thickness deviation distribution map, i.e., the spatial distribution data of the thickness deviation values ​​at each sampling point generated in past construction; and the interlayer bond strength value corresponding to the above three sets of data. This interlayer bond strength value is the concrete interlayer bond force value obtained through on-site core drilling and pull-out tests after the corresponding construction is completed. The control system archives and stores this historical construction data according to construction time or spatial location, forming a dataset for model construction.

[0057] Then, the control system performs a multivariate nonlinear regression analysis on the historical texture feature parameters, historical gloss feature parameters, historical thickness deviation distribution map, and interlayer bonding strength value. During the regression analysis, the control system first extracts the point thickness deviation values ​​of all sampling points from each historical thickness deviation distribution map and calculates the arithmetic mean of these point thickness deviation values ​​as the average thickness deviation of that set of historical data. Subsequently, the control system uses the historical texture feature parameters, historical gloss feature parameters, and the calculated average thickness deviation from each set of historical data as independent variables, and the corresponding interlayer bonding strength value as the dependent variable to construct a regression equation. The goal of the regression analysis is to find a set of coefficients that minimizes the error between the calculated value and the measured interlayer bonding strength value after substituting the independent variables into the regression equation. To achieve this goal, the control system uses an iterative optimization method, continuously adjusting the values ​​of the coefficients until the sum of squared errors between the calculated and measured values ​​of all historical data sets reaches its minimum. The coefficients obtained at this point are the model coefficients.

[0058] Next, based on the model coefficients, the control system constructs a model for determining the interlayer bonding state of the tunnel section to be supported. The interlayer bonding state determination model is a set of calculation rules stored in the control system's memory. This set of rules contains two parts: the first part is the model coefficients obtained from regression analysis, which are stored in memory as fixed values; the second part is the model's calculation structure, i.e., how to combine the input texture feature parameters, gloss feature parameters, and real-time thickness deviation values ​​to calculate the interlayer bonding timing coefficients in the correct order. Simultaneously with model construction, the control system also stores preset contrast reference values, energy reference values, gloss reference values, and thickness deviation tolerances in memory as fixed parameters to be called during model calculation.

[0059] Finally, the control system inputs the texture feature parameters, gloss feature parameters, and real-time thickness deviation values ​​calculated in real time into the interlayer bonding state determination model. After receiving these input parameters, the model first extracts the point thickness deviation values ​​of all sampling points from the real-time thickness deviation values ​​and calculates their average value to obtain the current average thickness deviation.

[0060] Then, the model reads preset contrast reference values, energy reference values, gloss reference values, thickness deviation tolerance, and model coefficients from memory. Next, following a predetermined calculation order, the model substitutes the input texture feature parameters, gloss feature parameters, the calculated average thickness deviation, and the read reference values ​​and coefficients into the calculation rules, sequentially performing ratio calculations, exponential calculations, and weighted summations, ultimately calculating a value that is the interlayer bonding timing coefficient of the sprayed concrete layer surface. The control system uses this coefficient as a quantitative representation of the current interlayer bonding state for subsequent start condition determination.

[0061] in, The interlayer bonding timing coefficient is obtained by inputting the texture feature parameters, gloss feature parameters, and real-time thickness deviation values ​​into a pre-constructed interlayer bonding state determination model. After ratio calculation, exponential operation, and weighted summation within the model, the output value is used to quantitatively characterize whether the currently sprayed concrete layer is suitable for the next layer of spraying.

[0062] The contrast of the gray-level co-occurrence matrix is ​​a statistical measure extracted from the gray-level co-occurrence matrix of the gray-level image during texture analysis. It reflects the depth of the texture grooves on the surface of the concrete layer and the image clarity.

[0063] The preset contrast reference value is a value that is determined in advance through experiments and stored in the control system's memory. This reference value corresponds to the mix proportion information of the concrete used in the current construction and is used as a standardized reference for contrast.

[0064] The first weighting coefficient of the interlayer bonding state determination model is obtained by minimizing the error between the calculated value and the measured interlayer bonding strength value when performing multivariate nonlinear regression analysis on the historical construction data. It reflects the degree of contribution of the relative contrast value to the interlayer bonding timing coefficient.

[0065] The second weighting coefficient of the interlayer bonding state determination model is a coefficient obtained by minimizing the error between the calculated value and the measured interlayer bonding strength value when performing multivariate nonlinear regression analysis on the historical construction data. It reflects the degree of contribution of the relative energy value to the interlayer bonding timing coefficient.

[0066] The third weighting coefficient of the interlayer bonding state determination model is a coefficient obtained by minimizing the error between the calculated value and the measured interlayer bonding strength value when performing multivariate nonlinear regression analysis on the historical construction data. It reflects the degree of contribution of the relative gloss value to the interlayer bonding timing coefficient.

[0067] The weighting coefficients of the interlayer bonding state determination model are obtained by minimizing the error between the calculated value and the measured interlayer bonding strength value when performing multivariate nonlinear regression analysis on the historical construction data. They reflect the degree of contribution of the thickness deviation index term to the interlayer bonding timing coefficient.

[0068] The energy of the gray-level co-occurrence matrix is ​​used to perform texture analysis on the gray-level image. First, the gray-level co-occurrence matrix of the gray-level image is constructed. Then, the energy statistics extracted from the gray-level co-occurrence matrix reflect the uniformity and roughness of the surface texture of the concrete layer.

[0069] The preset energy benchmark value is a value that is determined in advance through experiments and stored in the control system's memory. This benchmark value corresponds to the mix proportion information of the concrete used in the current construction and is used as a standardized benchmark for energy.

[0070] The gloss feature parameter at the current moment is obtained by counting the number of highlight pixels in the grayscale image whose grayscale value is higher than a preset grayscale threshold and calculating the proportion of the number of highlight pixels to the total number of pixels in the grayscale image when performing highlight region detection on the grayscale image. It reflects the wetness of the concrete layer surface and the distribution of cement paste.

[0071] The preset gloss reference value is a value that is determined in advance through experiments and stored in the control system's memory. This reference value corresponds to the mix proportion information of the concrete used in the current construction and is used as a standardized reference for gloss characteristic parameters.

[0072] The average value of the thickness deviation of all sampling points is calculated based on the real-time distance data and the preset spraying thickness at the corresponding position in the basic spraying path. The thickness deviation of each sampling point is then added together and divided by the number of sampling points. This value reflects the average level of the overall thickness deviation of the sprayed concrete layer surface.

[0073] The preset thickness deviation tolerance is a constant that is predetermined based on engineering experience and stored in the control system memory. It is used as a standardization benchmark for the average thickness deviation in the exponential calculation. When the average thickness deviation approaches or exceeds this tolerance, the exponential term will significantly decrease.

[0074] The thickness deviation index term is obtained by dividing the average thickness deviation by the thickness deviation tolerance to obtain an intermediate ratio, and then performing an exponential operation on the negative value of the intermediate ratio. When the average thickness deviation is much smaller than the thickness deviation tolerance, the index term approaches one. When the average thickness deviation approaches or exceeds the thickness deviation tolerance, the index term rapidly decays to approach zero, reflecting the influence of thickness deviation on the timing of interlayer bonding in a nonlinear manner.

[0075] First, the contrast of the gray-level co-occurrence matrix is ​​divided by its corresponding preset contrast benchmark value to obtain a relative contrast value. Then, this relative contrast value is multiplied by the first weighting coefficient in the interlayer bonding state determination model to obtain the first product term. The control system divides the energy of the gray-level co-occurrence matrix by its corresponding preset energy benchmark value to obtain a relative energy value. Then, this relative energy value is multiplied by the second weighting coefficient in the interlayer bonding state determination model to obtain the second product term. The control system divides the gloss characteristic parameter at the current moment by its corresponding preset gloss benchmark value to obtain a relative gloss value. Then, this relative gloss value is multiplied by the third weighting coefficient in the interlayer bonding state determination model to obtain the third product term.

[0076] The control system divides the calculated average thickness deviation by a preset thickness deviation tolerance to obtain an intermediate ratio. Then, the control system performs an exponential operation on the negative value of this intermediate ratio, that is, it calculates an exponential function with the natural constant as the base and the negative value of the intermediate ratio as the exponent, to obtain the thickness deviation exponential term. This thickness deviation exponential term is then multiplied by the fourth weighting coefficient in the interlayer bonding state determination model to obtain the fourth product term.

[0077] Finally, the control system adds the first, second, third, and fourth product terms together, and the sum is the interlayer combination timing coefficient for the current moment. This coefficient is stored in a temporary register of the control system and used to determine the subsequent lower-layer injection initiation conditions.

[0078] In summary, by performing median filtering and grayscale processing on real-time image data, the interference of dust and uneven lighting within the tunnel on the image was eliminated, resulting in a clear and accurate grayscale image, laying a reliable foundation for subsequent feature extraction. By performing texture analysis on the grayscale image and constructing a grayscale co-occurrence matrix, contrast and energy were used as texture feature parameters to quantitatively characterize the surface density and flow state of the concrete layer. By detecting highlight areas in the grayscale image and using the proportion of highlight pixels as a gloss feature parameter, the surface wettability and cement paste distribution of the concrete layer were quantitatively characterized. Furthermore, by using real-time distance data and pre-set spraying parameters at corresponding positions in the foundation spraying path... The process involves calculating real-time thickness deviation values, determining the nozzle attitude angle sequence based on the normal direction of the nozzle's movement trajectory, determining the preset spray thickness based on the total spray thickness and the number of layer spraying times in the design contour data, calculating point thickness deviation values ​​for sampling points, and integrating the point thickness deviation values ​​into a thickness deviation distribution map. This achieves a quantitative representation of the spatial distribution of the uniformity of the sprayed concrete layer surface thickness. These steps collectively transform the visual and geometric features of the concrete layer surface into a quantifiable and processable parameter system, providing accurate and reliable input for the scientific determination of the timing of interlayer bonding and the adaptive correction of the spraying path, thereby ensuring the construction quality of multi-layer shotcrete support.

[0079] Z5: When the interlayer bonding timing coefficient meets the lower layer spraying start condition of the tunnel section to be supported, the real-time thickness deviation value is coupled with the preset spraying thickness to obtain the dynamic compensation coefficient of the nozzle, and the unexecuted part in the basic spraying path is corrected in real time based on the dynamic compensation coefficient to obtain the secondary spraying path of the tunnel section to be supported. In this embodiment of the invention, when the interlayer bonding timing coefficient satisfies the lower layer injection initiation condition of the tunnel section to be supported, the real-time thickness deviation value is coupled with the preset injection thickness for calculation to obtain the dynamic compensation coefficient of the nozzle, and the unexecuted portion of the basic injection path is corrected in real time based on the dynamic compensation coefficient to obtain the secondary injection path of the tunnel section to be supported, including: When the interlayer bonding timing coefficient is greater than or equal to the preset target lower threshold, the lower layer spraying start condition of the tunnel section to be supported is triggered. The dynamic compensation coefficient of the nozzle is obtained by coupling the real-time thickness deviation value with the preset injection thickness. Based on the dynamic compensation coefficient, the unexecuted portion of the basic spraying path is corrected in real time to obtain the secondary spraying path of the tunnel section to be supported.

[0080] The formula for calculating the dynamic compensation coefficient is as follows: in, For any target location point on the basic injection path, For any point in the thickness deviation distribution map, For point The dynamic compensation coefficient at that location. The preset compensation strength coefficient, For the weight function, For any point in the thickness deviation distribution map The point thickness deviation value, For the point Centered on, with a preset radius The circular neighborhood, For point The preset spray thickness at the location.

[0081] The process of real-time correction of the unexecuted portion of the basic spraying path based on the dynamic compensation coefficient to obtain the secondary spraying path of the tunnel section to be supported includes: The preset moving speed of each point on the basic injection path is mathematically multiplied by the reciprocal of the dynamic compensation coefficient, and the result of the mathematical product is used as the corrected moving speed of the corresponding point on the secondary injection path. The product of the preset injection flow rate at each point on the basic injection path and the dynamic compensation coefficient is used as the corrected injection flow rate at the corresponding point on the secondary injection path. Keeping the nozzle attitude angle sequence of each point on the basic injection path unchanged, and making real-time corrections to the unexecuted parts of the basic injection path based on the corrected moving speed and the corrected injection flow rate, a secondary injection path for the tunnel section to be supported is obtained.

[0082] The control system first monitors the calculated interlayer coupling timing coefficient in real time and compares it with a target lower limit threshold pre-stored in the control system's memory. This target lower limit threshold is a fixed value determined based on engineering experience, used to define the minimum interlayer coupling timing coefficient value allowed for lower-layer spraying operations. When the interlayer coupling timing coefficient is greater than or equal to the preset target lower limit threshold, the control system determines that the current moment meets the lower-layer spraying initiation conditions and then triggers subsequent coupling calculations and path correction processes.

[0083] The control system couples the thickness deviation values ​​of each sampling point included in the thickness deviation distribution map with the preset spray thickness at the corresponding position on the basic spray path. This coupling calculation comprehensively considers the spatial distribution of local thickness deviation and the reference value of the preset spray thickness. Through specific calculation rules, it calculates the required adjustment range at each position and finally obtains the dynamic compensation coefficient of the nozzle in the entire area to be corrected. This dynamic compensation coefficient is a numerical field that varies with spatial position and reflects the proportional factor that needs to be adjusted in subsequent sprays to compensate for the thickness deviation of the sprayed layer.

[0084] The control system then performs real-time corrections on the unexecuted portions of the basic injection path based on the calculated dynamic compensation coefficient. During the correction process, the control system mathematically multiplies the preset moving speed of each point on the basic injection path with the reciprocal of the dynamic compensation coefficient at the corresponding position, using the result as the corrected moving speed of the corresponding point on the secondary injection path. This means that in areas with large thickness deviations, the dynamic compensation coefficient is larger, and its reciprocal is smaller, resulting in a slower corrected moving speed and thus increasing the injection time per unit area. The control system also mathematically multiplies the preset injection flow rate of each point on the basic injection path with the dynamic compensation coefficient at the corresponding position, using the result as the corrected injection flow rate of the corresponding point on the secondary injection path. This means that in areas with large thickness deviations, the dynamic compensation coefficient is larger, resulting in an increased corrected injection flow rate and thus increasing the amount of injection material supplied per unit area. Simultaneously, the control system maintains the nozzle attitude angle sequence at each point on the basic injection path unchanged, ensuring that the nozzles can still spray perpendicularly to the sprayed surface on the corrected path.

[0085] The control system integrates the complete path data, including the corrected moving speed, corrected jet flow rate, and original nozzle attitude angle sequence, obtained after the above corrections, into a secondary jetting path for the tunnel section to be supported. This secondary jetting path is stored in the working memory of the control system and is used to drive the subsequent next layer of jetting operations.

[0086] For any target position point on the basic injection path, it is a spatial coordinate point on the nozzle movement trajectory included in the basic injection path, representing a specific position that the nozzle needs to pass through in subsequent injection operations. Each point has a corresponding preset movement speed, preset injection flow rate and preset injection thickness.

[0087] Let A be any point in the thickness deviation distribution map. This point is a spatial coordinate point within the area covered by the thickness deviation distribution map. Each point has a corresponding point thickness deviation value. These points are located at points... Circular neighborhood centered Within, its thickness deviation value is used to calculate the point through a weighting function. The dynamic compensation coefficient.

[0088] For point The dynamic compensation coefficient mentioned above is obtained by coupling the real-time thickness deviation value with the preset injection thickness, specifically based on the points on the basic injection path. Define a circular neighborhood around the center, perform a spatially weighted integral on the thickness deviation values ​​of each point in the thickness deviation distribution map within this neighborhood, and divide the integral result by the number of points. The preset spray thickness at the location is multiplied by the preset compensation intensity coefficient and then added to obtain the result, which is used for subsequent point alignment. The preset moving speed and preset jet flow rate are corrected.

[0089] The preset compensation strength coefficient is a constant determined based on engineering experience and stored in the control system's memory before construction. It is used to adjust the degree of influence of thickness deviation on the dynamic compensation coefficient. The larger the value of this coefficient, the greater the compensation amplitude produced by the same thickness deviation.

[0090] This is the weighting function, which is a Gaussian function for calculating weights, used to determine points at different locations within a circular neighborhood. For the center point The magnitude of the compensation contribution is calculated using a point-based method. Measurement point at the origin Time The Euclidean distance is calculated, and this distance is substituted into the Gaussian function expression, which involves a preset influence range coefficient. The point is then calculated using this expression. The corresponding weight value.

[0091] For any point in the thickness deviation distribution map The point thickness deviation value is calculated by subtracting the real-time distance data from the preset injection thickness at the corresponding location for each sampling point on the nozzle's movement trajectory. This value is then integrated into the thickness deviation distribution map according to the spatial location of the sampling points, reflecting the point thickness deviation. The difference between the actual spray thickness at the location and the preset spray thickness.

[0092] For the point Centered on, with a preset radius A circular neighborhood, which is based on a preset radius. Determined, preset radius These are constants stored in the control system's memory, used to define the participation points. The range of thickness deviation points calculated by the dynamic compensation coefficient is limited to points located within this circular neighborhood. Only the point thickness deviation value will be included in the weighted integral calculation.

[0093] For point The preset spray thickness at a given point is determined based on the total spray thickness required by the design contour data and the preset number of layered sprays, and is stored in the control system as part of the basic spray path, representing the spray thickness at that point. The planned thickness of the sprayed concrete layer is used as a denominator in the calculation of the dynamic compensation coefficient to normalize the comprehensive impact of thickness deviation.

[0094] In summary, by comparing the interlayer bonding timing coefficient with the preset target lower threshold, the system automatically triggers the initial conditions for spraying the lower layer, ensuring that subsequent work is only carried out when the sprayed concrete layer reaches a suitable bonding state, thus avoiding insufficient interlayer bonding strength due to improper timing. By coupling the point thickness deviation values ​​in the thickness deviation distribution map with the preset spray thickness on the base spraying path and introducing a weighted integral of the thickness deviation at each point within a circular neighborhood centered on the spatial location, a dynamic compensation coefficient that varies with spatial location is generated. This coefficient accurately reflects the compensation magnitude required for thickness deviation in different local areas. In summary, by multiplying the preset moving speed of each point on the basic spraying path by the reciprocal of the dynamic compensation coefficient to obtain the corrected moving speed, and multiplying the preset spraying flow rate by the dynamic compensation coefficient to obtain the corrected spraying flow rate, while keeping the nozzle attitude angle sequence unchanged, a decoupled coordinated adjustment of moving speed and spraying flow rate is achieved. This allows the nozzle to automatically reduce its moving speed and increase its spraying flow rate to increase material supply in areas with thinner thickness, and automatically increase its moving speed and decrease its spraying flow rate to reduce material supply in areas with thicker thickness. The resulting secondary spraying path enables subsequent spraying operations to accurately adapt to the actual thickness distribution of the sprayed layer, achieving refined compensation for local thickness deviations. This significantly improves the overall thickness uniformity and interlayer bonding quality of multi-layer sprayed support, ensuring the load-bearing capacity and durability of the tunnel's initial support structure.

[0095] Z6: Based on the secondary spraying path, perform the next layer of spraying operation on the surface of the sprayed concrete layer, repeating steps Z2 to Z6 until the multi-layer spraying support of the tunnel section to be supported is completed.

[0096] In this embodiment of the invention, the step of performing the next layer of spraying operation on the surface of the already sprayed concrete layer based on the secondary spraying path, repeating steps Z2 to Z6 until the multi-layer sprayed support of the tunnel section to be supported is completed, includes: After the current spraying operation of the tunnel section to be supported is completed, the end time of the spraying of the current layer in the tunnel section to be supported is taken as the start time of the spraying operation of the next layer, and the process returns to step Z2. The newly completed shotcrete layer in the tunnel section to be supported is used as the new shotcrete layer, and the lower shotcrete operation of the tunnel section to be supported is carried out. When the cumulative spray thickness reaches the total spray thickness required in the design profile data, the repetition stops, and the support of the tunnel section to be supported is completed.

[0097] After the current spraying operation on the tunnel section to be supported is completed, that is, when the next layer of spraying operation driven by the secondary spraying path is completed, the control system records the end time of the spraying operation of that layer, takes that time as the start time of the next layer of spraying operation, and automatically returns to step Z2 to prepare for the start of a new cycle of data acquisition and processing.

[0098] The control system takes the newly completed sprayed concrete layer in the tunnel section to be supported as the new sprayed concrete layer. That is, it takes the concrete layer that has just been sprayed and has initially set as the object of the next round of monitoring and operation, and continues to carry out the lower layer spraying operation of the tunnel section to be supported. This means that the surface state of the new layer of concrete will become the target of real-time distance data and real-time image data acquisition in step Z2, and its thickness deviation will become the basis for subsequent compensation.

[0099] After each cycle, the control system accumulates the completed spray thickness of each layer in real time and compares the accumulated spray thickness with the total spray thickness required by the design profile data. When the accumulated spray thickness reaches the total spray thickness, the control system determines that the multi-layer spray support has been completed and then stops repeating steps Z2 to Z6, finally completing the support work of the tunnel section to be supported and forming a complete initial support structure that meets the design requirements.

[0100] In summary, the above steps, through a cyclical mechanism, enable seamless transitions between the status monitoring and operational preparation of the next layer after each spraying layer is completed, achieving continuous closed-loop control over the entire multi-layer shotcrete support process. By using the newly completed sprayed concrete layer as the new monitoring target, it ensures that the real-time distance and image data in each cycle originate from the current construction surface, ensuring that subsequent texture feature extraction, thickness deviation calculation, timing determination, and path correction are always based on the latest actual state of the concrete layer, thus guaranteeing the targeting and accuracy of each spraying layer. By accumulating the completed spraying thickness in real time and comparing it with the total spraying thickness required by the design contour data, the cycle automatically stops when the total thickness is reached, ensuring that the final concrete support layer accurately meets the design contour requirements, avoiding quality issues such as over-spraying or under-spraying. Thus, while ensuring the overall structural integrity and load-bearing capacity of the multi-layer shotcrete support, it achieves automation, precision, and traceability of the construction process.

[0101] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0102] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0103] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for initial support construction of tunnels based on a concrete wet spraying robot, characterized in that, The method includes: Z1: Obtain the design contour data of the tunnel section to be supported, and perform path planning on the tunnel section to be supported to generate the foundation spraying path; Z2: Real-time acquisition of real-time distance data between the nozzle and the surface of the sprayed concrete layer, as well as real-time image data of the surface of the sprayed concrete layer; Z3: Extract texture features and gloss features from the real-time image data to obtain texture feature parameters and gloss feature parameters, and calculate the real-time thickness deviation value of the sprayed concrete layer surface based on the real-time distance data and the preset spray thickness at the corresponding position in the basic spraying path. Z4: Input the texture feature parameters, gloss feature parameters and the real-time thickness deviation value into the pre-constructed interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface; Z5: When the interlayer bonding timing coefficient meets the lower layer spraying start condition of the tunnel section to be supported, the real-time thickness deviation value is coupled with the preset spraying thickness to obtain the dynamic compensation coefficient of the nozzle, and the unexecuted part in the basic spraying path is corrected in real time based on the dynamic compensation coefficient to obtain the secondary spraying path of the tunnel section to be supported. Z6: Based on the secondary spraying path, perform the next layer of spraying operation on the surface of the sprayed concrete layer, repeating steps Z2 to Z6 until the multi-layer spraying support of the tunnel section to be supported is completed.

2. The tunnel initial support construction method based on a concrete wet spraying robot as described in claim 1, characterized in that, The process of acquiring the design contour data of the tunnel section to be supported and performing path planning on the tunnel section to be supported to generate the foundation spraying path includes: Obtain the design contour data of the tunnel section to be supported, and extract the contour line of the tunnel section to be supported; According to the preset spray width and overlap width, the nozzle movement trajectory of the tunnel section to be supported is constructed, and the nozzle movement trajectory is used as the basic spray path.

3. The tunnel initial support construction method based on a concrete wet spraying robot as described in claim 2, characterized in that, The process of extracting texture and gloss features from the real-time image data to obtain texture and gloss parameters, and calculating the real-time thickness deviation of the sprayed concrete layer surface based on the real-time distance data and the preset spray thickness at the corresponding position in the base spraying path, includes: The real-time image data is preprocessed, including median filtering for noise reduction and grayscale processing, to obtain a grayscale image of the sprayed concrete layer surface. Texture analysis is performed on the grayscale image to obtain the grayscale co-occurrence matrix of the grayscale image, and the contrast and energy of the grayscale co-occurrence matrix are integrated into the texture feature parameters of the grayscale image; The grayscale image is subjected to highlight region detection. Pixels in the grayscale image whose grayscale value is higher than a preset grayscale threshold are identified as highlight pixels. The total percentage of the highlight pixels in the grayscale image is used as a glossiness feature parameter. Based on the real-time distance data and the preset spraying thickness at the corresponding position in the basic spraying path, the real-time thickness deviation value of the sprayed concrete layer surface is calculated.

4. The tunnel initial support construction method based on a concrete wet spraying robot as described in claim 3, characterized in that, The step of calculating the real-time thickness deviation value of the sprayed concrete layer surface based on the real-time distance data and the preset spraying thickness at the corresponding position in the basic spraying path includes: Based on the normal direction on the nozzle movement trajectory, determine the nozzle attitude angle sequence in which the nozzle axis is consistent with the normal direction; Based on the total spray thickness of the design contour data and the preset number of layered sprays, the preset spray thickness on the nozzle movement trajectory is determined; For the sampling points on the nozzle movement trajectory, the difference between the real-time distance data and the preset injection thickness is used as the point thickness deviation value of the sampling point; The thickness deviation values ​​at the sampling points are integrated according to their spatial locations to generate a thickness deviation distribution map of a local area on the surface of the sprayed concrete layer.

5. A method for initial tunnel support construction based on a concrete wet spraying robot as described in claim 4, characterized in that, The step of inputting the texture feature parameters, gloss feature parameters, and real-time thickness deviation value into a pre-constructed interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface includes: Obtain historical construction data of the tunnel section to be supported. The historical construction data includes historical texture feature parameters, historical gloss feature parameters, historical thickness deviation distribution map, and interlayer bonding strength value. A multivariate nonlinear regression analysis was performed on the historical texture feature parameters, historical gloss feature parameters, historical thickness deviation distribution map, and interlayer bonding strength value to obtain model coefficients; Based on the model coefficients, a model for determining the interlayer bonding state of the tunnel section to be supported is constructed. The texture feature parameters, the gloss feature parameters, and the real-time thickness deviation value are input into the interlayer bonding state determination model to obtain the interlayer bonding timing coefficient of the sprayed concrete layer surface.

6. The tunnel initial support construction method based on a concrete wet spraying robot as described in claim 5, characterized in that, The formula for calculating the interlayer bonding timing coefficient is as follows: in, This refers to the interlayer bonding timing coefficient. The contrast of the gray-level co-occurrence matrix. The preset contrast reference value, , , , These are the weighting coefficients of the interlayer bonding state determination model. The energy of the gray-level co-occurrence matrix. The preset energy baseline value, The gloss characteristic parameter at the current moment, The preset gloss reference value, The average of the point thickness deviation values ​​for all sampling points. This is the preset thickness deviation tolerance. It is an exponential function.

7. The tunnel initial support construction method based on a concrete wet spraying robot as described in claim 4, characterized in that, When the interlayer bonding timing coefficient satisfies the lower-layer spraying initiation condition of the tunnel section to be supported, the real-time thickness deviation value is coupled with the preset spraying thickness to obtain the dynamic compensation coefficient of the nozzle. Based on the dynamic compensation coefficient, the unexecuted part of the basic spraying path is corrected in real time to obtain the secondary spraying path of the tunnel section to be supported, including: When the interlayer bonding timing coefficient is greater than or equal to the preset target lower threshold, the lower layer spraying start condition of the tunnel section to be supported is triggered. The dynamic compensation coefficient of the nozzle is obtained by coupling the real-time thickness deviation value with the preset injection thickness. Based on the dynamic compensation coefficient, the unexecuted portion of the basic spraying path is corrected in real time to obtain the secondary spraying path of the tunnel section to be supported.

8. A method for initial tunnel support construction based on a concrete wet spraying robot as described in claim 7, characterized in that, The formula for calculating the dynamic compensation coefficient is as follows: in, For any target location point on the basic injection path, For any point in the thickness deviation distribution map, For point The dynamic compensation coefficient at that location. The preset compensation strength coefficient, For the weight function, For any point in the thickness deviation distribution map The point thickness deviation value, For the point Centered on, with a preset radius The circular neighborhood, For point The preset spray thickness at the location.

9. A method for initial tunnel support construction based on a concrete wet spraying robot as described in claim 7, characterized in that, The process of real-time correction of the unexecuted portion of the basic spraying path based on the dynamic compensation coefficient to obtain the secondary spraying path of the tunnel section to be supported includes: The preset moving speed of each point on the basic injection path is mathematically multiplied by the reciprocal of the dynamic compensation coefficient, and the result of the mathematical product is used as the corrected moving speed of the corresponding point on the secondary injection path. The product of the preset injection flow rate at each point on the basic injection path and the dynamic compensation coefficient is used as the corrected injection flow rate at the corresponding point on the secondary injection path. Keeping the nozzle attitude angle sequence of each point on the basic injection path unchanged, and making real-time corrections to the unexecuted parts of the basic injection path based on the corrected moving speed and the corrected injection flow rate, a secondary injection path for the tunnel section to be supported is obtained.

10. A method for initial tunnel support construction based on a concrete wet spraying robot as described in claim 1, characterized in that, Based on the secondary spraying path, the next layer of spraying operation is performed on the surface of the already sprayed concrete layer, and steps Z2 to Z6 are repeated until the multi-layer sprayed support of the tunnel section to be supported is completed, including: After the current spraying operation of the tunnel section to be supported is completed, the end time of the spraying of the current layer in the tunnel section to be supported is taken as the start time of the spraying operation of the next layer, and the process returns to step Z2. The newly completed shotcrete layer in the tunnel section to be supported is used as the new shotcrete layer, and the lower shotcrete operation of the tunnel section to be supported is carried out. When the cumulative spray thickness reaches the total spray thickness required in the design profile data, the repetition stops, and the support of the tunnel section to be supported is completed.

Images