A needle beam trolley construction digital management method

Abstract

The application provides a needle beam trolley construction digital management method, and relates to the technical field of digital management. The method comprises the following steps: scanning the wall surface of a tunnel to be constructed by using a laser scanner; obtaining the position information of a plurality of working windows of a needle beam trolley; performing operation sequence enumeration with the maximum number of collaborative working windows as a constraint to generate a plurality of working window operation sequences; based on a three-dimensional model of the tunnel wall surface, taking the expected construction quality as a constraint and maximizing the construction efficiency as a target, performing pouring parameter and vibrating parameter optimization according to the plurality of working window operation sequences, and outputting an optimal construction scheme to perform digital construction management on the needle beam trolley. The technical problem of low construction efficiency caused by unreasonable operation sequence and uneven resource allocation in the construction process is solved. Through spatial perception of the tunnel to be constructed by three-dimensional modeling, multi-aspect optimization is realized in combination with digital management, and the efficiency and precision of needle beam trolley construction are improved.

Description

Technical Field

[0001] This application relates to the field of digital management technology, and in particular to a digital management method for needle beam trolley construction. Background Technology

[0002] A needle beam trolley is a specialized piece of equipment used in tunnel construction, primarily for concrete pouring and formwork support during tunnel lining. It typically consists of a needle beam, formwork, a walking system, and a hydraulic system, enabling movement and positioning within the tunnel. However, tunnel walls often have complex geometries, and the construction environment of each tunnel (such as curvature, inclination, and dimensional irregularities) varies. Traditional construction methods often rely on manual labor or rough models to plan the work sequence and resource allocation, failing to fully consider the specific shape of the tunnel walls and construction conditions. This leads to a mismatch between the work area and the construction equipment, increasing construction difficulty and reducing efficiency. Furthermore, due to a lack of consideration for the actual construction conditions at different work windows, an unreasonable work sequence results in frequent equipment movement or waiting between different work windows, wasting time and resources and reducing construction efficiency.

[0003] In summary, existing technologies suffer from technical problems such as unreasonable work sequence and uneven resource allocation due to insufficient consideration of the specific shape of the tunnel wall and construction conditions, which further affect construction efficiency. Summary of the Invention

[0004] The purpose of this application is to provide a digital management method for needle beam trolley construction, in order to solve the technical problems in the prior art that, due to insufficient consideration of the specific shape of the tunnel wall and construction conditions, the work sequence is unreasonable and the resource allocation is uneven, which further affects the construction efficiency.

[0005] In view of the above problems, this application provides a digital management method for needle beam trolley construction, wherein the digital management method for needle beam trolley construction includes: scanning the wall of the tunnel to be constructed using a laser scanner to construct a three-dimensional model of the tunnel wall; obtaining the position information of several working windows of the needle beam trolley to construct the working window position distribution; using the maximum number of collaborative working windows as a constraint, enumerating the operation sequence according to the working window position distribution to generate multiple working window operation sequences; based on the three-dimensional model of the tunnel wall, using the expected construction quality as a constraint and maximizing construction efficiency as an objective, optimizing the pouring parameters and vibration parameters according to the multiple working window operation sequences, outputting the optimal construction scheme, and performing digital construction management of the needle beam trolley.

[0006] Optionally, a laser scanner is used to scan the wall of the tunnel to be constructed at multiple preset scanning positions to obtain multiple wall point cloud datasets; after point cloud registration and point cloud cleaning of the multiple wall point cloud datasets, mesh modeling is performed to generate a three-dimensional model of the tunnel wall.

[0007] Optionally, based on a preset construction start point, with the maximum number of collaborative work windows as the construction constraint for the same period, and with adjacent work windows as the sequence constraint, the operation path is enumerated according to the position distribution of the work windows to generate multiple work window operation sequences.

[0008] Optionally, using the three-dimensional model of the tunnel wall as a comparison constraint, a sample dataset is retrieved to train and obtain a concrete quality predictor; a first working window operation sequence is randomly selected from the multiple working window operation sequences; using the concrete quality predictor, with the constraint of meeting the expected construction quality and the goal of maximizing construction efficiency, the pouring parameters and vibration parameters are optimized according to the first working window operation sequence to output a first optimized construction scheme, and multiple optimized construction schemes of the multiple working window operation sequences are analyzed and obtained in sequence.

[0009] Optionally, using the three-dimensional model of the tunnel wall as a comparison constraint and satisfying a preset similarity threshold as a condition, historical pinion beam trolley construction records are retrieved to collect a sample dataset. The sample data includes sample work window operation sequences, sample pouring parameters, and sample vibration parameters. Concrete quality parameters under different sample work window operation sequences, sample pouring parameters, and sample vibration parameters are collected to obtain a sample quality parameter set. Using the sample dataset and sample quality parameter set, the generator and discriminator of a generative adversarial network are trained until convergence to obtain the concrete quality predictor.

[0010] Optionally, within the threshold values ​​for pouring parameters and vibration parameters, several pouring parameters and several vibration parameters are randomly selected and combined with the first work window operation sequence to obtain several construction schemes; using the concrete quality predictor, the several construction schemes are evaluated respectively, and several quality parameters are output; according to the expected construction quality, the several quality parameters are screened respectively, and multiple qualified construction schemes are output; with the goal of maximizing construction efficiency, the pouring parameters and vibration parameters are optimized based on the multiple qualified construction schemes, and a first optimized construction scheme is output.

[0011] Optionally, operation simulation is performed based on the multiple qualified construction schemes, outputting multiple simulated operation durations; the qualified construction schemes are considered as initial solutions, and the multiple qualified construction schemes are arranged in ascending order of simulated operation duration to generate an initial solution sequence; the first K solutions of the initial solution sequence are selected as optimal solutions, and the last J solutions are selected as inferior solutions, and equi-clustering is performed on the J inferior solutions centered on the K optimal solutions to obtain K solution sets, where J is a multiple of K and Q is an integer greater than 5; within each solution set, the inferior solution with the longest simulated construction duration is selected as a poor solution, determining K poor solutions; based on the K optimal solutions and K poor solutions, according to the optimization-avoidance strategy, the inferior solutions in the K solution sets are further optimized according to a preset optimization step size. One adjustment outputs K updated solution sets. If the updated inferior solution does not meet the threshold of the pouring parameters or the threshold of the vibration parameters, a construction scheme is randomly selected for replacement. The K updated solution sets are identified. If the simulation operation time of the inferior solution in the same solution set is less than the simulation operation time of the superior solution, the inferior solution replaces the superior solution. If the simulation operation time of the inferior solution in the same solution set is greater than the simulation operation time of the poor solution, the inferior solution replaces the poor solution. Iterative optimization is performed until the preset number of optimization attempts is met, and K current solution sets are output. The optimal solution set is evaluated and determined, and the superior solution of the optimal solution set is selected as the first optimized construction scheme. The optimal solution set is the solution set with the smallest sum of simulation operation times among the K current solution sets.

[0012] Optionally, the average simulation operation time of multiple inferior solutions in each solution set is calculated, and the deviations from the simulation operation time of the superior and inferior solutions are calculated respectively to determine the superior solution deviation and the inferior solution deviation. If the superior solution deviation is greater than or equal to the inferior solution deviation, the optimization strategy is set as the optimization strategy, and the inferior solutions in the K solution sets are adjusted once according to the preset optimization step size, with the superior solution as the adjustment direction. If the superior solution deviation is less than the inferior solution deviation, the optimization strategy is set as the inferior solution avoidance strategy, and the inferior solutions in the K solution sets are adjusted once according to the preset optimization step size, with the inferior solution being moved away from the inferior solution as the adjustment direction.

[0013] Optionally, the simulation operation time deviations between multiple inferior solutions and superior or inferior solutions within each solution set are calculated separately to obtain multiple superior solution time deviations and multiple inferior solution time deviations. The average superior solution time deviation and the average inferior solution time deviation are then calculated. If the optimization strategy is an optimization-oriented strategy, the ratio of the superior solution time deviation to the average superior solution time deviation is set as an adjustment coefficient to obtain multiple adjustment coefficients. The initial optimization step size is then adjusted to obtain multiple adjusted optimization step sizes as preset optimization step sizes. If the optimization strategy is an inferior-avoidance strategy, the ratio of the inferior solution time deviation to the average inferior solution time deviation is set as an adjustment coefficient to adjust the initial optimization step size to obtain multiple adjusted optimization step sizes as preset optimization step sizes.

[0014] Optionally, the scheme with the shortest simulated operation time is selected as the optimal construction scheme from among the multiple optimized construction schemes of the multiple work window operation sequences.

[0015] The technical solution provided in this application has at least the following beneficial effects:

[0016] A three-dimensional model of the tunnel wall is constructed by scanning the tunnel wall using a laser scanner. The position information of several working windows of the needle beam trolley is obtained, and the working window position distribution is constructed. Constrained by the maximum number of collaborative working windows, the operation sequence is enumerated based on the working window position distribution to generate multiple working window operation sequences. Based on the three-dimensional model of the tunnel wall, and with the constraint of meeting the expected construction quality and the goal of maximizing construction efficiency, the pouring and vibration parameters are optimized according to the multiple working window operation sequences to output the optimal construction plan and achieve digital construction management of the needle beam trolley. In other words, the laser scanner spatially perceives the tunnel wall, generates a three-dimensional model that realistically reflects the tunnel's specific shape and structure, constructs the working window position distribution, and generates multiple working window operation sequences considering the working window position distribution and the maximum number of collaborative working windows. Combined with the three-dimensional model of the tunnel wall, the pouring and vibration parameters for each working window operation sequence are optimized to ensure construction quality while maintaining the efficiency and accuracy of the needle beam trolley construction.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the digital management method for needle beam trolley construction according to this application.

[0020] Figure 2 This is a flowchart illustrating the process of obtaining the first optimized construction scheme in the digital management method for needle beam trolley construction of this application. Detailed Implementation

[0021] This application provides a digital management method for needle beam trolley construction, solving the technical problems in existing technologies where insufficient consideration of the specific shape and construction conditions of the tunnel wall leads to unreasonable work sequences and uneven resource allocation, further affecting construction efficiency. By using a laser scanner to spatially perceive the tunnel wall under construction, a 3D model is generated, realistically reflecting the tunnel's specific shape and structure. The method constructs the distribution of work windows, considering their location and the maximum number of collaborative work windows, generating multiple work window work sequences. Combined with the 3D model of the tunnel wall, the pouring and vibration parameters for each work window sequence are optimized, ensuring construction quality while improving the efficiency and accuracy of needle beam trolley construction.

[0022] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0023] For examples, please refer to the appendix. Figure 1 This application provides a digital management method for needle beam trolley construction, wherein the digital management method for needle beam trolley construction specifically includes the following steps:

[0024] S100: Use a laser scanner to scan the walls of the tunnel under construction to create a three-dimensional model of the tunnel walls.

[0025] Furthermore, this application S100 includes:

[0026] S110: At multiple preset scanning positions, a laser scanner is used to scan the wall of the tunnel to be constructed to obtain multiple wall point cloud datasets; S120: After point cloud registration and point cloud cleaning of the multiple wall point cloud datasets, mesh modeling is performed to generate and construct a three-dimensional model of the tunnel wall.

[0027] Specifically, a needle beam trolley is a specialized piece of equipment used in tunnel construction, primarily for concrete pouring and formwork support during tunnel lining. It typically consists of a needle beam, formwork, a walking system, and a hydraulic system, enabling movement and positioning within the tunnel to improve construction efficiency and quality. Pre-set scanning positions refer to planning multiple fixed locations for the laser scanner before tunnel construction. Scanning the tunnel walls from these locations ensures the acquisition of three-dimensional data for all parts of the tunnel. Multiple pre-set scanning positions need to be determined based on the actual structure and shape of the tunnel to ensure that scanning from each position covers all parts of the tunnel wall, avoiding blind spots or missed areas. Typically, scanning positions are set at different curvatures and angles of the tunnel to ensure comprehensiveness. For example, for a 500-meter-long tunnel with varied curves, a scanning position needs to be set every 50 meters, with scanning points also set at different heights on both sides of the tunnel.

[0028] At the preset scanning position, the laser scanner begins operation, performing spatial sensing of the tunnel under construction. By emitting and receiving reflected signals from a laser beam, it records point cloud data on the tunnel wall, recording the three-dimensional coordinates of the tunnel wall. The coordinates of each point are calculated by measuring the time difference of the reflected laser light. Each point cloud dataset contains the three-dimensional coordinates (x, y, z) of each point on the tunnel wall, which can represent the accurate geometry of the tunnel's inner wall. The laser scanner is a measuring tool that calculates the position of an object's surface by measuring the time difference between the emission and reception of reflected laser light, thereby obtaining high-precision three-dimensional spatial point cloud data.

[0029] A point cloud is a collection of multiple points in space acquired by a laser scanner, where each point represents the coordinates (x, y, z) of a specific location on the surface of an object. In tunnel construction, a wall point cloud dataset refers to the surface information of the tunnel wall collected from multiple scanning positions using a laser scanner, containing the three-dimensional coordinate data of the tunnel wall.

[0030] Point cloud registration is performed on multiple wall point cloud datasets to merge point cloud datasets from different scan locations into a unified 3D dataset. Since the coordinate systems at each scan location are different, a registration algorithm is needed to align these data sets so that they match within the same coordinate system. Point cloud registration is the process of spatially aligning multiple point cloud datasets from different locations using an algorithm. Because the coordinate systems at each scan location are different, registration ensures that these datasets are correctly aligned within the same coordinate system, guaranteeing data integrity and consistency.

[0031] Point cloud registration typically uses the Iterative Nearest Point (ICP) algorithm, which continuously optimizes the matching degree between point clouds, ultimately merging all point cloud data into a unified 3D dataset. The ICP algorithm is a commonly used point cloud registration method for aligning two or more point cloud datasets. The core idea of ​​the algorithm is to find an optimal spatial transformation (such as translation or rotation) that maximizes the matching degree between two point clouds through iterative processing. In each iteration, the ICP algorithm optimizes the matching by minimizing the point-to-point distance. For each pair of point cloud datasets, points from the first point cloud are paired with the points in the second point cloud with the smallest distance, using the nearest neighbor algorithm to select matching point pairs. Each time a pair of corresponding points is selected, their Euclidean distance is calculated. Once all matching point pairs are found, the ICP algorithm calculates a transformation matrix containing parameters for rotation and translation. The goal is to find an optimal transformation that aligns the points in the target point cloud as closely as possible to the points in the reference point cloud. The calculated transformation is applied to the target point cloud, and new matching point pairs are calculated. Then, a new transformation matrix is ​​calculated again to optimize the alignment between the point clouds. By repeatedly performing this process, the distance between point clouds is gradually reduced until the matching degree reaches a set threshold, or the number of iterations reaches a preset upper limit. When the matching degree meets the preset standard, or when the number of iterations exceeds the set limit, the ICP algorithm stops, and the registration process ends. At this point, the registration of the two point cloud datasets is complete. For multiple wall point cloud datasets, this step is repeated until all point cloud data are finally merged into a unified 3D dataset.

[0032] Point cloud data may contain noise or outliers during acquisition (such as erroneous data caused by poor reflection or equipment malfunction). Point cloud cleaning removes this invalid data and retains valid, accurate data. The cleaning process may use filtering algorithms, such as removing outliers or low-quality data. Some outliers (such as irrelevant data caused by equipment scanning angle issues) may appear in the point cloud data; these points will be removed. Noise points are usually caused by environmental factors (such as dust or light interference in the tunnel) or equipment issues (such as errors in the laser scanner), and their presence affects the accuracy of the point cloud data. Redundant points can be removed by comparing the distances between adjacent points. If the distance between two points is less than a preset threshold (usually determined based on the specific construction requirements of the tunnel to be constructed), they are considered redundant, and one of the points can be removed. Holes may exist in the point cloud data due to scanning angle or object occlusion. Hole filling can be achieved using various methods, such as interpolation methods based on neighboring points.

[0033] Mesh modeling technology is used to construct a 3D model of the tunnel wall based on registered and cleared point cloud data. Mesh modeling transforms point cloud data into a mesh model composed of triangular facets, accurately representing the shape of the tunnel wall and providing precise 3D visualization data for subsequent construction. Mesh modeling typically involves two steps: mesh generation and mesh optimization. Mesh generation creates a 3D mesh model from the point cloud data, while mesh optimization improves the geometric accuracy and smoothness of the model by adjusting the mesh's topology and node positions.

[0034] Point cloud data obtained through laser scanning can accurately capture the geometry of the tunnel wall. The registered and cleaned data will ensure the high accuracy of the model. Through 3D modeling, the actual shape of the tunnel is reflected, and precise construction can be carried out based on the model, thereby reducing the risks and losses caused by shape errors or equipment incompatibility.

[0035] S200: Obtain the position information of several working windows of the needle beam trolley and construct the working window position distribution.

[0036] Specifically, the position of each working window on the needle beam trolley is determined by measuring the position of each working window relative to the trolley frame. That is, based on the needle beam trolley structural design drawings, the relative fixed position parameters of each working window on the trolley frame are determined (such as the offset distance, height, and front-to-back position relative to the needle beam centerline). A needle beam trolley is a specialized piece of equipment used in tunnel construction, primarily for concrete pouring and formwork support in tunnel lining construction. It typically consists of needle beams, formwork, a walking system, and a hydraulic system, enabling it to move and position itself within the tunnel, assisting in tasks such as tunnel lining. In tunnel construction, a working window usually refers to the specific coordinates of each working window on the needle beam trolley within the tunnel; each working window has a defined coordinate system that determines the area where the trolley operates.

[0037] Once the scope of the working windows is defined, their precise locations within the tunnel are obtained, typically represented by three-dimensional coordinates. Using positioning sensors deployed in the tunnel (such as laser positioning devices and inertial measurement units, IMUs), the real-time position coordinates (X, Y, Z) and attitude (pitch, yaw, roll) of the needle beam trolley are acquired. After obtaining the coordinates of all working windows, these locations are integrated using a 3D visualization tool to generate a spatial distribution map of the working windows, showing their positions and distribution relationships within the tunnel. The working window location distribution refers to the spatial arrangement of multiple working windows within the tunnel. Based on the spatial distribution of the working windows, combined with the tunnel construction schedule, trolley operating capacity, and environmental constraints, the operating time windows, operating sequences, trolley movement paths, and resource allocation for each working window are determined. By acquiring the working window location information of the needle beam trolley and constructing its spatial distribution, precise operational area planning can be provided for tunnel construction.

[0038] S300: With the maximum number of collaborative work windows as a constraint, the job sequence is enumerated according to the position distribution of the work windows to generate multiple work window job sequences.

[0039] Based on the preset construction start point, with the maximum number of collaborative work windows as the construction constraint for the same period, and with adjacent work windows as the sequence constraint, the operation path is enumerated according to the position distribution of the work windows to generate multiple work window operation sequences.

[0040] Specifically, the maximum number of collaborative working windows refers to the maximum number of working windows that can operate simultaneously during construction. That is, the maximum number of working windows that the trolley can collaboratively operate across multiple working windows within a given time period. For example, assuming the trolley can handle a maximum of two working windows simultaneously during a certain time period, then the maximum number of collaborative working windows for that time period is 2. This number is a constraint, limited by factors such as the trolley's operational capacity, the spatial requirements of the working windows, the specific construction environment of the tunnel (such as tunnel size, curvature, and inclination), and construction needs. Specifically, the maximum simultaneous operation capacity of the trolley is assessed based on factors such as its moving speed, working accuracy, and the size of the working area; a reasonable maximum number of collaborative working windows is determined by considering the tunnel's geometry, spatial constraints, and operational difficulty.

[0041] The purpose of job sequence enumeration is to find the optimal job sequence based on the distribution of work windows, the constraint of the maximum number of collaborative work windows, and the job sequence constraints between adjacent work windows. The job sequence directly affects the movement path and work efficiency of the trolley; therefore, a correct job sequence can maximize construction efficiency and reduce the ineffective movement of the trolley within the tunnel. Different job sequences are generated based on the distribution of work window positions in the tunnel and the constraint of the maximum number of collaborative work windows. The optimal sequence is selected by enumerating all possible job sequences. Each sequence should follow the sequence constraints of adjacent work windows to ensure that the trolley can move reasonably to the next work window for work. For example, suppose a tunnel construction project has 4 work windows and a maximum number of collaborative work windows of 2. Through enumeration, the following possible job sequences are obtained: Sequence 1: Work windows 1 and 2 work simultaneously, then work windows 3 and 4 work simultaneously; Sequence 2: Work windows 1 and 3 work simultaneously, then work windows 2 and 4 work simultaneously; Sequence 3: Work windows 1 and 4 work simultaneously, then work windows 2 and 3 work simultaneously.

[0042] Work path enumeration, after determining the work sequence, plans the specific path for the trolley to move from one work window to another. The goal of work path enumeration is to generate multiple feasible paths from the preset construction start point to the work window, ensuring that the order of these paths conforms to preset requirements. Work path generation must consider at least the construction start point, the relative positions of the work windows, order constraints, and the maximum number of collaborative work windows. A path planning algorithm, such as the shortest path algorithm in graph theory, is used. Path optimization in tunnel construction can be achieved by treating the tunnel as a graph model, where work windows and construction equipment positions are vertices, and the paths between work windows are edges. The weight of each edge is determined based on distance, time, or other construction costs. The tunnel is modeled, with each work window, start point, and target point considered as vertices in the graph. Each path in the tunnel (such as a passage from one work window to another, the trolley's movement route, etc.) is considered as an edge in the graph. Each edge in the graph is assigned a weight based on the actual distance between work windows or the estimated travel time.

[0043] When planning the work path, the distance between adjacent work windows and the trolley's mobility must be considered to ensure the path is as short as possible and reduce the trolley's idle time. The starting point, work sequence, and work window location distribution are preset. A path planning algorithm is used to determine the shortest or optimal path for each work sequence. Shortest path algorithms from graph theory are used to calculate the shortest path for each work sequence. These algorithms can generate the optimal movement path for the trolley based on the work window's location and the distance between adjacent work windows. Work windows are considered nodes in a graph, and adjacent work windows are edges; shortest path algorithms from graph theory are used to optimize the trolley's movement path.

[0044] In tunnel construction, each work window can be considered a node in a graph. The location of the work window, the starting point of the trolley, and other work areas can be considered as multiple nodes in the graph. Nodes are connected by edges, and the weight of each edge is the distance between the work windows or the time required for the trolley to move. Adjacent work windows refer to work windows that are spatially close or connected to each other. During construction, when the trolley moves from one work window to another, their relative positions must be considered. Adjacent work windows can be represented by edges, and the weights of these edges are usually the spatial distance between the work windows or the time required for the trolley to travel. Once all work windows are considered as nodes in the graph, and adjacent nodes are connected by edges, the shortest path from one work window to another is calculated according to different work sequences, thus determining the trolley's movement route. Specifically, based on the location information of the work windows, a graph is constructed, with work windows as nodes, and the connections and relative distances between work windows as edge weights. Based on the characteristics of the graph (e.g., all edge weights are non-negative), a suitable shortest path algorithm is selected; Dijkstra's algorithm is typically used to calculate the shortest path from one work window to other work windows.

[0045] Based on the generated graph model, a suitable shortest path algorithm, such as Dijkstra's algorithm, is selected to optimize the job path. First, a graph (including nodes and edges) is defined, where each edge has a non-negative weight. A distance table is created to record the shortest distances from the source node to all nodes. The initial distance to the source node is 0, and the initial distances to other nodes are ∞. A set S of nodes with determined shortest paths is created, initially empty. A set U of nodes with undetermined shortest paths is created, initially containing all nodes. From node set U, select a node u that is not in set S and has the smallest distance (i.e., the node on the current shortest path), add node u to node set S, and remove node u from node set U. For each neighbor node v of the current node u, if the path to node v via node u is shorter than the original distance, update the value of node v in the distance table. Repeat the above steps until the shortest paths to all nodes are determined, or node set U is empty. Finally, the distance table stores the shortest paths from the source node to all other nodes. For example, suppose the graph contains nodes A, B, and C; the edge weights are as follows: A to B = 4, A to C = 2, B to C = 1, C to A = 2, B to A = 4, and C to B = 1. We need to find the shortest path from node A to all other nodes. Initialize the distance of node A to 0, and the distances of nodes B and C to ∞. Starting from node A, update the shortest paths of neighboring nodes B and C, resulting in a distance of 0 for node A, 4 for node B, and 2 for node C. Select node C, which has the shortest path, and update the shortest paths of neighboring nodes A and B. Node A's distance is 0 (no update, as 0 is already the shortest path), and node B's distance is 3 (A to C = 2 + B to C = 1). The updated distances are 0 for node A, 3 for node B, and 2 for node C, which is the final output shortest path.

[0046] For each work sequence, the shortest path algorithm is used to calculate the shortest path for the trolley from one work window to another. This ensures the trolley follows the order of adjacent work windows during movement and minimizes travel distance and time, resulting in multiple work window work sequences. By enumerating work sequences and work paths, the construction team can generate multiple possible construction plans, helping to evaluate the advantages and disadvantages of different construction strategies and select the most suitable plan to maximize construction efficiency and minimize costs.

[0047] S400: Based on the three-dimensional model of the tunnel wall, with the constraint of meeting the expected construction quality and the goal of maximizing construction efficiency, the pouring parameters and vibration parameters are optimized according to the multiple work window operation sequences, and the optimal construction scheme is output to digitally manage the needle beam trolley.

[0048] Furthermore, this application S400 includes:

[0049] S410: Using the three-dimensional model of the tunnel wall as a comparison constraint, retrieve the sample dataset and train a concrete quality predictor; S420: Randomly select a first working window operation sequence from the multiple working window operation sequences; S430: Using the concrete quality predictor, with the constraint of meeting the expected construction quality and the goal of maximizing construction efficiency, optimize the pouring parameters and vibration parameters according to the first working window operation sequence, output a first optimized construction scheme, and sequentially analyze and obtain multiple optimized construction schemes from the multiple working window operation sequences.

[0050] S411: Using the 3D model of the tunnel wall as a comparison constraint and satisfying a preset similarity threshold, retrieve historical pinion beam trolley construction records and collect a sample dataset. The sample data includes sample work window operation sequences, sample pouring parameters, and sample vibration parameters. S412: Collect concrete quality parameters under different sample work window operation sequences, sample pouring parameters, and sample vibration parameters to obtain a sample quality parameter set. S413: Using the sample dataset and sample quality parameter set, train the generator and discriminator of the generative adversarial network until convergence to obtain the concrete quality predictor.

[0051] Specifically, the comparison constraint refers to the requirement that historical construction data must match the current 3D model of the tunnel wall during retrieval. The similarity threshold is a standard value used to determine whether the current tunnel wall is sufficiently similar to historical data. By setting the threshold, the historical construction records that best match the current tunnel can be selected. Under the comparison constraint, past needle beam trolley construction records are retrieved. These records contain the work paths, work sequences, and corresponding construction parameters (such as pouring and vibration parameters) used in the historical construction process. From these, historical needle beam trolley construction records that match the 3D model of the tunnel wall are extracted to obtain a sample dataset, including sample work window work sequences, sample pouring parameters, and sample vibration parameters.

[0052] The sample work window operation sequence records the operation order of the trolley between different work windows under different construction scenarios; the sample pouring parameters record the relevant parameters used in the concrete pouring process, such as pouring speed and pouring depth; the sample vibration parameters record the parameters used in the vibration process, such as vibration frequency and vibration time.

[0053] Simultaneously, based on different sample work window operation sequences, sample pouring parameters, and sample vibration parameters, corresponding concrete quality parameters are collected, namely, concrete quality parameters (such as compressive strength, density, uniformity, etc.) under different operation sequences and parameter configurations. This reflects the construction quality under various conditions and forms a sample quality parameter set, including concrete quality parameters such as compressive strength, density, and durability.

[0054] Using the collected sample dataset and sample quality parameter set, a Generative Adversarial Network (GAN) is employed for training. During training, the generator produces concrete quality predictions based on the input sample data (such as work order and work parameters), while the discriminator judges the generated quality data to determine whether it meets the quality standards of the tunnel wall 3D model. The generator and discriminator continuously adjust their parameters through adversarial training, ultimately generating a predictor capable of efficiently predicting concrete quality. The training process continues until the model's loss function converges, meaning the generator can accurately predict concrete quality parameters, at which point training stops, resulting in the concrete quality predictor.

[0055] Generative Adversarial Networks (GANs) are deep learning models consisting of a generator and a discriminator. The generator produces data, while the discriminator determines the authenticity of the data. Through adversarial training, both the generator and discriminator improve, ultimately generating high-quality sample data. The generator produces fake data based on the input data, while the discriminator identifies whether the input data is real. During training, they compete and continuously improve to ultimately generate high-quality output.

[0056] Based on the actual construction environment, concrete material characteristics, and construction machinery performance, threshold values ​​for pouring and vibration parameters are determined. For each work window sequence, within the threshold values ​​for both pouring and vibration parameters, several pouring and vibration parameters are randomly generated. These are then combined with the work window sequence to generate multiple construction plans. A concrete quality predictor is used to evaluate these plans, selecting the one that meets the expected construction quality. The pouring and vibration parameters are then optimized to obtain the optimal construction plan for each work window sequence—multiple optimized construction plans. Finally, the optimal construction plan is output for digital construction management of the needle beam trolley.

[0057] By optimizing pouring and vibration parameters, construction efficiency can be improved while ensuring concrete quality, reducing unnecessary construction time and resource waste. Concrete quality predictors can accurately predict concrete quality under different work sequences and parameters based on historical data and actual construction conditions, thereby avoiding quality problems caused by improper parameter configuration.

[0058] Further details are attached. Figure 2 As shown, this application also includes the following steps:

[0059] S431: Within the threshold values ​​for pouring parameters and vibration parameters, randomly select several pouring parameters and several vibration parameters, and combine them with the first work window operation sequence to obtain several construction schemes; S432: Using the concrete quality predictor, evaluate the several construction schemes respectively and output several quality parameters; S433: Based on the expected construction quality, screen the several quality parameters respectively and output multiple qualified construction schemes; S434: With the goal of maximizing construction efficiency, optimize the pouring parameters and vibration parameters based on the multiple qualified construction schemes and output the first optimized construction scheme.

[0060] Specifically, the first work sequence is randomly selected from multiple preset work window sequences. To ensure the effectiveness of pouring and vibration operations, threshold values ​​for pouring and vibration parameters are determined based on the actual construction environment, concrete material characteristics, and construction machinery performance. The pouring parameter threshold defines the range limits for various parameters during the pouring operation, including pouring speed and pressure, determining the fluidity and stability of the concrete during pouring. The vibration parameter threshold defines the range limits for various parameters during the vibration operation, typically involving vibration frequency and vibration time, affecting the compactness and strength of the concrete. Within the set threshold range, multiple pouring and vibration parameters are randomly selected. These parameters are then combined with the selected work window sequence to obtain multiple different construction schemes, each corresponding to a specific parameter configuration. For example, assuming the threshold range for pouring rate is 3-5 m³ / h and the threshold range for vibration frequency is 1000-1500 times / min, three pouring rates (e.g., 3.5 m³ / h, 4.0 m³ / h, 4.5 m³ / h) and three vibration frequencies (e.g., 1050 times / min, 1200 times / min, 1350 times / min) are randomly selected and combined into several construction schemes.

[0061] Using a concrete quality predictor, the parameters corresponding to several construction schemes are input into the predictor to evaluate these schemes. The predictor forecasts the concrete quality (such as compressive strength and density) for each construction scheme by inputting data such as the work sequence, pouring parameters, and vibration parameters. For example, evaluating a construction scheme using the predictor yields the following quality parameters: Strength: 35 MPa, Uniformity: 0.98 (the closer the quality is to 1, the better), Durability: 70 (meets requirements).

[0062] Based on the operational requirements of the tunnel to be constructed, the expected construction quality is determined, including strength, density, and uniformity. Based on this expected quality, several quality parameters are screened, and multiple schemes that meet the expected quality are selected as qualified construction schemes. That is, every quality parameter of each construction scheme meets the expected construction quality parameters; if even one parameter is not met, the scheme is deemed unqualified and must be eliminated. With the goal of maximizing construction efficiency, the qualified construction schemes are further optimized by adjusting the pouring and vibration parameters to improve construction efficiency while ensuring concrete quality. For example, if the quality parameters of a construction scheme are lower than expected (e.g., strength only 28 MPa, uniformity 0.94), these schemes will be excluded. Schemes that meet the standards (e.g., strength 35 MPa, uniformity 0.98, durability 70) will be retained.

[0063] After optimizing the pouring and vibration parameters based on multiple qualified construction schemes, a first optimal construction scheme is output. This maximizes construction efficiency and reduces resource waste and time delays while meeting quality standards. For example, assuming a qualified construction scheme has a pouring rate of 4.0 m³ / h and a vibration frequency of 1200 times / min, the algorithm adjusts these parameters to obtain a new optimized scheme (e.g., a pouring rate of 4.2 m³ / h and a vibration frequency of 1150 times / min), improving construction efficiency while ensuring concrete quality. By optimizing pouring and vibration parameters, construction efficiency is improved and construction time is reduced while ensuring concrete quality. The prediction results from a concrete quality predictor ensure that each construction scheme meets quality standards, avoiding quality problems caused by improper work sequence or parameter configuration.

[0064] Furthermore, this application also includes:

[0065] Among the multiple optimized construction schemes of the multiple work window operation sequences, the scheme with the shortest simulated operation time is selected as the optimal construction scheme.

[0066] Specifically, the multiple optimized construction schemes are a series of schemes derived from optimizing different work window operation sequences. Each scheme is based on different optimization results of operation sequences, pouring parameters, and vibration parameters, and has different parameter settings (pouring, vibration, etc.) and corresponding operation durations. Simulations are performed on multiple optimized construction schemes corresponding to multiple work window operation sequences to calculate the time required to complete the construction task, including work window operation time, trolley movement time, and interval time between work windows. Work window operation time is the time required for pouring and vibration operations at each work window, calculated using parameters from the optimization process (such as pouring pressure and vibration frequency); trolley movement time is the time the trolley takes to move through the tunnel, depending on the tunnel geometry and trolley speed; interval time between work windows is the switching time between different work windows, including equipment adjustments and material transportation. By weighting and calculating each time factor, the simulated operation duration for each construction scheme is obtained.

[0067] Simulated work duration estimates the total construction time by considering factors such as the operational requirements of each work window, trolley movement time, and the time for pouring and vibration. Among multiple optimized construction schemes, the scheme with the shortest simulated work duration is selected as the optimal scheme by comparing the simulated work durations of each scheme. This means that the construction task can be completed in the shortest time while ensuring quality, thereby improving construction efficiency and reducing costs. For example, suppose there are three optimized construction schemes with simulated work durations of: Scheme 1: 8 hours, Scheme 2: 6.5 hours, and Scheme 3: 7.2 hours. Scheme 2 has the shortest simulated work duration and is therefore selected as the optimal construction scheme. The final construction plan is generated based on the selected scheme with the shortest simulated work duration. By selecting the construction scheme with the shortest work duration, construction efficiency can be significantly improved, construction time reduced, and project costs lowered.

[0068] Furthermore, step S434 of this application also includes the following steps:

[0069] Based on the multiple qualified construction schemes, operation simulations are performed, outputting multiple simulated operation durations. The qualified construction schemes are considered as initial solutions, and arranged in ascending order of simulated operation duration to generate an initial solution sequence. The first K solutions of the initial solution sequence are selected as optimal solutions, and the last J solutions as inferior solutions. Using the K optimal solutions as centers, the J inferior solutions are subjected to equal-value clustering to obtain K solution sets, where J is a multiple of K, and Q is an integer greater than 5. Within each solution set, the inferior solution with the longest simulated operation duration is selected as a poor solution, determining K poor solutions. Based on the K optimal solutions and K poor solutions, following a strategy of seeking the best and avoiding the worst, the inferior solutions within the K solution sets are processed once according to a preset optimization step size. Adjustments are made, and K updated solution sets are output. If the updated inferior solution does not meet the threshold of the pouring parameters or the threshold of the vibration parameters, a construction scheme is randomly selected for replacement. The K updated solution sets are identified. If the simulation operation time of the inferior solution in the same solution set is less than the simulation operation time of the superior solution, the inferior solution replaces the superior solution. If the simulation operation time of the inferior solution in the same solution set is greater than the simulation operation time of the poor solution, the inferior solution replaces the poor solution. Iterative optimization is performed until the preset number of optimizations is met, and K current solution sets are output. The optimal solution set is evaluated and determined, and the superior solution of the optimal solution set is selected as the first optimized construction scheme. The optimal solution set is the solution set with the smallest sum of simulation operation time among the K current solution sets.

[0070] Specifically, multiple qualified construction schemes are simulated. These schemes are input into a 3D model of the tunnel wall for construction operation simulation. This involves inputting specific parameters (such as work sequence, pouring parameters, and vibration parameters) of each qualified scheme into the 3D model of the tunnel wall to obtain simulation results for each scheme, and determining the corresponding simulated operation time accordingly. In other words, after the simulation is completed, the simulated operation time for each qualified construction scheme is output. Construction simulation typically involves simulating the entire construction process, with time units including hours and minutes. The simulation results for each scheme will provide the specific completion time for each construction task, thus obtaining the total operation time. The simulated operation time is estimated by considering factors such as the operational requirements of each work window, the trolley movement time, and the pouring and vibration times. Similar to the steps explained above, the time required to complete the construction tasks is calculated, including work window operation time, trolley movement time, and interval time between work windows, resulting in multiple simulated operation times.

[0071] For example, suppose there are three qualified construction schemes, A, B, and C, involving different pouring and vibration parameters: Scheme A: pouring speed of 1.5 m³ / h, vibration time of 12 minutes, transportation time of 10 minutes, and estimated operation time of 60 minutes; Scheme B: pouring speed of 2.0 m³ / h, vibration time of 15 minutes, transportation time of 8 minutes, and estimated operation time of 55 minutes; Scheme C: pouring speed of 1.8 m³ / h, vibration time of 10 minutes, transportation time of 12 minutes, and estimated operation time of 58 minutes. Simulation results show that the simulated operation time for Scheme A is 62 minutes, for Scheme B it is 54 minutes, and for Scheme C it is 60 minutes.

[0072] Based on multiple qualified construction schemes, an initial solution sequence is generated by arranging them in ascending order of simulated operation time. This helps identify which schemes have the greatest advantage in construction efficiency. The initial solution sequence is generated by arranging multiple qualified construction schemes in ascending order of simulated operation time. The first K solutions in the initial solution sequence are designated as optimal solutions, and the last J solutions as suboptimal solutions; K+J is the sum of the schemes in the initial solution sequence. Using the K optimal solutions as centers, the J suboptimal solutions are clustered using equal-value clustering. The ratio of K to J is set based on the Q value, where Q is typically greater than 5, and J is Q times K. Within each solution set, the suboptimal solution with the longest simulated operation time is selected as the worst solution. For example, assuming there are 14 solutions and Q is 6, then the number of K solutions is 2, and the number of J solutions is 12, meaning the first 2 solutions are optimal, and the last 12 are suboptimal.

[0073] Clustering inferior solutions ensures that solutions within each set are similar in simulated job duration or other criteria. Clustering algorithms (such as K-means clustering) are used to group inferior solutions, ensuring similarity within each set. Within each inferior solution set, the differences in job duration or other parameters between solutions are small, allowing them to be optimized together. The inferior solutions are then clustered into K solution sets, the same number as the number of superior solutions.

[0074] In each solution set, the worst solution with the longest simulated construction time is selected as the suboptimal solution, resulting in K suboptimal solutions. Suboptimal solutions are those that require special attention during the optimization process, representing the least desirable solutions. Each solution set contains one suboptimal solution. For example, if the durations of the worst solutions in a solution set are [60, 62, 63, 65], then the solution with the longest duration of 65 hours is the suboptimal solution.

[0075] Based on K optimal solutions and K suboptimal solutions, a strategy of seeking the best and avoiding the worst is adopted, i.e., moving away from suboptimal solutions and moving towards optimal solutions. According to a preset optimization step size, all suboptimal solutions in the K solution sets are adjusted, moving closer to the corresponding optimal solutions in each solution set and moving away from the corresponding suboptimal solutions in each solution set, resulting in K updated solution sets. If the updated suboptimal solution does not meet the aforementioned determined pouring parameter thresholds or vibration parameter thresholds, it is randomly replaced with another construction scheme. That is, it is determined whether the updated construction parameters fall within the set pouring parameter thresholds or vibration parameter thresholds. If the updated values ​​exceed the range, a new construction scheme is randomly selected from multiple qualified construction schemes (i.e., the initial solution sequence) to replace the suboptimal solution, preventing invalid parameters from participating in the optimization. For example, assuming the preset optimization step size is 0.5, if the simulation operation time of the current inferior solution is 65 hours and the simulation operation time of the superior solution is 40 hours, then the adjusted inferior solution time is: 65 - (0.5 * (65 - 40)) = 65 + 12.5 = 52.5 hours (optimization strategy).

[0076] The K updated solution sets are traversed to identify the superior and inferior solutions in each set. Specific parameters (such as work sequence, pouring parameters, and vibration parameters) for each qualified construction scheme are input into the 3D model of the tunnel wall to obtain a simulation result and determine the corresponding simulation operation time. If the simulation operation time of an inferior solution is less than that of a superior solution in the same solution set, it means that the inferior solution performs better than the original superior solution after the current round of optimization, and the inferior solution replaces the superior solution. If the simulation operation time of an inferior solution is greater than that of a poor solution in the same solution set, it means that the inferior solution performs worse after the current round of adjustment, even worse than the original worst solution, and the inferior solution replaces the poor solution.

[0077] Repeat the above process iteratively until the preset number of iterations is met, then stop the optimization and output the K current solutions. The preset number of iterations refers to the maximum number of iterations allowed in an iterative optimization algorithm. It is usually set before the algorithm starts execution and is used to control the termination condition. Different algorithms have different convergence speeds. The preset number of iterations needs to consider the convergence of the algorithm, ensuring that the algorithm can converge to the optimal solution within a finite number of iterations. Determining the preset number of iterations usually requires a trade-off based on the characteristics of the specific problem and the needs of the practical application. If the preset number of iterations is too small, the algorithm may not be able to find the optimal solution; if the preset number of iterations is too large, the algorithm may consume too many computational resources, reducing efficiency. Therefore, a suitable preset number of iterations should be able to improve the utilization of computational resources while ensuring algorithm performance.

[0078] Based on the K current solution sets, an optimal solution set is evaluated. That is, the total simulated operation time of the K current solution sets is calculated, and the solution set with the smallest total simulated operation time is selected as the optimal solution set. From the optimal solution set, the best solution is selected, i.e., the scheme that minimizes the simulated operation time, as the first optimal construction scheme. The first optimal construction scheme is the final best scheme selected through iterative optimization; it is based on the best solutions in the optimal solution set and represents the best scheme that can be implemented in actual construction.

[0079] Multiple work window sequences are analyzed sequentially to determine the optimized construction plan for each sequence, resulting in multiple optimized construction plans. Through these multiple work window sequences, the aforementioned pouring and vibration parameters are optimized to obtain the optimal construction plan. This optimized plan is then transmitted to the needle beam trolley for precise equipment control, including parameter adjustments for pouring and vibration operations and real-time monitoring of construction progress. During construction, the execution status of various parameters is monitored in real time and fed back to the control center. Equipment parameters are adjusted based on the real-time data to ensure the operation remains within the optimal plan. By optimizing pouring and vibration parameters, construction efficiency can be significantly improved while ensuring concrete quality.

[0080] Furthermore, this application also includes the following steps:

[0081] Calculate the average simulation operation time of multiple inferior solutions within each solution set, and calculate the deviation from the simulation operation time of the superior and inferior solutions respectively to determine the superior solution deviation and the inferior solution deviation. If the superior solution deviation is greater than or equal to the inferior solution deviation, the optimization strategy is set as the optimization-oriented strategy, and the inferior solutions within the K solution sets are adjusted once according to the preset optimization step size, with the superior solution as the adjustment direction. If the superior solution deviation is less than the inferior solution deviation, the optimization strategy is set as the inferior solution avoidance strategy, and the inferior solutions within the K solution sets are adjusted once according to the preset optimization step size, with the inferior solution being moved away from the inferior solution as the adjustment direction.

[0082] Specifically, the average simulation operation time of multiple inferior solutions within each solution set is calculated, and the deviation from the simulation operation time corresponding to the optimal solution is calculated as the optimal solution deviation; the deviation from the simulation operation time corresponding to the inferior solution is calculated as the inferior solution deviation. If the optimal solution deviation is greater than or equal to the inferior solution deviation, it indicates that the current inferior solution is closer to the inferior solution and farther from the optimal solution, requiring optimization through an optimization strategy, i.e., adjusting the inferior solution towards the direction of the optimal solution. Based on a predetermined preset optimization step size, an adjustment is performed on the inferior solutions within the K solution sets. That is, the new inferior solution parameters are equal to the original inferior solution parameters plus the product of the preset optimization step size and the optimal solution deviation. An adjustment refers to an update of the current inferior solution parameters according to the current strategy during one iteration.

[0083] Conversely, if the deviation of the optimal solution is less than the deviation of the inferior solution, it indicates that the current inferior solution is closer to the optimal solution. The optimization strategy is then set as an inferior solution avoidance strategy, adjusting away from inferior solutions as the direction of adjustment. Based on a preset optimization step size, the inferior solutions within the K solution sets are adjusted once. Dynamically setting the optimization strategy based on the deviations between inferior and optimal / inferior solutions, and effectively adjusting inferior solutions, helps improve the efficiency and accuracy of the optimization process, leading to a faster finding of the optimal solution. The preset optimization step sizes for the optimization-oriented strategy and the inferior solution avoidance strategy are different, specifically determined according to the corresponding strategy, which will be described in detail in subsequent steps and will not be elaborated here.

[0084] Furthermore, this application also includes the following steps:

[0085] Calculate the simulation operation time deviations between multiple inferior solutions and superior or inferior solutions within each solution set, obtaining multiple superior solution time deviations and multiple inferior solution time deviations. Then, calculate the average superior solution time deviation and the average inferior solution time deviation. If the optimization strategy is an optimization-oriented strategy, set the ratio of the superior solution time deviation to the average superior solution time deviation as an adjustment coefficient, obtaining multiple adjustment coefficients to adjust the initial optimization step size, resulting in multiple adjusted optimization step sizes as preset optimization step sizes. If the optimization strategy is an inferior-avoidance strategy, set the ratio of the inferior solution time deviation to the average inferior solution time deviation as an adjustment coefficient, adjusting the initial optimization step size, resulting in multiple adjusted optimization step sizes as preset optimization step sizes.

[0086] Specifically, each solution set includes one optimal solution, one suboptimal solution, and multiple inferior solutions. The deviations in simulated operation time between the multiple suboptimal solutions and the optimal solution (the construction scheme with the shortest simulated operation time) are calculated for each solution set, including the deviations in simulated operation time between each suboptimal solution and the corresponding construction scheme of the optimal solution, resulting in multiple optimal solution time deviations. Similarly, the deviations in simulated operation time between the multiple suboptimal solutions and the suboptimal solutions (the construction scheme with the longest simulated operation time) are calculated for each solution set, resulting in multiple suboptimal solution time deviations. The average of the multiple optimal solution time deviations is then calculated to obtain the average optimal solution time deviation; the average of the multiple suboptimal solution time deviations is also calculated to obtain the average suboptimal solution time deviation.

[0087] The initial optimization step size is a parameter set in an iterative optimization algorithm that determines the magnitude of solution adjustment in each iteration. Specifically, it is the step size by which the algorithm modifies or adjusts the solution at the beginning of the optimization process. The initial optimization step size is usually determined based on experience or through experimentation and has a significant impact on the efficiency and convergence speed of the optimization process.

[0088] If the optimization strategy is an optimization-oriented strategy, that is, moving closer to a better solution, the current inferior solution is gradually adjusted towards a better solution, thereby improving the overall efficiency of the solution set. Multiple adjustment coefficients are obtained by calculating the ratio of the time deviation of multiple optimal solutions to the average time deviation of optimal solutions. These coefficients are then used to adjust the initial optimization step size, resulting in multiple adjusted optimization step sizes, which serve as preset optimization compensations. The larger the deviation, the larger the corresponding adjustment coefficient, and therefore the larger the preset optimization step size obtained after multiplying by the initial optimization compensation. If the deviation is large, it indicates that the inferior solution is far from the target and requires a larger step size adjustment, i.e., a larger adjustment range; if the deviation is small, finer adjustment is needed, i.e., a smaller adjustment range.

[0089] If the optimization strategy is a suboptimal avoidance strategy, i.e., moving away from suboptimal solutions, the current suboptimal solution is gradually adjusted in a direction away from suboptimal solutions, thereby improving the overall efficiency of the solution set. Similarly, by calculating the ratio of the duration deviation of multiple suboptimal solutions to the average duration deviation of suboptimal solutions, multiple adjustment coefficients are obtained. These coefficients are then used to adjust the initial optimization step size, resulting in multiple adjusted optimization step sizes, which serve as preset optimization compensation.

[0090] The optimization step size is dynamically adjusted based on the deviation between the inferior solution and the superior or inferior solution. A larger deviation results in a larger adjustment range, which helps improve optimization efficiency. Conversely, a smaller deviation results in a smaller adjustment range, which helps improve optimization accuracy, quickly find the optimal solution, and enhance the level of intelligence in construction management.

[0091] In summary, the digital management method for needle beam trolley construction provided in this application has the following beneficial effects:

[0092] A three-dimensional model of the tunnel wall is constructed by scanning the tunnel wall using a laser scanner. The position information of several working windows of the needle beam trolley is obtained, and the working window position distribution is constructed. Constrained by the maximum number of collaborative working windows, the operation sequence is enumerated based on the working window position distribution to generate multiple working window operation sequences. Based on the three-dimensional model of the tunnel wall, and with the constraint of meeting the expected construction quality and the goal of maximizing construction efficiency, the pouring and vibration parameters are optimized according to the multiple working window operation sequences to output the optimal construction plan and achieve digital construction management of the needle beam trolley. In other words, the laser scanner spatially perceives the tunnel wall, generates a three-dimensional model that realistically reflects the tunnel's specific shape and structure, constructs the working window position distribution, and generates multiple working window operation sequences considering the working window position distribution and the maximum number of collaborative working windows. Combined with the three-dimensional model of the tunnel wall, the pouring and vibration parameters for each working window operation sequence are optimized to ensure construction quality while maintaining the efficiency and accuracy of the needle beam trolley construction.

[0093] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0094] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A digital management method for needle beam trolley construction, characterized in that, include: A laser scanner is used to scan the walls of the tunnel under construction to create a three-dimensional model of the tunnel walls. Obtain the position information of several working windows of the needle beam trolley and construct the working window position distribution; With the maximum number of collaborative work windows as a constraint, the work order is enumerated according to the work window position distribution to generate multiple work window work sequences; Based on the three-dimensional model of the tunnel wall, with the constraint of meeting the expected construction quality and the goal of maximizing construction efficiency, the pouring parameters and vibration parameters are optimized according to the multiple work window operation sequences, and the optimal construction scheme is output to digitally manage the needle beam trolley. Based on the three-dimensional model of the tunnel wall, and with the constraint of meeting the expected construction quality and the goal of maximizing construction efficiency, the pouring parameters and vibration parameters are optimized according to the multiple work window operation sequences, including: Using the three-dimensional model of the tunnel wall as a comparison constraint, a sample dataset is retrieved, and a concrete quality predictor is trained and obtained. Randomly select the first work window operation sequence from the plurality of work window operation sequences; Using the concrete quality predictor, with the constraint of meeting the expected construction quality and the goal of maximizing construction efficiency, the pouring parameters and vibration parameters are optimized according to the first work window operation sequence, and the first optimized construction scheme is output. Then, multiple optimized construction schemes of the multiple work window operation sequences are analyzed and obtained in sequence.

2. The digital management method for needle beam trolley construction according to claim 1, characterized in that, A laser scanner is used to scan the walls of the tunnel under construction to create a three-dimensional model of the tunnel walls, including: At multiple preset scanning locations, a laser scanner is used to scan the wall of the tunnel under construction to obtain multiple wall point cloud datasets; After point cloud registration and cleaning of the multiple wall point cloud datasets, mesh modeling is performed to generate a 3D model of the tunnel wall.

3. The digital management method for needle beam trolley construction according to claim 1, characterized in that, Based on the preset construction start point, with the maximum number of collaborative work windows as the construction constraint for the same period, and with adjacent work windows as the sequence constraint, the operation path is enumerated according to the position distribution of the work windows to generate multiple work window operation sequences.

4. The digital management method for needle beam trolley construction according to claim 1, characterized in that, Using the 3D model of the tunnel wall as a comparison constraint, a sample dataset is retrieved, and a concrete quality predictor is trained, including: Using the three-dimensional model of the tunnel wall as a comparison constraint and satisfying a preset similarity threshold as a condition, historical needle beam trolley construction records are retrieved, and a sample dataset is collected. The sample data includes sample work window operation sequence, sample pouring parameters, and sample vibration parameters. Collect concrete quality parameters under different sample work window operation sequences, sample pouring parameters, and sample vibration parameters to obtain a sample quality parameter set; Using the sample dataset and sample quality parameter set, the generator and discriminator of the generative adversarial network are trained until convergence, thus obtaining the concrete quality predictor.

5. The digital management method for needle beam trolley construction according to claim 1, characterized in that, Using the concrete quality predictor, constrained by meeting expected construction quality and aiming to maximize construction efficiency, the pouring and vibration parameters are optimized according to the first work window operation sequence, and a first optimized construction scheme is output, including: Within the threshold values ​​of pouring parameters and vibration parameters, several pouring parameters and several vibration parameters are randomly selected and combined with the first work window operation sequence to obtain several construction schemes. Using the concrete quality predictor, the various construction schemes are evaluated, and several quality parameters are output. Based on the expected construction quality, the various quality parameters are screened to output multiple qualified construction plans; With the goal of maximizing construction efficiency, the pouring and vibration parameters are optimized based on the multiple qualified construction schemes, and the first optimized construction scheme is output.

6. The digital management method for needle beam trolley construction according to claim 5, characterized in that, With the goal of maximizing construction efficiency, the pouring and vibration parameters are optimized based on the multiple qualified construction schemes to output the first optimized construction scheme, including: Based on the multiple qualified construction plans, conduct operation simulations and output multiple simulated operation durations; The qualified construction schemes are regarded as the initial solutions. The qualified construction schemes are arranged in ascending order of simulated operation time to generate an initial solution sequence. The first K solutions of the initial solution sequence are selected as excellent solutions and the last J solutions are selected as inferior solutions. Then, with the K excellent solutions as the center, the J inferior solutions are clustered equally to obtain K solution sets, where J is Q times K and Q is an integer greater than 5. Within each solution set, the inferior solution with the longest simulated construction time is selected as the inferior solution, and K inferior solutions are determined. Based on the K optimal solutions and K inferior solutions, following the optimization and avoidance strategy, the inferior solutions in the K solution sets are adjusted once according to the preset optimization step size, and K updated solution sets are output. If the updated inferior solution does not meet the pouring parameter threshold or the vibration parameter threshold, a construction scheme is randomly selected for replacement. Identify the K updated solution sets. If the simulation operation time of the inferior solution in the same solution set is less than that of the superior solution, then replace the superior solution with the inferior solution. If the simulation operation time of the inferior solution in the same solution set is greater than that of the poor solution, then replace the poor solution with the inferior solution. Perform iterative optimization until the preset number of optimization attempts is met, output K current solution sets, evaluate and determine the optimal solution set, and select the optimal solution set as the first optimized construction scheme, wherein the optimal solution set is the solution set with the smallest total simulated operation time among the K current solution sets.

7. The digital management method for needle beam trolley construction according to claim 6, characterized in that, Based on the K optimal solutions and K inferior solutions, following a strategy of seeking the best and avoiding the worst, the inferior solutions within the K solution set are adjusted once according to a preset optimization step size, including: Calculate the average simulation operation time of multiple inferior solutions in each solution set, and calculate the deviation from the simulation operation time of the superior and inferior solutions respectively, and determine the superior solution deviation and the inferior solution deviation; If the deviation of the optimal solution is greater than or equal to the deviation of the poor solution, the optimization strategy is set as the optimization strategy. Then, the optimal solution is used as the adjustment direction, and the inferior solutions in the K solution sets are adjusted once according to the preset optimization step size. If the deviation of the optimal solution is less than the deviation of the inferior solution, the optimization strategy is set as the inferior solution avoidance strategy. Then, the adjustment direction is to move away from the inferior solution, and the inferior solutions in the K solution sets are adjusted once according to the preset optimization step size.

8. The digital management method for needle beam trolley construction according to claim 7, characterized in that, The method for setting the preset optimization step size includes: Calculate the simulation operation time deviation between multiple inferior solutions and superior or inferior solutions in each solution set to obtain the time deviation of multiple superior solutions and multiple inferior solutions, and calculate the average time deviation of superior solutions and the average time deviation of inferior solutions. If the optimization strategy is an optimization-oriented strategy, the ratio of the deviation of the optimal solution time to the mean of the deviation of the optimal solution time is set as the adjustment coefficient, and multiple adjustment coefficients are obtained. The initial optimization step size is adjusted to obtain multiple adjusted optimization step sizes as the preset optimization step size. If the optimization strategy is a suboptimal strategy, the ratio of the difference solution duration deviation to the mean difference solution duration deviation is set as the adjustment coefficient, and the initial optimization step size is adjusted to obtain multiple adjusted optimization step sizes as preset optimization step sizes.

9. The digital management method for needle beam trolley construction according to claim 1, characterized in that, Among the multiple optimized construction schemes of the multiple work window operation sequences, the scheme with the shortest simulated operation time is selected as the optimal construction scheme.

Images