Intelligent control method and control system for tunneling path of roadheader

Abstract

The present invention discloses an intelligent control method and control system for the tunneling path of a tunnel boring machine, which relates to the technical field of tunnel boring machine control. The method includes: establishing a tunneling space risk distribution model; performing hierarchical encirclement collision detection on the tunneling space risk distribution model to construct a safe tunneling motion space model; performing tunneling path planning on the safe tunneling motion space model to build a first tunneling path planning domain; performing overlimit feedback optimization on the first tunneling path planning domain to establish a second tunneling path planning domain; performing joint optimization to determine the tunneling path optimization result; and performing tunneling path tracking control on the tunnel boring machine. The present invention solves the technical problems in the prior art of insufficient risk prediction of the tunneling area and insufficient consideration of the tunnel boring machine's technical parameters and tunneling objectives in path planning, resulting in poor tunneling safety and low efficiency. It achieves the technical effect of realizing accurate path tracking control of the tunnel boring machine and improving the safety and efficiency of tunneling operations.

Description

Technical Field

[0001] The present invention relates to the technical field of roadheader control, and in particular to an intelligent control method and control system for a roadheader excavation path. Background Art

[0002] In mining operations, the efficiency and safety of roadheaders (TBMs) directly impact mining progress and profitability. The geological conditions in mining areas are complex and highly variable, making risk predictions for the areas where the TBMs will be excavating often inaccurate and incomplete, making it difficult to detect potential hazards in advance. This leads to frequent safety hazards such as rock collapses and gas leaks during excavation. Furthermore, traditional path planning methods often rely on experience or simple algorithms, failing to fully consider the technical limitations of the TBMs themselves and failing to closely align with complex mining objectives. This results in low tunneling efficiency and wasted resources.

[0003] The existing technology has the problem of insufficient risk prediction in the excavation area and the path planning does not fully consider the technical parameters of the tunnel boring machine and the excavation goals, resulting in technical problems such as poor safety and low efficiency of the excavation operation. Summary of the Invention

[0004] The present application provides an intelligent control method and control system for the excavation path of a tunnel boring machine, which is used to solve the technical problems in the prior art of insufficient risk prediction of the excavation area and insufficient consideration of the technical parameters and excavation goals of the tunnel boring machine in path planning, resulting in poor safety and low efficiency of the excavation operation.

[0005] In view of the above problems, the present application provides an intelligent control method and control system for the excavation path of a tunnel boring machine.

[0006] In a first aspect of the present application, a method for intelligently controlling a tunneling path of a tunnel boring machine is provided, the method comprising:

[0007] Perform risk prediction on the area to be excavated by the tunnel boring machine and establish a tunnel boring space risk distribution model; perform hierarchical encirclement collision detection on the tunnel boring space risk distribution model based on the tunnel boring machine and construct a safe tunnel boring motion space model; perform tunnel boring path planning on the safe tunnel boring motion space model according to the excavation target of the area to be excavated and build a first tunnel boring path planning domain; perform over-limit feedback optimization on the first tunnel boring path planning domain according to the technical parameter constraints of the tunnel boring machine and establish a second tunnel boring path planning domain; perform joint optimization on the second tunnel boring path planning domain according to predetermined tunnel boring evaluation factors and determine the tunnel boring path optimization result; and perform tunnel boring path tracking control on the tunnel boring machine according to the tunnel boring path optimization result.

[0008] In a possible implementation, modeling is performed based on the tunnel boring machine to construct a tunnel boring machine model; based on the tunnel boring machine model, simulated collision detection is performed on the tunnel boring space risk distribution model according to a multi-level bounding box to construct a first safe tunnel boring space model, wherein the multi-level bounding box includes multiple box-shaped colliders and multiple grid colliders; point cleaning is performed on the tunnel boring space risk distribution model according to a point risk threshold to construct a second safe tunnel boring space model; the intersection of the first safe tunnel boring space model and the second safe tunnel boring space model is obtained to generate the safe tunnel boring motion space model.

[0009] In a possible implementation, based on the tunnel boring machine model, according to the multi-level bounding box, simulated collision detection is performed on the tunneling space risk distribution model respectively to construct a first simulated collision space model; according to the multi-level bounding box, random combination is performed to obtain a multi-level enclosing structure; based on the tunnel boring machine model, simulated collision detection is performed on the tunneling space risk distribution model according to the multi-level enclosing structure to construct a second simulated collision space model; the first simulated collision space model and the second simulated collision space model are union-identified to generate a simulated collision integrated space model; the difference set between the simulated collision integrated space model and the tunneling space risk distribution model is used as the first safe tunneling space model.

[0010] In a possible implementation, based on the multi-level bounding box, a first bounding box is extracted; based on the first bounding box, the tunnel boring machine model is abstractly bounded to obtain a first bounding tunnel boring machine model; based on the first bounding tunnel boring machine model, multiple simulated collisions are performed on the tunneling space risk distribution model to obtain multiple simulated collision spaces; the multiple simulated collision spaces are fused to obtain a first bounding simulated collision space, and the first bounding simulated collision space is added to the first simulated collision space model.

[0011] In a possible implementation, each excavation path within the first excavation path planning domain is evaluated for overlimit according to the technical parameter constraints to obtain multiple path overlimit coefficients; the first excavation path planning domain is classified according to the multiple path overlimit coefficients to obtain a first cluster of excavation paths with a less than predetermined path overlimit coefficient and a second cluster of excavation paths with a greater than or equal to predetermined path overlimit coefficient; the second cluster of excavation paths is adaptively adjusted according to the technical parameter constraints to obtain a third cluster of excavation paths; the first cluster of excavation paths and the third cluster of excavation paths are merged to generate the second excavation path planning domain.

[0012] In a possible implementation, an excavation evaluation expectation is set according to the predetermined excavation evaluation factor, and the predetermined excavation evaluation factor includes excavation efficiency and excavation safety; the second excavation path planning domain is optimized and analyzed according to the excavation evaluation expectation, and a third excavation path planning domain is established; weights are allocated according to the predetermined excavation evaluation factor, and an excavation optimality evaluation function is established; the third excavation path planning domain is optimized to maximize the excavation optimality according to the excavation optimality evaluation function, and the excavation path optimization result is generated.

[0013] In a possible implementation, based on the second excavation path planning domain, the nth excavation path is extracted, where n is a positive integer; feature prediction is performed based on the nth excavation path to determine the nth excavation prediction result, wherein the nth excavation prediction result includes the nth predicted excavation efficiency coefficient and the nth predicted excavation safety coefficient; it is judged whether the nth excavation prediction result meets the excavation evaluation expectation; if the nth excavation prediction result meets the excavation evaluation expectation, the nth excavation path is added to the third excavation path planning domain; if the nth excavation prediction result does not meet the excavation evaluation expectation, the nth excavation path is eliminated.

[0014] In a possible implementation, the above-ground and underground parameters of the area to be excavated are collected to obtain an excavation space data set; three-dimensional reconstruction is performed based on the excavation space data set to obtain a three-dimensional model of the excavation space; risk prediction is performed on each point in the area to be excavated based on the excavation space data set to obtain a risk prediction coefficient for each point; the risk prediction coefficient for each point is rendered to the three-dimensional model of the excavation space to generate the excavation space risk distribution model.

[0015] In a possible implementation, feature identification is performed on each point based on the excavation space data set to obtain feature data of each point; a point risk prediction network is trained based on a recurrent neural network; and the feature data of each point is input into the point risk prediction network to generate a risk prediction coefficient for each point.

[0016] A second aspect of the present application provides an intelligent control system for a tunneling path of a tunnel boring machine, the system comprising:

[0017] A risk prediction module is used to predict the risks of the area to be excavated by the tunnel boring machine and establish a tunnel boring space risk distribution model; a safe tunnel boring motion space model construction module is used to perform hierarchical encirclement collision detection on the tunnel boring space risk distribution model according to the tunnel boring machine and construct a safe tunnel boring motion space model; a tunnel boring path planning module is used to plan the tunnel boring path for the safe tunnel boring motion space model according to the tunnel boring target of the area to be excavated and build a first tunnel boring path planning domain; an over-limit feedback optimization module is used to perform over-limit feedback optimization on the first tunnel boring path planning domain according to the technical parameter constraints of the tunnel boring machine and establish a second tunnel boring path planning domain; a tunnel boring path optimization result determination module is used to perform joint optimization on the second tunnel boring path planning domain according to a predetermined tunnel boring evaluation factor and determine the tunnel boring path optimization result; a tunnel boring path tracking control module is used to perform tunnel boring path tracking control on the tunnel boring machine according to the tunnel boring path optimization result.

[0018] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0019] The system predicts risks in the area to be excavated by the tunnel boring machine and establishes a tunnel boring space risk distribution model. It then performs hierarchical collision detection on the tunnel boring space risk distribution model to construct a safe tunnel boring motion space model. It then performs tunnel boring path planning on the safe tunnel boring motion space model to establish a first tunnel boring path planning domain. It then performs overrun feedback optimization on the first tunnel boring path planning domain to establish a second tunnel boring path planning domain. It then performs joint optimization on the second tunnel boring path planning domain to determine the tunnel boring path optimization result. Finally, it performs tunnel boring path tracking control on the tunnel boring machine. This system achieves the technical effect of achieving precise path tracking control for the tunnel boring machine and improving the safety and efficiency of tunnel boring operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A schematic flow chart of an intelligent control method for a tunneling path of a tunnel boring machine provided in an embodiment of the present application;

[0022] Figure 2 A schematic diagram of the structure of an intelligent control system for a tunneling path of a tunnel boring machine provided in an embodiment of the present application.

[0023] Explanation of the accompanying symbols: risk prediction module 10, safe excavation movement space model construction module 20, excavation path planning module 30, overrun feedback optimization module 40, excavation path optimization result determination module 50, excavation path tracking control module 60. DETAILED DESCRIPTION

[0024] This application provides an intelligent control method and control system for the excavation path of a tunnel boring machine, which is used to solve the technical problems in the existing technology that the risk prediction of the excavation area is insufficient and the path planning does not fully consider the technical parameters and excavation goals of the tunnel boring machine, resulting in poor safety and low efficiency of the excavation operation.

[0025] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.

[0026] Example 1, as Figure 1 As shown, the present application provides an intelligent control method for a tunneling path of a tunnel boring machine, the method comprising:

[0027] Step S100: Predicting the risk of the area to be excavated by the tunnel boring machine and establishing a tunneling space risk distribution model.

[0028] Specifically, the system first collects above- and below-ground parameters of the area to be excavated, such as geological conditions, rock properties, and the surrounding environment, to form a tunneling space dataset. This data is then used for 3D reconstruction, generating a visual 3D model of the tunneling space. Based on the tunneling space dataset, feature recognition is performed on each point to obtain characteristic data. A point risk prediction network is trained using a recurrent neural network. This data is then input into the network to generate a risk prediction coefficient for each point. Finally, the risk prediction coefficients for each point are rendered onto the 3D tunneling space model, thereby establishing a tunneling space risk distribution model that clearly displays the risk level at each location in the area to be excavated.

[0029] Step S200: performing hierarchical encirclement collision detection on the tunneling space risk distribution model according to the tunneling machine to construct a safe tunneling motion space model.

[0030] Specifically, a TBM model is first constructed based on its actual shape, size, and operating characteristics. Based on this model, multi-level bounding boxes, namely multiple box-shaped colliders and multiple mesh colliders, are used to simulate collision detection against the tunneling space risk distribution model. Each bounding box is then used to abstractly enclose the TBM model, generating the corresponding bounding box model. This bounding box is then used to simulate collisions with the risk distribution model multiple times. The multiple simulated collision spaces are then fused together to construct a first simulated collision space model. Next, the multi-level bounding boxes are randomly combined to form a multi-level enclosing architecture. Collisions are then simulated against the risk distribution model again to construct a second simulated collision space model. These two simulated collision space models are then combined for identification to generate a simulated collision integrated space model. The difference between this model and the tunneling space risk distribution model constitutes the first safe tunneling space model. Furthermore, the tunneling space risk distribution model is cleaned based on point risk thresholds to remove points with excessively high risk, thus constructing a second safe tunneling space model. Finally, the intersection of the first safe excavation space model and the second safe excavation space model is obtained to construct a safe excavation motion space model that can ensure the safe operation of the tunnel boring machine, providing a safe spatial range basis for subsequent excavation path planning.

[0031] Step S300: planning a tunneling path for the safe tunneling motion space model according to the tunneling target of the area to be tunneled, and establishing a first tunneling path planning domain.

[0032] Specifically, the first excavation path planning domain is constructed by combining the excavation objectives of the area to be excavated with the safe excavation motion space model, employing an improved A* algorithm. First, the excavation objectives are clearly defined, such as the designated excavation endpoint coordinates and the desired tunnel shape, among other key information. The safe excavation motion space model is then gridded, dividing the space into small grid cells. Each grid cell contains attribute information such as risk level and traversability. Starting from the starting position of the tunnel boring machine, the A* algorithm is used to search within the grid. The A* algorithm calculates the actual cost g(n) from each grid node to the starting point and the estimated cost h(n) to the target point, i.e., f(n) = g(n) + h(n). The search then sorts the nodes in ascending order of f(n) values. During the search, the risk factors in the safe excavation motion space model are taken into account. The movement costs of grid cells with higher risk levels are appropriately increased, allowing the algorithm to avoid high-risk areas as much as possible. The path is also constrained by technical parameters such as the TBM's turning radius and minimum straight-line excavation distance. If the TBM turns at a certain node and the required turning radius doesn't meet the required radius, or if the continuous straight-line excavation distance is too short, the path branch is discarded. As the search progresses, the algorithm generates a series of paths from the starting point to the target point that meet safety and technical requirements. These paths together constitute the first tunneling path planning domain, providing the foundation for subsequent optimization.

[0033] Step S400: performing overrun feedback optimization on the first tunneling path planning domain according to technical parameter constraints of the tunneling machine to establish a second tunneling path planning domain.

[0034] Specifically, a detailed overrun assessment is performed on each tunneling path within the first tunneling path planning domain based on the technical parameters of the tunnel boring machine (TBM), such as maximum excavation speed, cutting power, equipment size limitations, and maximum climbing angle. The degree to which each parameter deviates from the normal range during execution is calculated for each path, resulting in multiple overrun coefficients. These coefficients intuitively reflect the technical risks and equipment loads each path may face. Based on the calculated overrun coefficients, the paths within the first tunneling path planning domain are classified. Paths with overrun coefficients less than the predetermined overrun coefficient are classified as the first cluster of tunneling paths. These paths generally meet the technical parameter requirements of the TBM and are considered relatively low-risk. Paths with overrun coefficients greater than or equal to the predetermined overrun coefficient are classified as the second cluster of tunneling paths. These paths have some parameter overruns and require adjustment. For this second cluster of tunneling paths, adaptive adjustments are made based on the technical constraints of the TBM. For example, if the excavation speed of a path is too high, causing the cutting power to exceed the limit, the excavation speed is appropriately reduced. If the turning angle of a path is too large, exceeding the equipment's tolerance, the path for the turning portion is replanned. Through adjustment and optimization measures, a third cluster of tunneling paths was obtained, which met the technical parameter requirements. Finally, the first cluster of tunneling paths and the optimized third cluster of tunneling paths were merged to generate the second tunneling path planning domain. The paths in the second tunneling path planning domain not only considered the tunneling target and safety space requirements, but also met the technical parameter constraints of the tunnel boring machine, further improving the accuracy and feasibility of tunneling path planning.

[0035] Step S500: performing joint optimization on the second excavation path planning domain according to a predetermined excavation evaluation factor to determine an excavation path optimization result.

[0036] Specifically, within the second tunneling path planning domain, a joint optimization process is conducted based on predetermined tunneling evaluation factors to determine the final tunneling path optimization result. These predetermined tunneling evaluation factors cover two key aspects: tunneling efficiency and tunneling safety. First, tunneling evaluation expectations corresponding to these factors are set based on actual project requirements, such as the desired minimum tunneling efficiency standard and the maximum acceptable risk level. Each n-th tunneling path (n is a positive integer) is extracted from the second tunneling path planning domain and a comprehensive feature prediction is performed. By analyzing factors such as path length, excavation speed, and the risk level of the traversed area, the n-th predicted tunneling efficiency coefficient and n-th predicted tunneling safety coefficient are determined to quantify the path's tunneling efficiency and safety performance. These predictions are then compared with the tunneling evaluation expectations. If the n-th predicted tunneling result meets the set tunneling evaluation expectations, indicating that the path meets the basic requirements for efficiency and safety, it is added to the third tunneling path planning domain. If it does not, the path is eliminated. After the screening is complete, appropriate weights are assigned to tunneling efficiency and tunneling safety based on their importance in the overall tunneling operation, thereby constructing a tunneling quality evaluation function. This function comprehensively considers these two key factors and quantitatively assesses the overall quality of the path. Finally, an optimization algorithm is used to optimize the paths in the third tunneling path planning domain to maximize their tunneling quality. During this optimization process, path parameters such as tunneling speed and turning methods are continuously adjusted to maximize the path's tunneling quality. The resulting tunneling path is the tunneling path optimization result, which not only meets the project's tunneling objectives but also strikes a balance between tunneling efficiency and safety.

[0037] Step S600: performing tunneling path tracking control on the tunnel boring machine according to the tunneling path optimization result.

[0038] Specifically, the optimized tunneling path is converted into control commands recognizable by the TBM, including parameters such as tunneling direction, speed, and cutting force. During TBM operation, various sensors mounted on the machine, such as laser sensors, inertial navigation sensors, and displacement sensors, acquire real-time operating status information such as the TBM's position, posture, and speed. These actual operating conditions are compared and analyzed with the theoretical parameters of the optimized path. If any deviation is detected, the magnitude and direction of the deviation are calculated. Appropriate adjustment strategies are implemented for each type of deviation. If the tunneling direction deviates, the TBM's steering mechanism is adjusted to gradually return it to the optimized path. If the tunneling speed deviates from the planned value, the drive system power is adjusted to achieve an appropriate speed. Furthermore, considering that geological conditions and environmental factors may change during tunneling, the control system continuously monitors these dynamic changes and optimizes control commands accordingly. Furthermore, to ensure the stability and reliability of tunneling path tracking control, a feedback control mechanism is introduced to monitor and evaluate the adjusted operating status, forming a closed-loop control process. If the deviation cannot be effectively eliminated after multiple adjustments, an automatic alarm will be issued, prompting the operator to intervene manually. This tunneling path tracking control ensures that the tunnel boring machine completes the tunneling operation efficiently and stably along the optimal path, reducing construction risks and improving project quality.

[0039] In one possible implementation, step S100 further includes:

[0040] Step S110: Collecting above-ground and underground parameters of the area to be excavated to obtain an excavation space dataset.

[0041] Step S120: Perform three-dimensional reconstruction based on the tunneling space dataset to obtain a three-dimensional model of the tunneling space.

[0042] Step S130: performing risk prediction on each point in the area to be excavated based on the excavation space data set to obtain a risk prediction coefficient for each point.

[0043] Step S140: Rendering the risk prediction coefficient of each point to the three-dimensional model of the tunneling space to generate the tunneling space risk distribution model.

[0044] Specifically, in the early stages of tunneling, comprehensive data collection of above- and below-ground parameters of the area to be excavated is required. For above-ground parameters, satellite remote sensing and drone mapping can be used to obtain information on topography, building distribution, and vegetation cover. For underground parameters, geological drilling and geophysical exploration are used to collect data on geological structure, rock properties, groundwater levels, and soil mechanical properties. By integrating this rich data set, a tunneling spatial dataset is generated.

[0045] 3D reconstruction is performed based on the obtained tunneling space dataset. Using professional 3D modeling software such as 3ds Max and Maya, a 3D model of the area to be excavated is constructed based on the terrain, landforms, and underground structure information contained in the dataset. During this process, the software converts the data into intuitive 3D graphics, accurately representing the actual spatial form of the area to be excavated, ultimately resulting in a 3D model of the tunneling space.

[0046] A recurrent neural network (RNN) is used to predict risk at each point in the tunneling area based on a tunneling spatial dataset, thereby obtaining a risk prediction coefficient for each point. First, the tunneling spatial dataset is preprocessed. This dataset contains various above- and below-ground parameters, such as geological structure, rock properties, groundwater levels, and the surrounding environment. The data is cleaned to remove noise and missing values, and then normalized to ensure that the data is within a suitable numerical range. Next, the processed data is organized according to the spatial order of the points and their temporal sequence (if time-correlated data exists), creating a format suitable for RNN input. A RNN model is constructed, selecting a classic RNN architecture, determining hyperparameters such as the number of layers and neurons, and initializing the model. The RNN model is trained using the preprocessed dataset. During training, the model learns features and patterns in the data, as well as the potential relationships between various parameters and risk. A backpropagation algorithm continuously adjusts the model's weights and biases to minimize the error between the predicted and actual risk values. After training is complete, data from each point in the excavation area is fed into the trained recurrent neural network model for prediction. Based on the learned patterns, the model outputs a risk prediction result for each point. These results are post-processed and normalized to convert them into a risk prediction coefficient between 0 and 1, with a higher coefficient indicating a higher risk at the point. Ultimately, the risk prediction coefficient for each point in the excavation area is obtained.

[0047] The risk prediction coefficient for each point is rendered onto the 3D model of the tunneling space. Using visualization technology, each point in the model is assigned a different color or texture based on the risk prediction coefficient. For example, high-risk points are represented in red, and low-risk points in green. This allows the abstract risk data to be intuitively displayed within the 3D model, generating a risk distribution model for the tunneling space.

[0048] In one possible implementation, step S130 further includes:

[0049] Step S131: performing feature recognition on each of the points according to the tunneling space data set to obtain feature data of each point.

[0050] Step S132: Based on the recurrent neural network, train the point risk prediction network.

[0051] Step S133: inputting the feature data of each point into the point risk prediction network to generate the risk prediction coefficient of each point.

[0052] Specifically, the tunneling spatial dataset contains a rich and diverse range of information, including geological, topographic, and groundwater level data. To obtain characteristic data for each location, we first apply geostatistical methods to the geological data, analyzing the variations in stratum lithology at each location and identifying the specific manifestations of different rock types, fault orientations, and fold characteristics at each location. For example, by analyzing the mineral composition of rock samples, we determine characteristic data such as rock hardness and compressive strength at each location. For topographic data, we use digital elevation models (DEMs) combined with geographic information system (GIS) technology to extract characteristics such as slope, aspect, and topographic relief at each location, reflecting the complexity of the terrain at that location. For groundwater level data processing, we use hydrogeological simulation software to analyze the dynamic changes in groundwater levels at different locations, obtaining characteristic data such as water level depth and amplitude. By combining the results of these different types of data, we compile comprehensive and targeted characteristic data for each location.

[0053] To train the point risk prediction network based on a recurrent neural network, the previously acquired tunneling spatial dataset is preprocessed. This includes data cleaning to remove outliers and noise, followed by normalization to bring the data of different dimensions into a uniform numerical range, facilitating model training. Subsequently, a recurrent neural network model is constructed, and after determining key hyperparameters such as the number of layers and neurons, the model is trained using the preprocessed dataset. During training, the network's weights and biases are continuously adjusted, and the backpropagation algorithm is used to minimize the error between the predicted and actual risk values. This gradually improves the model's predictive accuracy until training converges, completing the training of the point risk prediction network.

[0054] The resulting feature data for each point is fed into a trained point risk prediction network. The network analyzes and calculates the input feature data based on the mapping relationship between features and risks learned during training, ultimately outputting a risk prediction result for each point. These results undergo appropriate post-processing and normalization to generate a risk prediction coefficient for each point, providing an accurate quantitative basis for subsequent risk assessment and tunneling path planning.

[0055] In one possible implementation, step S200 further includes:

[0056] Step S210: Modeling the roadheader to construct a roadheader model.

[0057] Step S220: Based on the tunnel boring machine model, simulate collision detection is performed on the tunnel boring space risk distribution model according to a multi-level bounding box to construct a first safe tunnel boring space model, wherein the multi-level bounding box includes multiple box-shaped colliders and multiple grid colliders.

[0058] Step S230: performing point cleaning on the excavation space risk distribution model according to the point risk threshold, and constructing a second safe excavation space model.

[0059] Step S240: Obtain the intersection of the first safe excavation space model and the second safe excavation space model to generate the safe excavation motion space model.

[0060] Specifically, modeling software was used to accurately construct a model of the TBM based on its actual size, shape, structure, and the range of motion of its components. This model accurately reflects the TBM's characteristics. This model forms the basis for subsequent simulated collision detection, and its accuracy directly impacts the effectiveness of creating a safe space.

[0061] Based on the TBM model's contours, structural characteristics, and the locations of key components, multiple box colliders and mesh colliders were arranged to form a multi-level bounding box. For the TBM's more regular main body, box colliders were used to provide a preliminary, large-scale bounding box to quickly define the approximate area of ​​potential collision. For complex and flexible components like the cutting head and boom, mesh colliders were used for precise alignment to capture more subtle collision scenarios. The TBM model was then placed in the three-dimensional space represented by the tunneling space risk distribution model, simulating various possible motion postures and paths. Using collision detection algorithms such as the Separating Axis Theorem (SAT), real-time collision detection was performed between the multi-level bounding box and high-risk areas in the tunneling space risk distribution model. At each simulated motion step, the box colliders and mesh colliders were checked for intersection with the risk area. If an intersection was detected, the location was marked as a potential danger zone. After collision detection for all simulated motions, the entire three-dimensional space was traversed to filter out any portions of the space not marked as danger zones. Next, spatial topology analysis methods were used to remove isolated, narrow, and non-traversable spatial fragments, and connectivity analysis and optimization were performed on the selected safe spaces. Finally, the optimized safe spaces were integrated to construct the first safe tunneling space model, which clearly defined the spatial range within which the tunnel boring machine could safely operate while avoiding collisions with high-risk areas.

[0062] Clearly define the point risk thresholds that have been pre-determined based on multiple factors, including geological conditions, tunnel boring machine performance, and construction safety standards. Each point in the established tunneling space risk distribution model is evaluated to determine its corresponding risk coefficient. These risk coefficients are then compared against the point risk thresholds. When the risk coefficient of a point is greater than or equal to the point risk threshold, the point is determined to be high-risk. Subsequently, all points identified as high-risk are precisely removed from the tunneling space risk distribution model. After removing the high-risk points, the remaining low-risk points are re-integrated and reconstructed based on their spatial scope, ultimately forming a second safe tunneling space model.

[0063] Convert the data of the two models into a unified spatial coordinate system to ensure the consistency of the position information. Select an appropriate spatial analysis algorithm, such as the Boolean intersection algorithm, to traverse the grid cells of the two models, find the grid cells that exist in both models, and mark these cells as the intersection. During the calculation process, the boundary parts must be accurately processed to avoid gaps or overlapping errors. For the obtained intersection data, perform topological inspection and optimization to remove isolated small areas and unreasonable tiny spaces to ensure the connectivity and rationality of the intersection space. Finally, based on the optimized intersection data, use 3D modeling software to reconstruct the safe excavation motion space model, providing a reliable spatial basis for subsequent excavation path planning and control.

[0064] In one possible implementation, step S220 further includes:

[0065] Step S221: Based on the tunnel boring machine model and according to the multi-level bounding box, simulated collision detection is performed on the tunneling space risk distribution model to construct a first simulated collision space model.

[0066] Step S222: randomly combining the multi-level bounding boxes to obtain a multi-level bounding structure.

[0067] Step S223: Based on the tunnel boring machine model, simulated collision detection is performed on the tunneling space risk distribution model according to the multi-level encirclement architecture to construct a second simulated collision space model.

[0068] Step S224: performing union identification on the first simulated collision space model and the second simulated collision space model to generate a simulated collision integrated space model.

[0069] Step S225: taking the difference between the simulated collision integrated space model and the tunneling space risk distribution model as the first safe tunneling space model.

[0070] Specifically, based on the established TBM model, a multi-level bounding box consisting of multiple box colliders and multiple mesh colliders is strategically placed around the TBM model, ensuring that the multi-level bounding box fully and accurately represents the possible spatial area occupied by the TBM. Subsequently, a simulated collision detection process is performed, simulating various possible postures and positions of the TBM within the tunneling space. In each simulated state, the multi-level bounding box is checked for overlap with the risk area in the tunneling space risk distribution model. Any overlap indicates a potential collision between the TBM and the risk area in that posture and position. After completing the test for all simulated states, the spatial portions where the multi-level bounding box does not overlap with the risk area are marked. Finally, these marked spatial portions are integrated and processed to remove any isolated patches and unreasonable gaps, thereby constructing the first simulated collision space model.

[0071] Given a multi-level bounding box composed of multiple box colliders and mesh colliders, some of these colliders are randomly selected and combined. Because each box collider and mesh collider has a unique size, shape, and preset position in space, their combinations are extremely diverse. For example, several box colliders of different sizes are selected from multiple boxes and paired with a specific mesh collider. During the selection process, a random algorithm is used to determine the type and number of colliders involved in each combination. After multiple rounds of random selection and combination, a series of different multi-level bounding architectures are obtained. Due to the random combination of colliders, these architectures exhibit different bounding shapes and coverage, capable of simulating a variety of different roadheader collision detection scenarios.

[0072] Based on the constructed TBM model, a multi-level bounding structure, derived from the random combination of multiple bounding boxes, is applied to simulated collision detection within the tunneling space risk distribution model, thereby constructing a second simulated collision space model. During the simulation, the simulated TBM moves in various possible postures and positions within the tunneling space. In each simulation state, the multi-level bounding structure is checked for overlap with the risk area in the tunneling space risk distribution model. If overlap is detected, it indicates that the TBM is at risk of colliding with the risk area in that specific state. After completing the test for all simulation states, the spatial portions where the multi-level bounding structure does not overlap with the risk area are marked. These marked spatial regions are then integrated and optimized to remove isolated, small, and insignificant spatial fragments, ultimately constructing the second simulated collision space model.

[0073] Convert the data of the two models into unified three-dimensional grid data, and assign a unique identifier and spatial coordinate information to each grid cell. Next, create an empty set to store the union results. For each grid cell in the first simulated collision space model, check whether it already exists in the result set. If not, add it. After that, perform the same operation on the second simulated collision space model, traverse each grid cell therein, and add it if the cell is not in the result set. During the addition process, special attention should be paid to boundary cases, such as overlapping parts between grid cells. For overlapping grid cells, only one copy is retained to avoid repeated calculations. Finally, based on the set storing the union results, the spatial model is reconstructed to form a simulated collision integrated spatial model.

[0074] The model is represented in the form of a three-dimensional grid, and the grids of the two models are first uniformly indexed. For each grid cell in the simulated collision integrated space model, its spatial coordinates are obtained. Using a spatial index structure, such as an octree, a quick search is performed in the excavation space risk distribution model to determine whether there are risk grid cells that coincide with their spatial coordinates. If no overlapping risk grid cells are found, the grid cells of the simulated collision integrated space model are marked as belonging to the difference set, and their relevant attribute information is recorded. If an overlapping risk grid cell is found, the cell is discarded. After traversing all the grid cells of the simulated collision integrated space model, all the grid cells marked as belonging to the difference set are recombined to construct the first safe excavation space model.

[0075] In one possible implementation, step S221 further includes:

[0076] Step S2211: extracting a first bounding box according to the multi-level bounding boxes.

[0077] Step S2212: abstractly enclose the roadheader model according to the first bounding box to obtain a first enclosed roadheader model.

[0078] Step S2213: performing multiple simulated collisions on the tunneling space risk distribution model according to the first surrounding tunnel boring machine model to obtain multiple simulated collision spaces.

[0079] Step S2214: merging the multiple simulated collision spaces to obtain a first enclosed simulated collision space, and adding the first enclosed simulated collision space to the first simulated collision space model.

[0080] Specifically, the first bounding box is extracted from a multi-level bounding box composed of multiple box colliders and mesh colliders. This process is based on pre-defined extraction rules. For example, given the external characteristics of the TBM, a focus is placed on large, regularly shaped box colliders that cover the TBM's primary components. The fit of different bounding boxes with key TBM components is also considered, prioritizing those that closely align with the TBM's collision-prone areas. Furthermore, based on an analysis of the risk distribution model for the tunneling space, bounding boxes with good coverage of common high-risk areas are selected. By evaluating each multi-level bounding box individually, the one that best meets the requirements is selected from the numerous colliders and designated as the first bounding box.

[0081] Determine the position and posture of the first bounding box in space so that it can cover the main structure and key parts of the tunnel boring machine model to the greatest extent possible, such as the tunnel boring machine's fuselage, cutting head, and major moving parts. During the encirclement process, ignore the complex detailed features of the tunnel boring machine model and regard it as a whole contained by the first bounding box. Through the spatial transformation algorithm, unify the coordinate system of the first bounding box with the coordinate system of the tunnel boring machine model to ensure accurate correspondence between the two in spatial position. Then, combine the contour information of the first bounding box with the tunnel boring machine model to generate the first encircled tunnel boring machine model. Although this model discards some of the tunnel boring machine's details, it retains its key spatial occupancy information required for simulated collision detection. In the subsequent simulated collision link, it can efficiently and accurately represent the tunnel boring machine and the tunneling space risk distribution model to interact, providing a more concise and effective data basis for simulated collision detection.

[0082] Based on the first encircling TBM model, multiple collision simulations were performed against all points in the tunneling space risk distribution model, generating multiple simulated collision spaces. Using simulation software such as ANSYS and ADAMS, the carefully constructed first encircling TBM model and the complete tunneling space risk distribution model were imported into the simulation environment to ensure that their spatial coordinate systems were aligned. Through programming or the software's built-in parameter setting functions, multiple sets of different motion parameters were generated, including the first encircling TBM model's starting position, direction, speed, rotation angle, and posture changes within the tunneling space. A loop instruction was used to initiate multiple simulations. Each time, the first encircling TBM model's initial state was set based on the current parameters, causing it to move along a pre-set trajectory within the virtual tunneling space. During this motion, collision detection was performed in real time between the first encircling TBM model and every point in the tunneling space risk distribution model (regardless of risk level) using a collision detection method based on spatial gridding. Once a collision was detected, the collision information was immediately recorded. If no collision was detected, the spatial area currently occupied by the first encircling TBM model was marked as a safe zone. After each simulation, the safety zone data obtained from that simulation is collated and stored to form a simulated collision space. This process is repeated, and as multiple sets of motion parameters are applied in sequence, multiple simulations are completed, accumulating multiple different simulated collision spaces.

[0083] Once multiple simulated collision spaces have been obtained, the union operation in Boolean operations is used to process the multiple simulated collision spaces. The algorithm traverses the boundaries and internal areas of each simulated collision space, identifies overlapping and adjacent parts, and merges these parts into a continuous, complete space. During the fusion process, the boundaries are precisely processed to eliminate gaps and overlaps caused by simulation errors, ensuring the continuity and accuracy of the space. At the same time, for any isolated small areas that may exist, if their area is too small and has little impact on the overall space, they are discarded to simplify the spatial structure. After the fusion operation, the resulting first enclosing simulated collision space comprehensively and accurately reflects the safe movement range of the tunnel boring machine based on the first bounding box when avoiding collisions with risk areas. Finally, this first enclosing simulated collision space is added to the first simulated collision space model being constructed.

[0084] In one possible implementation, step S400 further includes:

[0085] Step S410: performing an over-limit evaluation on each excavation path within the first excavation path planning domain according to the technical parameter constraints to obtain a plurality of path over-limit coefficients.

[0086] Step S420: Classify the first excavation path planning domain according to the multiple path excess coefficients to obtain a first cluster of excavation paths with a smaller path excess coefficient than a predetermined path excess coefficient and a second cluster of excavation paths with a greater than or equal to predetermined path excess coefficient.

[0087] Step S430: Adaptively adjust the second cluster of tunneling paths according to the technical parameter constraints to obtain a third cluster of tunneling paths.

[0088] Step S440: merging the first cluster of tunneling paths and the third cluster of tunneling paths to generate the second tunneling path planning domain.

[0089] Specifically, technical parameter constraints are categorized into different types, such as equipment performance parameters, geological condition parameters, and safety standard parameters. Weights are assigned to each parameter to reflect its relative importance in the overall evaluation. For each excavation path, a series of sampling points are selected at equal intervals along the path. At each sampling point, actual technical parameter values, such as excavation speed, slope, and distance to surrounding obstacles, are collected. For each technical parameter, the deviation ratio between the actual value and the constraint value is calculated: (actual value - constraint value) / constraint value (when the actual value exceeds the constraint value) or (constraint value - actual value) / constraint value (when the actual value is less than the constraint value). Each parameter's deviation ratio is then multiplied by its corresponding weight, and the resulting products are summed to determine the local exceedance coefficient for that sampling point. The local exceedance coefficients for all sampling points along the path are then weighted averaged (weights can be assigned based on the spacing between sampling points or the importance of their respective regions) to determine the path exceedance coefficient for the excavation path. The algorithm is executed sequentially on all excavation paths within the first excavation path planning domain, thereby obtaining a plurality of path excess coefficients.

[0090] The first tunneling path planning domain is classified based on the multiple path violation coefficients obtained. A predetermined path violation coefficient is set as a classification threshold. Tunneling paths with path violation coefficients less than the threshold are grouped together to form the first cluster of tunneling paths. This cluster of paths is more consistent with technical parameter constraints and has a lower degree of violation. Tunneling paths with path violation coefficients greater than or equal to the predetermined path violation coefficient are grouped together to form the second cluster of tunneling paths. This cluster of paths has more significant violation.

[0091] Adaptive adjustments were performed on the second cluster of tunneling paths. Using an optimization algorithm, each path in the second cluster was adjusted based on technical parameter constraints and actual project requirements. For example, parameters such as path direction, slope, and curvature were modified to achieve optimal results while still meeting technical parameter constraints. This adaptive adjustment resulted in the third cluster of tunneling paths, which had improved overruns and better met technical requirements.

[0092] The first and third clusters of tunneling paths are merged, and all the tunneling paths contained in these clusters are integrated to form a new set of paths. This set is the second tunneling path planning domain. This domain combines the original paths that meet the requirements with the paths that have been adjusted to meet the technical parameter constraints, providing a more reasonable and optimized range of choices for subsequent determination of the optimal tunneling path.

[0093] In one possible implementation, step S500 further includes:

[0094] Step S510: setting excavation evaluation expectations according to the predetermined excavation evaluation factors, wherein the predetermined excavation evaluation factors include excavation efficiency and excavation safety.

[0095] Step S520: performing optimization analysis on the second excavation path planning domain according to the excavation evaluation expectation to establish a third excavation path planning domain.

[0096] Step S530: performing weight allocation according to the predetermined excavation evaluation factors to establish an excavation quality evaluation function.

[0097] Step S540: Optimizing the excavation optimality of the third excavation path planning domain according to the excavation optimality evaluation function to generate the excavation path optimization result.

[0098] Specifically, tunneling evaluation expectations should be set based on the core impact of two key tunneling evaluation factors: tunneling efficiency and tunneling safety. For tunneling efficiency, a target tunneling rate or amount of tunneling per unit time should be set based on project schedule requirements, equipment performance, and previous construction experience. For tunneling safety, specific safety expectation indicators, such as maximum allowable risk probability and minimum safe distance, should be established, taking into account multiple factors, including geological conditions, equipment operational stability, and personnel safety.

[0099] The nth excavation path is extracted from the second excavation path planning domain in sequential order, where n is a positive integer. This ensures that each path is individually evaluated. Next, for each extracted nth excavation path, a prediction model is used to perform feature prediction, leveraging historical data, geological conditions, equipment performance, and other information. This results in an nth excavation prediction result. This result includes an nth predicted excavation efficiency coefficient, reflecting factors such as the expected excavation speed and workload, as well as an nth predicted excavation safety coefficient, which takes into account factors such as the geological stability of the path's surroundings and its proximity to hazardous areas. The nth predicted excavation result is then compared with pre-defined excavation evaluation expectations to determine whether it meets the desired excavation efficiency and safety standards. If the excavation evaluation expectations are met, the path performs well in terms of efficiency and safety and is added to the third excavation path planning domain being constructed. If not, the path presents efficiency or safety risks and is eliminated. By repeating this process for all paths in the second excavation path planning domain, a third excavation path planning domain consisting of paths that meet the excavation evaluation expectations is ultimately formed.

[0100] According to the predetermined excavation evaluation factors, weights are assigned and then the excavation quality evaluation function is established. The predetermined excavation evaluation factors are mainly excavation efficiency and excavation safety. Considering the differences in the importance of these two factors in different engineering scenarios, the hierarchical analysis method is used to determine their weights. Assume that through reasonable evaluation, the weight of excavation efficiency is determined to be W1, the weight of excavation safety is determined to be W2, and W1+W2=1 is satisfied. For excavation efficiency, it is measured by calculating the ratio of the expected completion time t of the path to the expected excavation time t0, which is recorded as E= , the larger the value, the higher the excavation efficiency. For excavation safety, taking into account the geological risks around the path, the distance to obstacles and other factors, a safety factor S can be obtained by establishing a risk assessment model. Its value range is between 0 and 1. The larger the value, the higher the safety. Based on the above analysis, an excavation suitability evaluation function is established. The excavation suitability evaluation function is: F=W1×E+W2×S, where F represents the excavation suitability, W1 represents the weight of excavation efficiency, W2 represents the weight of excavation safety, W1 and W2 together constitute the importance distribution of the overall evaluation, E represents excavation efficiency, which is calculated as the ratio of the expected completion time t of the path to the expected excavation time t0, is the expected excavation time, t is the estimated completion time, and S represents the excavation safety factor, with a value range of 0 to 1. This function comprehensively considers the two key factors of excavation efficiency and excavation safety and their corresponding weights, and can quantify the optimality of each path.

[0101] Using the established excavation quality evaluation function, the third excavation path planning domain is optimized to maximize excavation quality. During this process, an optimization algorithm is applied to calculate the excavation quality value for each path in the third excavation path planning domain. The algorithm continuously adjusts path parameters (such as path direction and slope) to find the path that maximizes the excavation quality evaluation function. After multiple rounds of iteration and optimization, the excavation path optimization result is finally generated. This result represents the most ideal excavation path solution that comprehensively considers excavation efficiency and excavation safety while maximizing quality.

[0102] In one possible implementation, step S520 further includes:

[0103] Step S521: extracting the nth excavation path according to the second excavation path planning domain, where n is a positive integer.

[0104] Step S522: Perform feature prediction based on the nth excavation path to determine the nth excavation prediction result, where the nth excavation prediction result includes the nth predicted excavation efficiency coefficient and the nth predicted excavation safety coefficient.

[0105] Step S523: Determine whether the nth excavation prediction result meets the excavation evaluation expectation.

[0106] Step S524: If the nth excavation prediction result meets the excavation evaluation expectation, the nth excavation path is added to the third excavation path planning domain.

[0107] Step S525: If the nth excavation prediction result does not meet the excavation evaluation expectation, the nth excavation path is eliminated.

[0108] Specifically, the nth excavation path is extracted from the second excavation path planning domain, where n is a positive integer. To evaluate and screen these paths individually and determine the optimal path plan, the paths are processed in numerical order. First, a numbering rule is defined: each path in the second excavation path planning domain is assigned a unique positive integer number, starting from 1 and increasing in sequence. Then, based on the current number n to be processed, an indexing mechanism is used to accurately extract the corresponding nth excavation path from the path set.

[0109] A random forest algorithm is used to predict features based on the nth tunneling path and determine the predicted result. A large amount of historical tunneling data is collected. This data contains numerous path characteristics, such as path length, slope, geological type, and surrounding environmental conditions. The corresponding actual tunneling efficiency and safety indicators are also recorded. Tunneling efficiency is measured by tunneling length per unit time; safety is comprehensively assessed based on factors such as accident probability and risk level. The collected data is divided into training and test sets, and a random forest model is constructed using the training set data. A random forest model consists of multiple decision trees. When constructing each decision tree, a portion of samples and features are randomly sampled from the training data, which increases the model's diversity and generalization ability. To predict the nth predicted tunneling efficiency coefficient, the features of the nth tunneling path are input into the trained random forest model. Each decision tree within the model generates a predicted tunneling efficiency value based on its own rules. The predictions from all decision trees are averaged to obtain the estimated tunneling efficiency for the nth tunneling path. This value is then compared with the historical optimal tunneling efficiency and normalized to obtain the nth predicted tunneling efficiency coefficient. Similarly, when predicting the safety factor for the nth predicted excavation, the characteristics of the nth excavation path are also input into the model. Each decision tree outputs a safety prediction result, which is then averaged to obtain the estimated safety index. Then, through normalization and other operations, the nth predicted excavation safety factor is obtained. Finally, these two factors together constitute the prediction result for the nth excavation.

[0110] Compare the nth predicted excavation result with pre-set excavation evaluation expectations, which cover the expected excavation efficiency and excavation safety standards. Determine whether the nth predicted excavation efficiency coefficient meets the expected efficiency level and whether the nth predicted excavation safety coefficient meets the expected safety requirements.

[0111] If the nth excavation prediction result meets the excavation evaluation expectations, the nth excavation path is added to the third excavation path planning domain being constructed. This means that the path meets expectations in terms of excavation efficiency and safety and is a relatively high-quality choice.

[0112] Conversely, if the nth excavation prediction result fails to meet the excavation evaluation expectations, the nth excavation path is eliminated because it fails to meet the established standards and may pose a risk to efficiency or safety in subsequent excavation work. By repeating the above process for each path in the second excavation path planning domain, a third excavation path planning domain consisting of paths that meet the excavation evaluation expectations is eventually formed, providing a more valuable set of candidate paths for subsequent path optimization.

[0113] The second embodiment is based on the same inventive concept as the intelligent control method for the tunneling path of the tunnel boring machine in the above embodiment. Figure 2As shown, the present application provides an intelligent control system for the excavation path of a roadheader. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0114] The risk prediction module 10 is used to predict the risk of the area to be excavated by the tunnel boring machine and establish a tunneling space risk distribution model.

[0115] The safe excavation motion space model construction module 20 is used to perform hierarchical encirclement collision detection on the excavation space risk distribution model according to the tunnel boring machine, and construct a safe excavation motion space model.

[0116] The excavation path planning module 30 is used to plan the excavation path for the safe excavation motion space model according to the excavation target of the area to be excavated, and to build a first excavation path planning domain.

[0117] The overrun feedback optimization module 40 is configured to perform overrun feedback optimization on the first tunneling path planning domain according to the technical parameter constraints of the tunneling machine to establish a second tunneling path planning domain.

[0118] The excavation path optimization result determination module 50 is used to perform joint optimization on the second excavation path planning domain according to a predetermined excavation evaluation factor to determine the excavation path optimization result.

[0119] The tunneling path tracking control module 60 is used to perform tunneling path tracking control on the tunneling machine according to the tunneling path optimization result.

[0120] Furthermore, the system is also used to implement the following functions:

[0121] Modeling is performed based on the tunnel boring machine to construct a tunnel boring machine model; based on the tunnel boring machine model, simulated collision detection is performed on the tunnel boring space risk distribution model according to a multi-level bounding box to construct a first safe tunnel boring space model, wherein the multi-level bounding box includes multiple box-shaped colliders and multiple grid colliders; point cleaning is performed on the tunnel boring space risk distribution model according to a point risk threshold to construct a second safe tunnel boring space model; the intersection of the first safe tunnel boring space model and the second safe tunnel boring space model is obtained to generate the safe tunnel boring motion space model.

[0122] Furthermore, the system is also used to implement the following functions:

[0123] Based on the tunnel boring machine model, according to the multi-level bounding boxes, simulated collision detection is performed on the tunneling space risk distribution model respectively to construct a first simulated collision space model; according to the multi-level bounding boxes, random combinations are performed to obtain a multi-level enclosing structure; based on the tunnel boring machine model, simulated collision detection is performed on the tunneling space risk distribution model according to the multi-level enclosing structure to construct a second simulated collision space model; the first simulated collision space model and the second simulated collision space model are union-identified to generate a simulated collision integrated space model; the difference set between the simulated collision integrated space model and the tunneling space risk distribution model is used as the first safe tunneling space model.

[0124] Furthermore, the system is also used to implement the following functions:

[0125] According to the multi-level bounding box, a first bounding box is extracted; according to the first bounding box, the tunnel boring machine model is abstractly bounded to obtain a first bounding tunnel boring machine model; according to the first bounding tunnel boring machine model, multiple simulated collisions are performed on the tunneling space risk distribution model to obtain multiple simulated collision spaces; the multiple simulated collision spaces are fused to obtain a first bounding simulated collision space, and the first bounding simulated collision space is added to the first simulated collision space model.

[0126] Furthermore, the system is also used to implement the following functions:

[0127] According to the technical parameter constraints, each excavation path in the first excavation path planning domain is evaluated for overlimit, and multiple path overlimit coefficients are obtained; the first excavation path planning domain is classified according to the multiple path overlimit coefficients to obtain a first cluster of excavation paths with less than a predetermined path overlimit coefficient, and a second cluster of excavation paths with greater than or equal to the predetermined path overlimit coefficient; according to the technical parameter constraints, the second cluster of excavation paths is adaptively adjusted to obtain a third cluster of excavation paths; the first cluster of excavation paths and the third cluster of excavation paths are merged to generate the second excavation path planning domain.

[0128] Furthermore, the system is also used to implement the following functions:

[0129] According to the predetermined excavation evaluation factors, excavation evaluation expectations are set, and the predetermined excavation evaluation factors include excavation efficiency and excavation safety; according to the excavation evaluation expectations, the second excavation path planning domain is optimized and analyzed to establish a third excavation path planning domain; according to the predetermined excavation evaluation factors, weights are allocated to establish an excavation optimality evaluation function; according to the excavation optimality evaluation function, the third excavation path planning domain is optimized to maximize the excavation optimality to generate the excavation path optimization result.

[0130] Furthermore, the system is also used to implement the following functions:

[0131] According to the second excavation path planning domain, the nth excavation path is extracted, where n is a positive integer; feature prediction is performed based on the nth excavation path to determine the nth excavation prediction result, wherein the nth excavation prediction result includes the nth predicted excavation efficiency coefficient and the nth predicted excavation safety coefficient; it is judged whether the nth excavation prediction result meets the excavation evaluation expectation; if the nth excavation prediction result meets the excavation evaluation expectation, the nth excavation path is added to the third excavation path planning domain; if the nth excavation prediction result does not meet the excavation evaluation expectation, the nth excavation path is eliminated.

[0132] Furthermore, the system is also used to implement the following functions:

[0133] Collect above-ground and underground parameters of the area to be excavated to obtain an excavation space data set; perform three-dimensional reconstruction based on the excavation space data set to obtain a three-dimensional model of the excavation space; perform risk prediction on each point in the area to be excavated based on the excavation space data set to obtain a risk prediction coefficient for each point; render the risk prediction coefficient for each point to the three-dimensional model of the excavation space to generate the excavation space risk distribution model.

[0134] Furthermore, the system is also used to implement the following functions:

[0135] Perform feature recognition on each point according to the excavation space data set to obtain feature data of each point; train a point risk prediction network based on a recurrent neural network; input the feature data of each point into the point risk prediction network to generate a risk prediction coefficient for each point.

[0136] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0137] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0138] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. An intelligent control method for a tunneling path of a roadheader, characterized in that: The method comprises: Conduct risk prediction for the area to be excavated by the roadheader and establish a risk distribution model for the excavation space; Performing hierarchical encirclement collision detection on the tunneling space risk distribution model according to the tunneling machine to construct a safe tunneling motion space model; Performing excavation path planning on the safe excavation motion space model according to the excavation target of the area to be excavated, and building a first excavation path planning domain; Performing overrun feedback optimization on the first tunneling path planning domain according to technical parameter constraints of the tunnel boring machine to establish a second tunneling path planning domain; performing joint optimization on the second excavation path planning domain according to a predetermined excavation evaluation factor to determine an excavation path optimization result; Performing tunneling path tracking control on the tunnel boring machine according to the tunneling path optimization result; Performing overrun feedback optimization on the first tunneling path planning domain according to technical parameter constraints of the tunnel boring machine to establish a second tunneling path planning domain includes: Performing an overrun evaluation on each excavation path within the first excavation path planning domain according to the technical parameter constraints to obtain a plurality of path overrun coefficients; Classifying the first excavation path planning domain according to the plurality of path excess coefficients to obtain a first cluster of excavation paths having less than a predetermined path excess coefficient and a second cluster of excavation paths having greater than or equal to the predetermined path excess coefficient; Adaptively adjusting the second cluster of tunneling paths according to the technical parameter constraints to obtain a third cluster of tunneling paths; Merging the first cluster of tunneling paths and the third cluster of tunneling paths to generate the second tunneling path planning domain; Performing joint optimization on the second excavation path planning domain according to a predetermined excavation evaluation factor to determine an excavation path optimization result includes: Setting excavation evaluation expectations according to the predetermined excavation evaluation factors, wherein the predetermined excavation evaluation factors include excavation efficiency and excavation safety; performing an optimization analysis on the second excavation path planning domain according to the excavation evaluation expectation to establish a third excavation path planning domain; Perform weight allocation according to the predetermined excavation evaluation factors to establish an excavation quality evaluation function; The excavation optimality is maximized in the third excavation path planning domain according to the excavation optimality evaluation function to generate the excavation path optimization result.

2. The intelligent control method for the tunneling path of a tunnel boring machine according to claim 1, characterized in that: Performing hierarchical encirclement collision detection on the tunneling space risk distribution model according to the tunneling machine to construct a safe tunneling motion space model, including: Modeling the tunnel boring machine to construct a tunnel boring machine model; Based on the tunnel boring machine model, simulated collision detection is performed on the tunnel boring space risk distribution model according to a multi-level bounding box to construct a first safe tunnel boring space model, wherein the multi-level bounding box includes a plurality of box-shaped colliders and a plurality of mesh colliders; Performing point cleaning on the tunneling space risk distribution model according to the point risk threshold to construct a second safe tunneling space model; The intersection of the first safe excavation space model and the second safe excavation space model is obtained to generate the safe excavation motion space model.

3. The intelligent control method for the tunneling path of a tunnel boring machine according to claim 2, characterized in that: Based on the tunnel boring machine model, a simulated collision detection is performed on the tunnel boring space risk distribution model according to a multi-level bounding box to construct a first safe tunnel boring space model, including: Based on the tunnel boring machine model and the multi-level bounding box, simulated collision detection is performed on the tunneling space risk distribution model to construct a first simulated collision space model; Randomly combining the multi-level bounding boxes to obtain a multi-level bounding architecture; Based on the roadheader model, simulated collision detection is performed on the tunneling space risk distribution model according to the multi-level encirclement architecture to construct a second simulated collision space model; performing union identification on the first simulated collision space model and the second simulated collision space model to generate a simulated collision integrated space model; The difference between the simulated collision integrated space model and the excavation space risk distribution model is used as the first safe excavation space model.

4. The intelligent control method for the tunneling path of a tunnel boring machine according to claim 3, characterized in that: Based on the roadheader model and according to the multi-level bounding box, simulated collision detection is performed on the tunneling space risk distribution model to construct a first simulated collision space model, including: Extracting a first bounding box according to the multi-level bounding box; Abstractly enclosing the roadheader model according to the first bounding box to obtain a first enclosed roadheader model; Perform multiple simulated collisions on the tunneling space risk distribution model according to the first surrounding tunnel boring machine model to obtain multiple simulated collision spaces; The multiple simulated collision spaces are merged to obtain a first enclosed simulated collision space, and the first enclosed simulated collision space is added to the first simulated collision space model.

5. The intelligent control method for the tunneling path of a tunnel boring machine according to claim 1, characterized in that: Performing an optimization analysis on the second excavation path planning domain according to the excavation evaluation expectation to establish a third excavation path planning domain, including: Extracting an nth excavation path according to the second excavation path planning domain, where n is a positive integer; Performing feature prediction based on the nth excavation path to determine an nth excavation prediction result, wherein the nth excavation prediction result includes an nth predicted excavation efficiency coefficient and an nth predicted excavation safety coefficient; Determining whether the nth excavation prediction result meets the excavation evaluation expectation; If the nth excavation prediction result meets the excavation evaluation expectation, adding the nth excavation path to the third excavation path planning domain; If the nth excavation prediction result does not meet the excavation evaluation expectation, the nth excavation path is eliminated.

6. The intelligent control method for the tunneling path of a tunnel boring machine according to claim 1, characterized in that: Risk prediction is conducted for the area to be excavated by the roadheader, and a risk distribution model for the excavation space is established, including: Collecting above-ground and underground parameters of the area to be excavated to obtain an excavation space data set; Performing three-dimensional reconstruction based on the excavation space data set to obtain a three-dimensional model of the excavation space; Perform risk prediction on each point in the area to be excavated based on the excavation space data set to obtain a risk prediction coefficient for each point; The risk prediction coefficients of each point are rendered to the three-dimensional model of the tunneling space to generate the tunneling space risk distribution model.

7. The intelligent control method for the tunneling path of a tunnel boring machine according to claim 6, characterized in that: Perform risk prediction on each point in the area to be excavated based on the excavation space data set to obtain a risk prediction coefficient for each point, including: Performing feature recognition on each of the points according to the tunneling space data set to obtain feature data of each point; Based on recurrent neural network, train the point risk prediction network; The feature data of each point is input into the point risk prediction network to generate the risk prediction coefficient of each point.

8. An intelligent control system for a tunnel boring machine's excavation path, characterized in that: The system is used to implement the intelligent control method for a tunneling path of a tunnel boring machine according to any one of claims 1 to 7, and the system includes: The risk prediction module is used to predict the risk of the area to be excavated by the roadheader and establish a risk distribution model for the excavation space; a safe excavation motion space model construction module, configured to perform hierarchical encirclement collision detection on the excavation space risk distribution model according to the excavation machine, and construct a safe excavation motion space model; An excavation path planning module is used to plan an excavation path for the safe excavation motion space model according to the excavation target of the area to be excavated, and to build a first excavation path planning domain; an overrun feedback optimization module, configured to perform overrun feedback optimization on the first tunneling path planning domain according to technical parameter constraints of the tunneling machine to establish a second tunneling path planning domain; a tunneling path optimization result determination module, configured to perform joint optimization on the second tunneling path planning domain according to a predetermined tunneling evaluation factor to determine a tunneling path optimization result; A tunneling path tracking control module, configured to perform tunneling path tracking control on the tunneling machine according to the tunneling path optimization result; The system is also used to implement the following functions: performing an overrun evaluation on each excavation path within the first excavation path planning domain according to the technical parameter constraints to obtain a plurality of path overrun coefficients; classifying the first excavation path planning domain according to the plurality of path overrun coefficients to obtain a first cluster of excavation paths having less than a predetermined path overrun coefficient and a second cluster of excavation paths having greater than or equal to the predetermined path overrun coefficient; adaptively adjusting the second cluster of excavation paths according to the technical parameter constraints to obtain a third cluster of excavation paths; and merging the first cluster of excavation paths and the third cluster of excavation paths to generate the second excavation path planning domain; According to the predetermined excavation evaluation factors, excavation evaluation expectations are set, and the predetermined excavation evaluation factors include excavation efficiency and excavation safety; according to the excavation evaluation expectations, the second excavation path planning domain is optimized and analyzed to establish a third excavation path planning domain; according to the predetermined excavation evaluation factors, weights are allocated to establish an excavation optimality evaluation function; according to the excavation optimality evaluation function, the third excavation path planning domain is optimized to maximize the excavation optimality to generate the excavation path optimization result.

Images