A method for designing and constructing a 3D printed tunnel lining structure by jetting

Abstract

The application is a kind of jet 3D printing tunnel lining structure and its path design and construction method. The method first fuses the tunnel design BIM model, geological exploration data and initial support scanning point cloud to obtain a tunnel lining digital twin model containing design target surface, surrounding rock grade partition, surrounding rock pressure boundary and initial deformation state; based on the tunnel lining digital twin model, output mechanical partition results, optimized three-dimensional printing path set, regional material parameter table and robot control instruction set; according to the robot control instruction set, implement layered and partitioned jet printing, and obtain printed structure scanning point cloud and forming quality evaluation results after each printing unit is completed; compare the printed structure scanning point cloud with the design target surface in the tunnel lining digital twin model to obtain a forming error field, and generate a compensation control instruction set to update the robot control instructions that have not been executed. The application realizes closed-loop compensation and improves continuous construction efficiency.

Description

Technical Field

[0001] This invention belongs to the field of intelligent construction and additive manufacturing technology for tunnel engineering, specifically relating to a jet 3D printed tunnel lining structure and its path design and construction method, a central processing system for implementing the method, and a tunnel lining structure constructed using the method. Background Technology

[0002] The tunnel lining structure bears the load of the surrounding rock and maintains the stability of the tunnel's operating space. Its geometric accuracy, density, and material continuity directly affect the tunnel's safety and durability. Current shotcrete construction usually relies on manual or semi-automatic robotic arms to operate according to a preset reciprocating trajectory. There is no clear mapping relationship between the spraying path and the distribution of surrounding rock pressure and the local stress state of the lining.

[0003] In areas with significant curvature changes, such as the arch crown and arch waist, fixed row spacing and fixed spraying parameters can easily lead to localized buildup, increased rebound, or insufficient thickness. Continuing to use high-density paths in relatively low-stress areas such as sidewalls will increase material consumption and idle travel time. Furthermore, traditional paths typically only describe the geometric trajectory and do not incorporate the material's early-age support capacity, robotic arm accessibility, spraying distance constraints, and the initial setting time window of the concrete into the planning process.

[0004] Furthermore, existing spraying operations often involve offline measurements after the completion of a lining section, making it difficult to promptly use scanning results to correct subsequent printed layers. This leads to the accumulation of localized depressions or protrusions layer by layer. Therefore, it is necessary to establish a spraying 3D printing tunnel lining structure path design and construction method that can clearly convey the "digital model—mechanical zoning—printing path—control commands—scanning error—compensation commands". Summary of the Invention

[0005] In view of the problems in the existing technology, such as the disconnect between path planning and stress requirements, unclear meaning of layering and zoning, lack of quantitative utilization of the early-age support capacity of materials, and lack of closed-loop compensation for forming errors, the technical problem to be solved by the present invention is to provide a path design and construction method, a central processing system and a lining structure for jet 3D printing tunnel lining.

[0006] The technical solution adopted by the present invention to solve the aforementioned technical problem is as follows: A method for designing and constructing a jet-printed 3D-printed tunnel lining structure includes the following steps: Step S1, Multi-source data fusion modeling: Register the tunnel design BIM model, geological exploration data and initial support scanning point cloud to a unified coordinate system to obtain a digital twin model of the tunnel lining that includes the design target surface, the actual surface of the initial support, the tunnel design centerline, the lining design thickness, the surrounding rock grade zoning, the surrounding rock pressure boundary and the initial deformation state. Step S2, Mechanical-Process Collaborative Path Planning: Finite element analysis is performed on the digital twin model of the tunnel lining to obtain mechanical partitioning results; based on the mechanical partitioning results, axial stepped segmentation, cross-sectional partitioning coupling, thickness progressive layering, gradient density path planning, and stress streamline path generation are implemented to obtain an optimized 3D printing path set; based on the mechanical partitioning results and the optimized 3D printing path set, a regional material parameter table is obtained, and the optimized 3D printing path set is discretized into a robot control instruction set; Step S3, Adaptive jet printing construction: Supply jet concrete according to the material parameter table of the area, and carry out jet printing according to the robot control instruction set and the layered and partitioned printing strategy. After each printing unit is completed, obtain the scanned point cloud of the printed structure and the evaluation result of the forming quality. Step S4, Closed-loop quality control: The printed structure scan point cloud is compared with the design target surface in the digital twin model of the tunnel lining to obtain the forming error field. A layered progressive compensation strategy is used to generate a compensation control instruction set, and the compensation control instruction set is used to update the instructions in the robot control instruction set that have not yet been executed.

[0007] Furthermore, in the layered and partitioned printing strategy, the printing unit is printed along the tunnel axis according to the longitudinal printing unit length L. u It is divided into M vertical printing units, and along the circumferential direction into the arch foot area, side wall area, arch waist area and arch top area, and along the thickness direction into N thickness printing layers; the intersection of a thickness printing layer, a vertical printing unit and a circumferential functional area constitutes a printing unit. The axial stepped segmentation causes the axial starting point of the i-th thickness printing layer to be moved back by ΔL relative to the axial starting point of the (i-1)-th thickness printing layer. i ΔL i 300–800 mm; The progressive thickness layering adjusts the target thickness or interlayer waiting time of the next printing layer based on the material's early-age support capacity and the measured deformation of the already printed layers.

[0008] Furthermore, the gradient density path planning outputs a gradient density path parameter matrix, which includes at least the path row spacing, path overlap rate, path direction, printing order, and thickness printing layer number for each printing unit. The row spacing gradually increases from the arch crown to the arch waist to the side walls and arch feet, while the path overlap rate gradually decreases from the arch crown to the arch waist to the side walls and arch feet. Specifically, the path row spacing in the high-stress area of ​​the arch crown ranges from 300 to 400 mm, and the path overlap rate ranges from 20% to 30%. The path row spacing in the medium-stress area of ​​the arch waist ranges from 400 to 500 mm, and the path overlap rate ranges from 10% to 20%. The path row spacing in the relatively low-stress areas of the side walls and arch feet ranges from 500 to 800 mm, and the path overlap rate ranges from 5% to 15%. The stress streamline path generation includes principal stress direction field interpolation, seed point arrangement, streamline integration, robotic arm accessibility constraints, minimum radius of curvature constraints, spray distance constraints, minimum spacing constraints between adjacent paths, as well as path smoothing and resampling.

[0009] Furthermore, the material parameter table for the area includes at least the material formula number, yield stress, plastic viscosity, setting time, and early-age strength; the high-stress area of ​​the arch crown uses early-strength shotcrete with a setting time of 3 to 5 minutes, while the relatively low-stress areas of the sidewalls and arch foot use standard shotcrete with a setting time of 8 to 12 minutes. The material formula for each area is independently adjusted by the path design and construction central processing system.

[0010] Furthermore, in step S3, the injection parameter-forming index mapping model Qp=a0+a1v is adopted. n +a2p n +a3d n +a4q n Calculate the predicted value of the forming index; Among them, v n p n d n and q n These are the normalized values ​​for nozzle moving speed, injection pressure, spray distance, and injection flow rate, respectively, with a0 to a4 being calibration coefficients. Within the allowable parameter range, select a combination of spraying parameters with a forming index value Qp not lower than the forming index threshold and low material consumption, and write it into the robot control instructions; in areas with large path curvature, reduce the nozzle moving speed by 10% to 30% and reduce the spray flow rate by 5% to 15% relative to the reference value.

[0011] Furthermore, the layered progressive compensation strategy divides the forming error field into segments along the axial direction according to the longitudinal printing units and into blocks along the circumferential direction according to 30° sectors, and allocates the target compensation amount of each error block to the subsequent thickness printing layers that have not yet been printed; when the absolute value of the negative error exceeds 20% of the target thickness of the current thickness printing layer, a local reinforcement path is inserted in the next thickness printing layer; when the positive error exceeds 15% of the target thickness of the current thickness printing layer, the subsequent path is shifted inward along the design normal; 90% of the target compensation amount is allocated to no more than 3 subsequent thickness printing layers.

[0012] Furthermore, the branch and bound algorithm is used to optimize the execution order of the path segments to be printed. The sum of the idle travel time, attitude adjustment time, inter-layer waiting time and penalty value of exceeding the initial condensation time window between path segments is used as the cost. The upper bound is determined according to the current feasible order, and the lower bound is determined according to the minimum entry cost of the unvisited path segment and the necessary printing time. Branches with a lower bound not less than the current upper bound are pruned.

[0013] This invention also protects a central processing system for implementing the path design and construction method, comprising a data fusion module, a mechanical analysis engine, an intelligent path planner, a process parameter optimizer, and a real-time compensation controller; the data fusion module is used to output a digital twin model of the tunnel lining, the mechanical analysis engine is used to output mechanical zoning results, the intelligent path planner is used to output an optimized 3D printing path set, the process parameter optimizer is used to output a regional material parameter table and a robot control instruction set, and the real-time compensation controller is used to output a compensation control instruction set based on the forming error field.

[0014] This invention also protects a jet-printed 3D-printed tunnel lining structure, constructed using the aforementioned path design and construction method. The tunnel lining structure has a stepped oblique overlapping interface along the axial direction formed by the retreat of the starting point of adjacent thickness printing layers, a gradient density printing path along the circumferential direction that matches the mechanical zoning results, and is formed by the superposition of multiple thickness printing layers that satisfy the early-age support criterion along the thickness direction.

[0015] Compared with the prior art, the present invention has the following beneficial effects: (1) Clear data transfer relationships. Each step provides a definite output result and subsequent usage location, avoiding disconnect between the digital model, path planning results and compensation results.

[0016] (2) Match the path with the stress requirements. Adjust the path row spacing, overlap rate and path direction according to the mechanical zoning results to obtain a higher path density in high stress areas and reduce unnecessary material accumulation in relatively low stress areas.

[0017] (3) Layered and partitioned rules can be executed. The thickness printing layer, the vertical printing unit and the circumferential functional area together constitute the printing unit. The robotic arm can perform jetting, scanning and compensation in a unified numbered sequence.

[0018] (4) Timely closed-loop compensation. The scanning error is distributed along the longitudinal printing unit, the circumferential sector and the remaining printing layer, and the main error compensation is completed step by step in the subsequent printing layers.

[0019] (5) The coordinated optimization of the robotic arm trajectory and the initial setting time window can reduce idle travel, inter-layer reversal and waiting time, and improve the efficiency of continuous construction. Attached Figure Description

[0020] Figure 1 This is a structural block diagram of a central processing system according to an embodiment of the present invention; Detailed Implementation

[0021] The present invention will be further described below with reference to specific embodiments and accompanying drawings. The parameter ranges described below are used to illustrate possible embodiments and do not constitute the sole limitation on the scope of protection of the present invention.

[0022] The present invention provides a method for path design and construction of jet 3D printed tunnel lining structures, comprising the following steps: Step S1 integrates the tunnel design BIM model, geological exploration data and initial support scanning point cloud to obtain a tunnel lining digital twin model that includes the design target surface, surrounding rock grade zoning, surrounding rock pressure boundary and initial deformation state, and outputs the tunnel lining digital twin model. Step S2 reads the digital twin model of the tunnel lining and performs finite element numerical analysis to obtain mechanical zoning results. Then, according to the rules of axial stepped segmentation, cross-sectional zoning coupling and thickness progressive layering, an optimized three-dimensional printing path set is generated, and a regional material parameter table and robot control instruction set are formed. Step S3: The intelligent trolley equipped with a seven-degree-of-freedom robotic arm performs layered and partitioned jet printing according to the robot control instruction set, and obtains the scanned point cloud of the printed structure and the forming quality evaluation results after each printing unit is completed; Step S4 reads the printed structure scan point cloud and compares it with the design target surface in the tunnel lining digital twin model output in step S1, outputs the forming error field and compensation control instruction set, and then feeds the compensation control instruction set back to step S3 to update the robot control instructions that have not yet been executed.

[0023] By clearly defining the output results of each step and their relationship to subsequent steps, the printing path density, material properties, and spraying parameters can be matched with the stress requirements of the lining, and closed-loop compensation for forming errors can be achieved.

[0024] This invention also provides a central processing system, including a data fusion module, a mechanical analysis engine, an intelligent path planner, a process parameter optimizer, and a real-time compensation controller. Each module transmits the tunnel lining digital twin model, mechanical zoning results, optimized 3D printing path set, regional material parameter table, robot control instruction set, forming error field, and compensation control instruction set.

[0025] The present invention also provides a jet-printed 3D printed tunnel lining structure, wherein the tunnel lining structure has a stepped oblique overlapping interface along the axial direction, a gradient density printing path that matches the mechanical zoning results along the circumferential direction, and is formed by stacking multiple printing layers that satisfy the early-age support criterion along the thickness direction.

[0026] Example 1 This embodiment describes a path design and construction method for jet 3D printed tunnel lining structures, including the following steps: Step S1: Multi-source data fusion modeling Acquire the tunnel design BIM model, geological exploration data, and initial support scanning point cloud. Use the design centerline in the tunnel design BIM model as a unified axial reference. Transform the initial support scanning point cloud to the BIM coordinate system through coarse registration of feature points and fine registration of iterative nearest points. Map the geological exploration data to the corresponding lining surface according to mileage and circumferential orientation.

[0027] After data cleaning, coordinate registration, and attribute mapping, a digital twin model of the tunnel lining is obtained. This digital twin model includes at least: the design target surface, the actual surface of the initial support, the tunnel design centerline, the lining design thickness, the surrounding rock grade zoning, the surrounding rock pressure boundary, material mechanical parameters, and the initial deformation state. In step S2, the digital twin model serves as the input for finite element analysis and path planning, and in step S4, it serves as the design benchmark for comparing scanning point cloud errors.

[0028] Step S2: Mechanics-Process Collaborative Path Planning Step S2 includes mechanical partitioning, 3D printing unit division, gradient density path planning, stress streamline path generation, axial stepped segmentation, thickness progressive layering, spatial continuous spiral path generation, regional material optimization, and control command discretization.

[0029] 1. Mechanical Division.

[0030] Read the digital twin model of the tunnel lining obtained in step S1, establish a lining-surrounding rock finite element model, and calculate the principal stress, equivalent stress, deformation, and safety reserve of the lining surface. Divide the lining surface into zones according to stress level, deformation level, and functional location to obtain mechanical zoning results. These mechanical zoning results include the zone number, stress level, principal stress direction, design thickness, allowable forming error, and early-age material support requirements for each zone, and are used in subsequent gradient density path planning, stress streamline path generation, and zone material optimization.

[0031] 2. 3D printing unit division.

[0032] Print the unit length L along the tunnel axis in the longitudinal direction. u Divide into segments, with a vertical printing unit length L. u The number of longitudinal printing units M is determined based on the effective working range of the robotic arm, the stepping distance of the intelligent trolley, and the initial setting time window of the concrete; for a lining section with an axial length of L, the number of longitudinal printing units M is not less than L / L. u The smallest integer. The area is uniformly divided along the circumferential direction into the arch foot area, sidewall area, arch waist area, and arch crown area; and along the thickness direction into N thickness printing layers. A printing unit is defined as the intersection of a thickness printing layer, a vertical printing unit, and a circumferential functional area.

[0033] 3. Gradient density path planning.

[0034] The regional stress values ​​in the mechanical partitioning results are normalized to a stress coefficient ξ between 0 and 1. For each printing unit, the path line spacing s and path overlap rate λ are determined based on the stress coefficient ξ: when the stress coefficient increases, the path line spacing decreases and the path overlap rate increases; when the stress coefficient decreases, the path line spacing increases and the path overlap rate decreases. The path density D is defined as D = 1 / s. The region number, path line spacing, path overlap rate, path direction, printing order, and thickness printing layer number of each printing unit together form the gradient density path parameter matrix.

[0035] Taking horseshoe-shaped tunnel lining as an example, the row spacing gradually increases from the arch crown to the arch waist, sidewalls, and arch foot, while the path overlap rate gradually decreases from the arch crown to the arch waist, sidewalls, and arch foot. Specifically, the row spacing in the high-stress area of ​​the arch crown ranges from 300 to 400 mm, and the path overlap rate ranges from 20% to 30%; the row spacing in the medium-stress area of ​​the arch waist ranges from 400 to 500 mm, and the path overlap rate ranges from 10% to 20%; while the row spacing in the relatively low-stress areas of the sidewalls and arch foot ranges from 500 to 800 mm, and the path overlap rate ranges from 5% to 15%. Linear interpolation or spline interpolation is used to smoothly transition between adjacent areas to avoid abrupt changes in path density. The gradient density path parameter matrix is ​​the output of gradient density path planning and is used to generate an optimized 3D printing path set.

[0036] 4. Stress streamline path generation.

[0037] In high-stress or weak areas, an improved stress streamline generation algorithm can be used. This algorithm first interpolates the maximum principal stress direction of the finite element nodes as a continuous direction field of the lining surface. Then, seed points are placed in the high-stress area, and candidate stress streamlines are generated by integrating along the continuous direction field. Subsequently, constraints such as robotic arm accessibility, minimum radius of curvature, allowable spray distance, minimum spacing between adjacent paths, and region boundary are added to trim, smooth, and resample the candidate stress streamlines. The constrained stress streamline paths are merged with gradient density paths to enhance material continuity along the principal stress directions.

[0038] 5. Axial stepped segmentation.

[0039] Let z be the axial starting point of the nozzle of the i-th thickness printing layer. i,o The axial backlash between adjacent thickness printing layers is ΔL. i Then, the starting point of the upper path moves backward relative to the starting point of the lower path along the axial direction, satisfying: z i+1,o = z i,o - ΔL i i = 1, 2, 3, ..., N Where, ΔL i It can be set to 300–800 mm, where N is the number of axial stepped segments, and z i+1,o This is the axial endpoint of the nozzle of the i-th thickness printing layer, which is also the axial starting point of the nozzle of the (i+1)-th thickness printing layer. The lower printing layer advances axially first. Once its early-age support capacity meets the support capacity criterion for progressively thicker layers, the upper printing layer begins printing from its retreating starting point, thus forming a stepped, obliquely overlapping interface between adjacent thickness printing layers. This interface disperses the interlayer seams axially, preventing the start and end points of all layers from concentrating on the same cross-section.

[0040] 6. Layered thickness progression.

[0041] Let the total thickness of the lining be H, and the target thickness of the i-th printed layer be h. i ,satisfy: h1 + h2 + ... + h N = H The early-age support capacity of the material is characterized by the early-age compressive strength σg(t) or equivalent support strength obtained from the calibration of the same batch of materials. Before printing the next thickness layer, it is determined whether the already printed layers simultaneously meet the early-age support criterion and the deformation criterion. σg(t i ) ≥ γρgh i+1 , and δi ≤ δlim Among them, t i Let σg(t) be the age after the i-th thickness printing layer is completed. i ) is the t-th i Early-age compressive strength at each age; γ is the safety factor, ρ is the material density, g is the gravitational acceleration, h i+1 δ is the target thickness for the next thickness printing layer. i The measured deformation of the printed layer is δlim, which is the allowable deformation limit. When the criterion is not met, the central processing system reduces the target thickness of the next printed layer, extends the interlayer waiting time, or increases the early strength grade of the material in the corresponding area; when the criterion is met, the printing of the next printed layer continues. The above process is the specific meaning of "dynamically adjusting the single-layer thickness according to the early-age support capacity of the material".

[0042] 7. Generation of spatially continuous spiral paths.

[0043] For lining sections suitable for continuous circumferential operations, the tunnel's design centerline is taken as the z-axis, and x-axis and y-axis are established in a section perpendicular to the z-axis. The spatially continuous spiral path can be represented by the circumferential parameter θ as follows: x(θ) = x c + R(θ) cos[θ + φ(z(θ))] y(θ) = y c + R(θ) sin[θ + φ(z(θ))] z(θ) = z o + kθ Where x(θ) and y(θ) are the coordinates of the path point within the tunnel cross-section, and z(θ) is the coordinate of the path point along the tunnel axis; x c and y c The coordinates of the current cross-section center are: θ is the circumferential parameter; R(θ) is the inner contour radius of the lining that varies with the circumferential parameter θ; φ(z) is the axial phase correction function used to avoid construction joints, embedded parts, or local obstacles; z o is the axial starting coordinate of the spiral path; k is the axial advance amount corresponding to each radian.

[0044] The helix angle α and the axial thrust k satisfy the following relationship: k=R e tanα Among them, R e This is the equivalent radius of the current lining section. The helix angle α is determined based on the material's yield stress τ, plastic viscosity η, and injection velocity v. s Determined through a pre-defined constructibility mapping function: α = f(τ, η, v) s ) The central processing system first obtains the helical angle α from the constructibility mapping function, and then uses k=R e tanα is used to calculate the axial thrust k, and k is then substituted into the spatial continuous helical path parameter equation. Therefore, the helix angle α is directly used in the aforementioned parameter equation, rather than being an unused parameter independent of the path equation.

[0045] 8. Discretization of regional material optimization and control commands.

[0046] Based on the mechanical zoning results, gradient density path parameter matrix, target thickness, and interlayer waiting time, the required yield stress, plastic viscosity, setting time, and early-age strength for each printing unit are calculated to form a regional material parameter table. Early-strength shotcrete with a setting time of 3–5 minutes can be used in the high-stress area of ​​the arch crown, while standard shotcrete with a setting time of 8–12 minutes can be used in the relatively low-stress areas of the sidewalls and arch feet. The material parameters in the arch waist area transition smoothly between adjacent areas.

[0047] The axial stepped path, cross-sectional partitioned coupling path, thickness progressive layered path, gradient density path, stress streamline path, and spatial continuous spiral path are merged and subjected to collision checks, curvature checks, and spray distance checks to obtain an optimized 3D printing path set. This optimized 3D printing path set is then discretized according to a set sampling interval. For each discrete point, position, attitude, spray flow rate, spray pressure, nozzle movement speed, spray distance, material formula number, and printing unit number are written to form a robot control instruction set. This robot control instruction set is directly executed in step S3.

[0048] Step S3: Adaptive Inkjet Printing Application The intelligent trolley, equipped with a seven-DOF robotic arm, moves to the current longitudinal printing unit. It calculates the transformation matrix from the model coordinate system to the construction coordinate system using a laser target or on-site reference points, and converts the path points in the robot control command set into executable coordinates for the robotic arm. Simultaneously, the central processing system reads the regional material parameter table and supplies matching shotcrete to the corresponding printing unit.

[0049] The specific meaning of the layered and zoned printing strategy is as follows: Print layer by layer along the thickness direction in the order of the 1st thickness printing layer to the Nth thickness printing layer; within each thickness printing layer, advance along the axial direction in the order of the 1st longitudinal printing unit to the Mth longitudinal printing unit; within each longitudinal printing unit, print along the circumferential direction in the order of the arch foot area, side wall area, arch waist area, and arch crown area from bottom to top, with symmetrical alternating operations on the left and right sides. The intersection of a thickness printing layer, a longitudinal printing unit, and a circumferential functional area constitutes a printing unit.

[0050] Before executing each path segment, the process parameter optimizer calculates the normalized values ​​of nozzle movement speed v, injection pressure p, nozzle distance d, and injection flow rate q. n p n d n and q n Calculate the predicted molding index value Qp. The following locally linear mapping model can be used: Qp=a0+a1v n +a2p n +a3d n +a4q n Among them, v n p n d n and q n These are the normalized values ​​for nozzle moving speed, injection pressure, spray distance, and injection flow rate, respectively, with a0 to a4 being calibration coefficients.

[0051] The process parameter optimizer generates candidate parameter combinations within the allowable parameter range, calculates the predicted forming index (Qp) for each candidate parameter combination, and prioritizes parameter combinations that satisfy Qp not lower than the forming index threshold and have low material consumption. The selected parameters are then written into the robot control instructions for the current path segment. Therefore, the spraying parameter-forming quality mapping model is used to select parameters before spraying, rather than just for post-spraying scoring. The forming index value Qp is generally between 0 and 10, and the specific value of the forming index threshold can be determined based on previous experiments.

[0052] For areas with significant path curvature, the nozzle movement speed can be reduced by 10%–30% from the baseline value (the system's default spray parameters), and the spray flow rate can be reduced by 5%–15% simultaneously to minimize buildup on the inner side of the curve. For straight or gently curving path sections, a constant speed and constant flow rate mode can be used. After each printing unit is completed, the 3D laser scanner at the end of the robotic arm acquires the point cloud of the printed structure for that unit. The central processing system then uses this data to simultaneously calculate thickness deviation, surface flatness, and continuity indicators. The forming quality evaluation result is composed of thickness deviation, surface flatness and continuity index, and forming index value Qp.

[0053] Step S4: Closed-loop quality control The printed structure scan point cloud obtained in step S3 is transformed into the unified coordinate system established in step S1, and the nearest point distance between it and the design target surface in the digital twin model of the tunnel lining is calculated to obtain a signed forming error field. It is agreed that when the scanned surface is located inside the design target surface, it is a negative error, i.e., a depression; when the scanned surface is located outside the design target surface, it is a positive error, i.e., a convexity.

[0054] The specific meaning of the layered progressive compensation strategy is as follows: The system is divided into segments along the axial direction according to the longitudinal printing units, and into blocks along the circumferential direction according to 30° sectors. The target compensation amount for each error block is allocated to subsequent thickness printing layers that have not yet been printed, and implemented layer by layer. This strategy is first executed in step S4 and fed back to step S3 through the compensation control instruction set.

[0055] When the absolute value of the negative error of an error block exceeds 20% of the target thickness of the current thickness printing layer, a local reinforcement path is inserted at the corresponding position in the next thickness printing layer, and the corresponding jet flow rate is increased or the nozzle movement speed is reduced. When the positive error of an error block exceeds 15% of the target thickness of the current thickness printing layer, the subsequent path is shifted inward by a corresponding distance along the design normal, and the jet flow rate is reduced accordingly. To avoid new accumulation or slippage caused by a one-time correction, 90% of the target compensation amount can be allocated to no more than 3 subsequent thickness printing layers, and the remaining compensation amount is used to continue correction based on subsequent scanning results.

[0056] The real-time compensation controller writes the path offset, injection flow correction, movement speed correction, injection pressure correction, and spray distance correction into the compensation control instruction set, and updates the instructions in the robot control instruction set that have not yet been executed using the compensation control instruction set. After the update is completed, step S3 continues printing; the newly printed structure scan point cloud re-enters step S4, forming a closed loop.

[0057] Printing efficiency optimization To reduce the idle travel time and inter-layer reversal time of the robotic arm, a branch and bound algorithm can be used to optimize the execution order of printing units and path segments. Each path segment to be printed is treated as a node, and the sum of the idle travel time, attitude adjustment time, inter-layer waiting time, and penalty value exceeding the initial coagulation time window between path segments is used as the edge cost. First, a feasible order is obtained using the nearest neighbor rule and used as the current upper bound. Then, the lower bound is calculated based on the minimum entry cost of unvisited nodes and the necessary printing time. When the lower bound of a branch is not less than the current upper bound, it is pruned; otherwise, it continues to expand until the path order with the minimum total cost or the termination accuracy is obtained is obtained.

[0058] The length of the longitudinal printing unit Lᵤ can also be dynamically adjusted according to the initial setting time window of the concrete: when the estimated time required to complete a longitudinal printing unit is close to the upper limit of the initial setting time, the length of the longitudinal printing unit is reduced; when the robotic arm operation is continuous and the material remains sprayable, the length of the longitudinal printing unit is appropriately increased.

[0059] Central Processing System The central processing system includes: a data fusion module for outputting a digital twin model of the tunnel lining; a mechanical analysis engine for outputting mechanical zoning results; an intelligent path planner for performing gradient density path planning, an improved stress streamline generation algorithm, axial stepped segmentation, cross-sectional zoning coupling, thickness progressive layering, and spatial continuous spiral path generation, and outputting an optimized 3D printing path set; a process parameter optimizer for outputting regional material parameter tables and robot control instruction sets; and a real-time compensation controller for outputting a compensation control instruction set based on the forming error field.

[0060] The gradient density path planning does not only output the abstract concept of "gradient density", but also outputs a gradient density path parameter matrix that includes the path row spacing, path overlap rate, path direction, printing order and thickness printing layer number of each printing unit. The improved stress streamline generation algorithm adds robotic arm accessibility, minimum radius of curvature, spray distance and minimum spacing between adjacent paths to the conventional principal stress streamline integral. Its output is a stress streamline path that can be continuously executed by the robotic arm.

[0061] Example The implementation process of this invention is illustrated using a horseshoe-shaped tunnel lining as an example. First, the tunnel design BIM model of the lining section is imported, the initial support scanning point cloud obtained by on-site 3D laser scanning is registered to the BIM coordinate system, and the surrounding rock grade, surrounding rock pressure and initial deformation data are mapped to the lining surface according to the mileage to obtain a digital twin model of the tunnel lining.

[0062] Finite element analysis was performed based on the digital twin model of the tunnel lining to obtain the high-stress zone at the arch crown, the medium-stress zone at the arch waist, and the relatively low-stress zones at the sidewalls and arch feet. The length of the longitudinal printing unit was determined according to the effective operating range of the intelligent trolley and the initial setting time window of the concrete. Each longitudinal printing unit was divided circumferentially into the arch foot zone, sidewall zone, arch waist zone, and arch crown zone, and further divided into multiple thickness printing layers along the thickness direction.

[0063] In the high-stress zone of the arch crown, smaller path row spacing and larger path overlap are used, while in the relatively low-stress zones of the sidewalls and arch feet, larger path row spacing and smaller path overlap are used. Stress flow lines are generated in locally weak areas based on the direction of the maximum principal stress. The axial starting points of adjacent thickness printing layers are successively moved back, creating a stepped, obliquely overlapping interface between layers. The central processing system determines whether the support criteria for the next thickness printing layer are met based on the early-age strength calibration curve of the same batch of materials.

[0064] The central processing system merges the above paths into an optimized set of 3D printing paths, and generates a regional material parameter table and a robot control instruction set. After the intelligent trolley arrives at the current longitudinal printing unit, it establishes a construction coordinate system and performs jet printing sequentially according to the hierarchy of "thickness printing layer - longitudinal printing unit - circumferential functional area". After each printing unit is completed, the 3D laser scanner acquires the scanned point cloud of the printed structure.

[0065] The real-time compensation controller compares the scanned point cloud of the printed structure with the design target surface in the digital twin model of the tunnel lining to obtain the forming error field. For recessed areas, local reinforcement paths are added to the next thickness of the printed layer; for protruding areas, subsequent paths are shifted inward toward the design target surface. After the compensation control instruction set updates the robot control instructions that have not yet been executed, the robotic arm continues printing and scanning again until the current lining segment is completed.

[0066] In this embodiment, the usage relationships of the output results of each step are as follows: the digital twin model of the tunnel lining is used for finite element analysis, path planning, and comparison of the design target surface; the mechanical zoning results are used for gradient density path planning, stress flow line path generation, and regional material optimization; the optimized 3D printing path set is used to generate the robot control instruction set; the regional material parameter table and the robot control instruction set are used for adaptive jet printing construction; the scanned point cloud of the printed structure is used to generate the forming error field; and the compensation control instruction set is used to update the robot control instructions that have not yet been executed.

[0067] The above embodiments are used to illustrate the technical concept of the present invention. Those skilled in the art can make equivalent adjustments to the length of the vertical printing unit, the path line spacing, the path overlap rate, the material parameter range, and the error threshold without departing from the essence of the present invention.

Claims

1. A method for path design and construction of jet-printed 3D-printed tunnel lining structures, characterized in that, Includes the following steps: Step S1, Multi-source data fusion modeling: Register the tunnel design BIM model, geological exploration data and initial support scanning point cloud to a unified coordinate system to obtain a digital twin model of the tunnel lining that includes the design target surface, the actual surface of the initial support, the tunnel design centerline, the lining design thickness, the surrounding rock grade zoning, the surrounding rock pressure boundary and the initial deformation state. Step S2, Mechanics-Process Collaborative Path Planning: Read the digital twin model of the tunnel lining and perform finite element analysis to obtain the mechanical zoning results; Based on the mechanical partitioning results, axial stepped segmentation, cross-sectional partitioning coupling, thickness progressive layering, gradient density path planning, and stress streamline path generation are implemented to obtain an optimized 3D printing path set; based on the mechanical partitioning results and the optimized 3D printing path set, a regional material parameter table is obtained, and the optimized 3D printing path set is discretized into a robot control instruction set; Step S3, Adaptive jet printing construction: Supply jet concrete according to the material parameter table of the area, and carry out jet printing according to the robot control instruction set and the layered and partitioned printing strategy. After each printing unit is completed, obtain the scanned point cloud of the printed structure and the evaluation result of the forming quality. Step S4, Closed-loop quality control: The printed structure scan point cloud is compared with the design target surface in the digital twin model of the tunnel lining to obtain the forming error field. A layered progressive compensation strategy is used to generate a compensation control instruction set, and the compensation control instruction set is used to update the instructions in the robot control instruction set that have not yet been executed.

2. The path design and construction method according to claim 1, characterized in that, In the layered and partitioned printing strategy, the printing unit length L is calculated along the tunnel axis. u It is divided into M vertical printing units, and along the circumferential direction into the arch foot area, side wall area, arch waist area and arch top area, and along the thickness direction into N thickness printing layers; the intersection of a thickness printing layer, a vertical printing unit and a circumferential functional area constitutes a printing unit. The axial stepped segmentation causes the axial starting point of the i-th thickness printing layer to be moved back by ΔL relative to the axial starting point of the (i-1)-th thickness printing layer. i ΔL i 300–800 mm; The progressive thickness layering adjusts the target thickness or interlayer waiting time of the next printed layer based on the material's early-age support capacity and the measured deformation of the already printed layer.

3. The path design and construction method according to claim 1, characterized in that, The gradient density path planning outputs a gradient density path parameter matrix, which includes at least the path row spacing, path overlap rate, path direction, printing order, and thickness printing layer number for each printing unit. The row spacing gradually increases from the arch crown to the arch waist to the side walls and arch feet, while the path overlap rate gradually decreases from the arch crown to the arch waist to the side walls and arch feet. Specifically, the path row spacing in the high-stress area of ​​the arch crown ranges from 300 to 400 mm, and the path overlap rate ranges from 20% to 30%. The path row spacing in the medium-stress area of ​​the arch waist ranges from 400 to 500 mm, and the path overlap rate ranges from 10% to 20%. The path row spacing in the relatively low-stress areas of the side walls and arch feet ranges from 500 to 800 mm, and the path overlap rate ranges from 5% to 15%. The stress streamline path generation includes principal stress direction field interpolation, seed point arrangement, streamline integration, robotic arm accessibility constraints, minimum radius of curvature constraints, spray distance constraints, minimum spacing constraints between adjacent paths, as well as path smoothing and resampling.

4. The path design and construction method according to claim 1, characterized in that, The optimized 3D printing path set also includes a spatially continuous spiral path, which satisfies the following: x(θ)=x c +R(θ)cos[θ+φ(z(θ))]、y(θ)=y c +R(θ)sin[θ+φ(z(θ))] and z(θ)=z o +kθ; Where x and y are the coordinates within the tunnel cross section, and z is the tunnel axial coordinate. c and y c Let θ be the coordinates of the cross-section center, θ be the circumferential parameter, R(θ) be the inner contour radius of the lining, and φ(z) be the axial phase correction function. o Let R be the starting coordinate along the axis, and k be the axial advance per radian; k = R e tanα, R e The equivalent radius of the cross-section, the helix angle α, is determined by the material yield stress τ, plastic viscosity η, and injection velocity v. s The constructibility mapping function α=f(τ, η, v) s )Sure.

5. The path design and construction method according to claim 1, characterized in that, The material parameter table for the area includes at least the material formula number, yield stress, plastic viscosity, setting time, and early-age strength; the high-stress area of ​​the arch crown uses early-strength shotcrete with a setting time of 3 to 5 minutes, while the relatively low-stress areas of the side walls and arch feet use standard shotcrete with a setting time of 8 to 12 minutes. The material formula for each area is independently adjusted by the path design and construction central processing system.

6. The path design and construction method according to claim 1, characterized in that, In step S3, the injection parameter-forming index mapping model Qp=a0+a1v is used. n +a2p n +a3d n +a4q n Calculate the predicted value of the forming index; Among them, v n p n d n and q n These are the normalized values ​​for nozzle moving speed, injection pressure, spray distance, and injection flow rate, respectively, with a0 to a4 being calibration coefficients. Within the allowable parameter range, select a combination of spraying parameters with a forming index value Qp not lower than the forming index threshold and low material consumption, and write it into the robot control instructions; in areas with large path curvature, reduce the nozzle moving speed by 10% to 30% and reduce the spray flow rate by 5% to 15% relative to the reference value.

7. The path design and construction method according to claim 1, characterized in that, The layered progressive compensation strategy divides the forming error field into segments along the axial direction according to the longitudinal printing units and into blocks along the circumferential direction according to 30° sectors, and allocates the target compensation amount of each error block to the subsequent thickness printing layers that have not yet been printed. When the absolute value of the negative error exceeds 20% of the target thickness of the current thickness printing layer, a local reinforcement path is inserted in the next thickness printing layer; when the positive error exceeds 15% of the target thickness of the current thickness printing layer, the subsequent path is shifted inward along the design normal; 90% of the target compensation amount is allocated to no more than 3 subsequent thickness printing layers.

8. The path design and construction method according to claim 1, characterized in that, The branch and bound algorithm is used to optimize the execution order of the path segments to be printed. The sum of the idle travel time, attitude adjustment time, inter-layer waiting time and penalty value of exceeding the initial condensation time window between path segments is used as the cost. The upper bound is determined according to the current feasible order, and the lower bound is determined according to the minimum entry cost of the unvisited path segment and the necessary printing time. Branches with a lower bound not less than the current upper bound are pruned.

9. A central processing system for implementing the path design and construction method according to any one of claims 1 to 8, characterized in that, It includes a data fusion module, a mechanical analysis engine, an intelligent path planner, a process parameter optimizer, and a real-time compensation controller. The data fusion module is used to output a digital twin model of the tunnel lining, the mechanical analysis engine is used to output mechanical zoning results, the intelligent path planner is used to output an optimized 3D printing path set, the process parameter optimizer is used to output a regional material parameter table and a robot control instruction set, and the real-time compensation controller is used to output a compensation control instruction set based on the forming error field.

10. A jet-printed 3D-printed tunnel lining structure, characterized in that, The tunnel lining structure is constructed using the path design and construction method described in any one of claims 1 to 8. The tunnel lining structure has a stepped oblique overlapping interface along the axial direction formed by the retreat of the starting point of adjacent thickness printing layers, a gradient density printing path along the circumferential direction that matches the mechanical zoning results, and is formed by the superposition of multiple thickness printing layers that satisfy the early-age support criterion along the thickness direction.

Images