An intelligent vibration method and system for directional exhaust of precast concrete segments based on arc migration vibration field

Abstract

This invention discloses a method and system for directional air venting and intelligent vibration compaction of precast concrete segments based on an arc-shaped migration vibration field. The method includes: generating a smooth curve that conforms to the curvature of the mold's inner wall based on mold data as the target guiding path for air bubble migration and discharge; obtaining the spatial location of a narrow, confined area inside the mold; arranging control nodes based on the overlap between the target guiding path and the confined area; determining the initial start time and initial vibration frequency of the vibrator corresponding to each node based on the parameters of each control node; obtaining the spatial distribution of vibration energy generated by each vibrator, determining energy scattering characteristics, and identifying the coordinates of the energy accumulation center; correcting the initial start time and initial vibration frequency based on the deviation direction and distance of the energy accumulation center coordinates to obtain the target start time and target vibration frequency of each vibrator; and controlling the coordinated operation of each vibrator based on the target start time and target vibration frequency to generate a directional driving force field in the arc apex region.

Description

Technical Field

[0001] This invention belongs to the field of intelligent manufacturing and vibration compaction technology of precast concrete components, and particularly relates to a method and system for directional air venting and intelligent vibration compaction of precast concrete segments based on arc-shaped migration vibration field. Background Technology

[0002] Precast concrete segments, as key load-bearing components of major infrastructure projects such as underground tunnels and integrated utility tunnels, directly determine the permeability grade, load-bearing capacity, and service life of the engineering structure through their concrete density and appearance quality. In the segment production process, vibration compaction is a core step in eliminating air bubbles within the concrete and ensuring uniform aggregate distribution, thus having a decisive impact on the finished product qualification rate. Currently, precast plants generally adopt a batch control mode where multiple vibrators are started and stopped simultaneously or in a preset zone sequence. Although some advanced production lines have introduced vibration sensors for feedback adjustment, the overall approach still primarily relies on fixed-cycle or threshold-triggered strategies. These methods can maintain basic compaction effects when dealing with planar or shallow-ribbed components, and their simple control logic and ease of engineering implementation mean they are still widely used in the industry.

[0003] However, when the aforementioned traditional vibration control strategies are directly applied to curved segment molds, the large curvature of the mold causes complex refraction, reflection, and interference of vibration waves at the curved boundaries and variable cross-sections, resulting in a significant disordered scattering of vibration energy within the mold cavity. Operators cannot predict the specific location and direction of energy concentration, and can only rely on experience to extend the vibration time for a conservative compaction. This is not only energy-intensive and inefficient, but also fails to create a directional driving force to guide the orderly discharge of air bubbles. In fact, after the segment concrete is poured, air bubbles should rise along the curved surface to the top free surface under buoyancy. However, if the vibration energy is randomly distributed, the bubble's path will be frequently interrupted. Particularly problematic are the narrow spaces such as the segment's apex, the corners of the waterstop groove, the bottom of embedded parts, and areas with dense reinforcement. Once air bubbles enter these narrow spaces, they are easily blocked by mechanical forces or captured by viscous resistance. Without targeted vibration guidance, these air bubbles will permanently solidify inside the concrete, forming deep pores or surface pitting, severely weakening the segment's impermeability and structural continuity. Even more challenging is that existing control systems cannot plan the bubble movement path based on the geometric features of the mold arc, nor can they dynamically correct the vibration field shift caused by energy scattering. Furthermore, they lack the ability to form a focused driving force field with the same frequency and phase in the arc apex region. This results in significant quality fluctuations among different segments in the same batch, greatly restricting the upgrading of precast segment production towards higher reliability and consistency.

[0004] Therefore, this invention proposes a method and system for directional air venting and intelligent vibration compaction of precast concrete segments based on an arc-directional migration vibration field. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a precast concrete segment directional air venting intelligent vibration method and system based on an arc-directional migration vibration field, thereby resolving the problems existing in the prior art.

[0006] To achieve the above objectives, this invention provides a method for directional air venting and intelligent vibration compaction of precast concrete segments based on an arc-directional migration vibration field, comprising: A smooth curve that conforms to the curvature of the inner wall of the mold is generated based on the mold data and serves as the target guiding path for bubble migration and discharge. Obtain the spatial location of the narrow restricted area inside the mold, and arrange control nodes based on the overlapping relationship between the target guide path and the restricted area; The initial start time and initial vibration frequency of the vibrator corresponding to each node are determined based on the parameters of each control node. Obtain the spatial distribution of vibration energy generated by each vibrator, determine the energy scattering characteristics, and identify the coordinates of the energy accumulation center. Based on the deviation direction and deviation distance of the energy concentration center coordinates, the initial opening time and initial vibration frequency are corrected to obtain the target opening time and target vibration frequency of each vibrator. Based on the target start time and target vibration frequency, the vibrators are controlled to operate in a coordinated manner, generating a directional driving force field in the arc apex region.

[0007] Optionally, the process of generating a smooth curve that conforms to the curvature of the mold inner wall based on mold data as the target guiding path for bubble migration and discharge includes: Obtain the three-dimensional structural data and concrete pouring sequence data of the arc-shaped mold, and construct the initial connection line with the bottom pouring start point of the mold as the starting point and the top vent as the ending point; Based on the three-dimensional structural data, the curvature features and arc tangent direction of the inner wall of the mold are extracted to generate a geometric fit curve; A physical constraint model is established based on the pouring sequence data and concrete rheological parameters, and the geometric fit curve is corrected to obtain the physical correction path. The geometric fit curve and the physical correction path are weighted and fused, and then smoothed to obtain the smooth curve.

[0008] Optionally, the expression for the weighted fusion of the geometric fit curve and the physical correction path is: ; In the formula, , To preset the curvature threshold, For geometric fit curves, For physical correction path, This refers to the local curvature of the inner wall of the mold.

[0009] Optionally, the process of arranging control nodes based on the overlapping relationship between the target guidance path and the restricted area includes: Obtain the spatial coordinates of the narrow restricted area inside the mold, compare the spatial coordinates with the target guide path, and extract the intersection line segments of the target guide path that enter each restricted area; Based on the starting and ending arc length positions of each intersecting line segment on the target guidance path, the target guidance path is segmented along the arc length direction to obtain multiple independent path segments with regional attribute labels; From each independent path segment, select the path segments located in the arc top area, water stop groove area, bottom area of ​​embedded part and dense reinforcement area, and arrange control nodes at equal intervals on these path segments according to the preset densification spacing. Arrange control nodes at regular intervals on the remaining path segments to obtain the initial control node set. Extract the target nodes located in the arc apex region from the initial control node set, calculate the rheological resistance based on the rheological parameters of the concrete at each target node, determine the vibration attenuation coefficient based on the rheological resistance, and perform power compensation on the target nodes that exceed the threshold based on the vibration attenuation coefficient to obtain the compensated target nodes. The compensated target node and the uncompensated regular node are merged in order of arc length to obtain multiple vibration control nodes distributed along the target guidance path.

[0010] Optionally, the expression for power compensation of target nodes exceeding the threshold based on the vibration attenuation coefficient is as follows: ; In the formula, This is the reference operating power of the vibrator. m As a compensation index, To compensate for the power value, d The vibration attenuation coefficient is... This is the threshold value for the vibration attenuation coefficient.

[0011] Optionally, the process of determining the initial start time and initial vibration frequency of the vibrator corresponding to each node based on the parameters of each control node includes: Obtain the spatial coordinates, local radius of curvature, concrete slump, and node sequence number of each vibration control node; A spatial curvature tensor is constructed based on the spatial coordinates and local curvature radius of each node, and the slump attenuation value is calculated based on the node's arrangement number, cumulative path distance, and concrete slump. The spatial curvature tensor is expanded and then concatenated with the collapse attenuation value to form a fused feature vector. The fused feature vector is predicted using a random forest regression model to obtain the initial start time and initial vibration frequency of the vibrator corresponding to each node.

[0012] Optionally, the process of obtaining the spatial distribution of vibration energy generated by each vibrator, determining the energy scattering characteristics, and identifying the coordinates of the energy accumulation center includes: The vibration energy sampling values ​​of the inner region of the mold are obtained when each vibrator is working at the initial opening time and initial vibration frequency, and the vibration energy spatial distribution matrix is ​​constructed. The coordinates of the energy peak points are extracted from the vibration energy spatial distribution matrix at each time step and time-series fitting is performed to obtain the time-series motion trajectory of the energy peak points; The presence of energy scattering characteristics is determined based on the maximum deviation distance between the temporal motion trajectory of the energy peak point and the target guidance path. When energy scattering characteristics are determined, the spatial coordinates of each vibrator are used as the initial cluster center to perform clustering on the vibration energy spatial distribution matrix to obtain the coordinates of the energy aggregation center.

[0013] Optionally, the process of correcting the initial start-up time and initial vibration frequency based on the deviation direction and deviation distance of the energy accumulation center coordinates includes: The deviation direction is determined based on the spatial orientation of the energy gathering center coordinates relative to the target guidance path, and the deviation distance is determined based on the minimum distance between the energy gathering center coordinates and the target guidance path; Based on the deviation distance, the difference scaling ratio is determined, and the initial opening time difference of adjacent vibrators is scaled according to the deviation direction, and the sum is used to obtain the corrected opening time of each vibrator. Based on the deviation distance, the arc apex compensation coefficient and arc trough compensation coefficient are determined respectively. The arc trough compensation coefficient is used to compensate for the initial vibration frequency, and the arc apex compensation coefficient is used to compensate for the initial vibration power, so as to obtain the target vibration frequency and target vibration power. The target start time is obtained by fine-tuning the corrected start time based on the power compensation amount and the deviation direction.

[0014] Optionally, the process of controlling the coordinated operation of each vibrator based on the target start time and target vibration frequency to generate a directional driving force field in the arc apex region includes: Each vibrator is activated based on its target start time and target vibration frequency. The vibration waves generated by each vibrator are superimposed to form a strong vibration zone that moves along the target guidance path. The movement trajectory of the strong vibration area is monitored. When the movement trajectory enters the arc apex area, the percussion device is controlled to stay for a preset duration. Within the duration, the phase difference of each percussion device is adjusted to zero and switched to the same frequency and phase operation mode. The percussion devices are synchronously superimposed to form a directional driving force field.

[0015] The present invention also provides a precast concrete segment directional venting intelligent vibration compaction system based on an arc-directional migration vibration field, for implementing the method described above, the system comprising: The target guidance path generation module is used to generate a smooth curve that fits the curvature of the inner wall of the mold based on the mold data, which serves as the target guidance path for bubble migration and discharge. The control node arrangement module is used to obtain the spatial position of the narrow restricted area inside the mold and arrange the control nodes based on the overlapping relationship between the target guide path and the restricted area; The initial parameter analysis module is used to determine the initial start time and initial vibration frequency of the vibrator corresponding to each node based on the parameters of each control node. The energy distribution analysis module is used to obtain the spatial distribution of vibration energy generated by each vibrator, determine the energy scattering characteristics, and identify the coordinates of the energy accumulation center. The parameter correction module is used to correct the initial start time and initial vibration frequency based on the deviation direction and deviation distance of the coordinates of the energy gathering center, so as to obtain the target start time and target vibration frequency of each vibrator. The collaborative operation control module is used to control the collaborative operation of each vibrator based on the target start time and target vibration frequency, and to generate a directional driving force field in the arc apex region.

[0016] Compared with the prior art, the present invention has the following advantages and technical effects: This invention provides a smart vibration method for directional air venting of precast concrete segments based on an arc-shaped migration vibration field. By generating a smooth path that conforms to the curvature of the mold's inner wall as a target guide path for air bubble migration, it fundamentally solves the problem of disordered air bubble movement in traditional vibration methods, achieving effective guidance of air bubble venting direction. Furthermore, this invention differentiates the arrangement of control nodes based on the overlap relationship between the path and narrow areas, and accurately determines the initial operating parameters of the vibrator by combining the parameters of each node, enabling the vibration energy to be distributed specifically according to the mold's geometry and the concrete state. Further, this invention autonomously identifies and corrects field offsets caused by energy scattering by sensing the spatial distribution of vibration energy in real time, ensuring that the vibration energy is always focused on the target guide path. Finally, this invention forms a directional driving force field of the same frequency and phase in the arc apex region, effectively solving the problem of air bubbles being easily captured and solidified in the arc apex, water-stop groove, and densely reinforced areas, significantly improving the density and quality consistency of the segment concrete. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of a method according to an embodiment of the present invention; Figure 2 This is a system structure diagram of an embodiment of the present invention. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0020] Example 1 like Figure 1 As shown, this embodiment provides a method for directional air venting and intelligent vibration compaction of precast concrete segments based on an arc-directional migration vibration field, including the following steps: Step S101: Obtain the three-dimensional structural data of the arc mold and the concrete pouring sequence data. Based on the three-dimensional structural data and the concrete pouring sequence data, take the concrete pouring start point at the bottom of the mold as the starting point and the vent at the top of the mold as the ending point, and generate a smooth curve with curvature that fits the inner wall of the mold along the arc tangent. This curve is then determined as the target guiding path for the migration and discharge of air bubbles.

[0021] Furthermore, the three-dimensional structural data of the arc-shaped mold and the concrete pouring sequence data are first obtained. Based on this, the starting point of concrete pouring at the bottom of the mold is taken as the starting point of the bubble migration trajectory, and the vent at the top of the mold is taken as the ending point. After constructing the initial connection, the present invention innovatively introduces a dual-branch parallel computing framework: the first branch generates a geometric fit curve based on the curvature of the inner wall of the mold and the arc tangent, and the second branch constructs a physical constraint mathematical model driven by the pouring sequence, local slope and concrete rheological parameters. The two output the initial candidate trajectory through an adaptive weight fusion layer, and then smooth it through cubic spline interpolation. Finally, the unique target guiding path is determined and output based on the curvature radius threshold.

[0022] As a specific implementation of this embodiment, a lightweight graph attention network is constructed to adaptively extract the spatial distribution relationship between key curvature features and arc-direction tangents of the mold's inner wall from 3D structural data. The network input is the coordinate matrix of the vertices of the triangular mesh of the mold's inner wall. and normal vector matrix Where N is the total number of grid vertices. After three layers of graph attention convolution (each layer contains 4 attention heads, with output dimensions of 64, 128, and 64 respectively), each layer uses a multi-head attention mechanism to calculate the adjacency weights between vertices. Then, after LeakyReLU activation and softmax normalization, a weighted adjacency matrix is ​​obtained. ; and then utilize The vertex features are aggregated and updated, and the encoded curvature feature tensor is output. At the same time, the concrete pouring sequence data (time step T, corresponding pouring height for each step) will be included. Input a one-dimensional temporal convolutional network to extract the temporal dimension features of the pouring process. Finally, The eigenvectors corresponding to the pouring start point and the vent are in the middle. End feature concatenation is fed into a fully connected regression layer, outputting the spatial direction vector of the arc tangent. , serving as the initial directional reference for the geometric fit curve.

[0023] After generating the geometric fit curve, a physical constraint model based on fluid dynamics and bubble motion dynamics is introduced to verify and fine-tune the fit curve, making it more consistent with actual bubble migration patterns. The model expression is: ; in, This is the sequence of discrete point coordinates on the aforementioned geometric fit curve. P The final path point sequence to be optimized. L Let be the total arc length of the path. s For arc length parameters; This refers to the density of concrete (unit: kg / m³). g It is the acceleration due to gravity. The angle between the tangent at the path point and the horizontal plane (unit: rad) represents the buoyancy driving term; The plastic viscosity of concrete (unit: Pa·s). ξ is the path curvature vector, representing the viscous drag term; ξ is the friction coefficient between the bubble and the inner wall. Let be the local curvature of the inner wall of the mold (unit: 1 / m), and v(s) be the rising velocity of the air bubble relative to the concrete (unit: m / s). Calculate (r is the equivalent radius of the bubble, (This refers to air density). These are weighting coefficients used to balance geometric fit and physical plausibility. The model is discretized using the finite difference method and then iteratively solved using gradient descent. Output the initial candidate path after physical constraint correction.

[0024] Adaptive weight fusion layer: fits geometry to curves With physical correction path The input fusion layer has the following fusion rules: ; in, , The preset curvature threshold (unit: 1 / m, default value is 0.02). This is the scaling factor (default value is 50). When the local curvature is greater than... hour, When the curvature approaches 1, the weighting favors the geometrically conforming curve, ensuring the path strictly adheres to the mold's inner wall in sharp bends; when the curvature is small, the weighting favors the physically correcting path, allowing the bubble to fully utilize buoyancy and rise naturally in gentler areas. (The fused bubble...) After global smoothing using a cubic spline interpolation algorithm, a continuously second-differentiable smooth curve is obtained. .

[0025] Curvature radius threshold determination: Multiple sampling points are taken at equal intervals along the entire smooth curve. At each sampling point, the radius of the inscribed circle of the small arc segment containing that point is calculated, i.e., the local curvature radius. This value reflects the severity of the curve's curvature at that location—a larger value indicates a gentler path, while a smaller value indicates a sharper bend. Subsequently, the minimum curvature radius is selected from all sampling points as the weakest link in the entire path. A dynamically adjustable curvature radius threshold is preset. This threshold is not a fixed constant but is comprehensively set based on the actual size of the mold and the maximum particle size of the aggregate in the currently poured concrete: the larger the mold size or the coarser the aggregate particle size, the higher the threshold is, and vice versa, to ensure that the threshold matches the actual working conditions. The aforementioned minimum curvature radius is compared with this dynamic threshold. If the minimum curvature radius is greater than or equal to the threshold, it indicates that the curvature of the entire path is within an acceptable range. Under the action of buoyancy, the air bubbles can migrate continuously upward along the path without detaching from the inner wall of the mold due to centrifugal effect or causing flow dead zones due to sharp bends. At this point, the smooth curve is officially determined as the target guiding path for air bubble discharge and is output. Conversely, if the minimum radius of curvature is less than the threshold, the path is determined to be too steep locally. In this case, the result is not output directly, but is fed back to the aforementioned fusion layer. By readjusting the weight ratio between the geometric fitting curve and the physical correction path, the optimization direction is guided towards the goal of increasing the minimum radius of curvature until the threshold condition is met, thereby ensuring that the final output target guidance path meets the engineering application requirements in terms of both geometric construction and physical feasibility.

[0026] Step S102: Obtain the spatial coordinates of the narrow restricted area inside the mold. Based on the target guide path and the spatial coordinates of the narrow restricted area, segment the target guide path. Densify the arrangement of control nodes in the path segments corresponding to the arc top area, water stop groove area, bottom area of ​​embedded part, and dense reinforcement area. Perform vibration power compensation preprocessing on the nodes corresponding to the arc top area to obtain multiple vibration control nodes distributed on the target guide path.

[0027] Among them, the vibration control node is the logical carrier unit of the vibration control command, and each node corresponds to an independently controlled vibrator; the actual installation position of the vibrator is the physical coordinate corresponding to the projection of each node onto the outer wall of the mold along the normal direction.

[0028] When the number of control nodes exceeds the number of vibrators, each vibrator is assigned a nearest node based on the principle of proximity by arc length. The remaining nodes obtain their command values ​​through linear interpolation of the control parameters of adjacent vibrators. When the number of vibrators exceeds the number of control nodes, a Thiessen polygon partition is established with each node as the center. All vibrators within the same partition share the control parameters of that node, achieving one-to-many group control. This mapping relationship is calculated once before the vibration operation starts, generating a node-vibrator lookup table matrix. Subsequent closed-loop corrections only update the parameter values ​​in the table, without needing to rebuild the mapping structure.

[0029] Furthermore, using the target guidance path generated in step S101 and the spatial coordinates of the narrow confined area inside the mold as input, the path is physically segmented by performing spatial intersection analysis between the target guidance path and the narrow area. Differentiated node densification is implemented for key path segments such as the arc apex, water stop groove, bottom of embedded parts, and areas with dense reinforcement. At the same time, for nodes in the arc apex area, a quantitative vibration power compensation mechanism based on concrete rheological parameters is introduced. By constructing a nonlinear mapping mathematical model between rheological resistance and vibration attenuation coefficient, power pre-compensation is performed on nodes whose attenuation exceeds the threshold. Finally, multiple vibration control nodes distributed along the target guidance path are output.

[0030] As a specific implementation method of this embodiment, firstly, the spatial coordinate set of all narrow and restricted areas inside the mold is obtained, including areas with dense steel reinforcement skeletons, installation positions of water-stop grooves, bottom areas of embedded parts, and chamfered areas. The spatial coordinate data of these areas are then compared with the target guide path. A spatial geometric intersection algorithm is used to calculate the coordinate points of the target guide path that enter or exit the boundaries of each narrow and restricted area, and the intersection lines where the target guide path coincides with each restricted area are extracted.

[0031] Based on the start and end arc lengths of each extracted intersection segment, the entire target guidance path is segmented along the arc length direction. Each intersection segment is treated as an independent restricted path segment, and its area type is marked (arc apex, waterstop groove, bottom of embedded part, densely reinforced area, or regular area); the path portion not covered by any narrow area is treated as a regular path segment. Through this segmentation process, the complete target guidance path is divided into multiple independent path segments that are connected end-to-end and have area attribute labels.

[0032] From the aforementioned independent path segments, path segments located in the arc apex area, water-stop groove area, bottom area of ​​embedded parts, and densely reinforced areas are selected and marked as specific path segments. For these specific path segments, control nodes are arranged equidistantly according to a preset densified node spacing (default 15 mm), making the node distribution density in critical areas significantly higher than in regular areas; while for regular path segments, a larger regular node spacing (default 30 mm) is used. After this step, an initial set of control nodes distributed along the entire target guidance path is obtained.

[0033] Extract all nodes located in the apex region from the initial control nodes and designate them as target nodes for vibration power compensation assessment. For each target node, obtain the rheological parameters of the surrounding local concrete, including yield stress, plastic viscosity, and aggregate volume fraction. Calculate the rheological resistance at the target node based on the Bingham fluid constitutive model. ; In the formula, For yield stress, Plastic viscosity, This represents the local shear rate.

[0034] To establish a quantitative relationship between rheological drag and vibration damping coefficient, a nonlinear mapping mathematical model in the form of a power function is constructed as follows: ; d The vibration attenuation coefficient is a value that characterizes the degree of energy loss during the transmission of vibration energy from the vibrator to the target node. A larger value indicates more severe attenuation. k This is a scaling factor (default value is 0.02), obtained by fitting historical experimental data, which controls the overall scaling of the mapping curve; m The exponent (default 0.6) controls the difference in the growth rate of the curve between the low-resistance and high-resistance regions, enabling the model to sensitively reflect the impact of changes in rheological drag within a small numerical range on the damping coefficient. Substituting the calculated rheological drag into the model outputs the vibration damping coefficient corresponding to the target node.

[0035] Set vibration attenuation coefficient threshold (The default value is 0.65. This value is set based on engineering experience; exceeding this value indicates that energy loss has significantly affected the vibration effect.) The determined... d Compare with this threshold: If d ≤ If the node does not require compensation, the original node settings should be retained; otherwise... d > If so, the node is determined to require power compensation, and the compensation power value is... Calculate using the following formula: ; in, This is the reference operating power of the vibrator. m The compensation index controls the non-linear growth rate of the compensation amount as the attenuation coefficient exceeds the threshold, preventing overcompensation and energy waste when the threshold is slightly exceeded. The calculated... The vibration power compensation preprocessing of the target node is completed by algebraically superimposing the target node's base power setting value to obtain the compensated target node.

[0036] The compensated target nodes, after power compensation processing, are merged with the uncompensated regular nodes in the initial control nodes. These nodes are then rearranged according to their arc length on the target guidance path to form a complete vibration control node sequence. Each node records its three-dimensional spatial coordinates, its region type label, and its final compensated power setting value. This sequence is ultimately identified as multiple vibration control nodes distributed along the target guidance path, serving as the standard input data for the initial parameter analysis of the vibrator in step S103.

[0037] Step S103: A machine learning model is used to extract and fuse features of the spatial coordinates, local radius of curvature, concrete slump, and node sequence number of each vibration control node to obtain the initial start time and initial vibration frequency of the vibrator corresponding to each vibration control node.

[0038] Furthermore, using the spatial coordinates, local radius of curvature, concrete slump, and node sequence number of each vibration control node output in step S102 as input, a spatial curvature tensor is constructed to quantify the degree of local geometric distortion of the mold. The slump attenuation value is calculated by combining the path distance and slump to reflect the loss of concrete flowability along the path. These two are then concatenated into a fused feature vector. That is, this invention introduces a pre-trained random forest model to perform nonlinear regression prediction on this fused feature, outputting the initial start time and initial vibration frequency of the vibrator corresponding to each node. Furthermore, a frequency compensation mathematical model based on the delay in start time is constructed to adaptively correct the vibration frequency of lagging nodes, ultimately outputting the initial start time and corrected final vibration frequency of each vibrator.

[0039] As a specific implementation method of this embodiment, the construction of the spatial curvature tensor is as follows: The three-dimensional spatial coordinates of each vibration control node and the local radius of curvature of the target guidance path at that node are obtained. A local orthogonal coordinate system is established with the node as the center, along the tangent direction, normal direction, and binormal direction of the path. The central difference method is used to calculate the second-order rates of change of the node's spatial coordinates in the horizontal, vertical, and longitudinal directions, i.e., the rates of change of curvature in each direction. The local radius of curvature is used as the reference scalar component of the tensor and combined with the rates of change of curvature in the three directions to construct a three-dimensional spatial curvature tensor. This tensor is a symmetric matrix, where the diagonal elements are all local radii of curvature, and the off-diagonal elements are the rates of change of curvature coupling between each direction.

[0040] Calculation of slump attenuation: Obtain the initial slump of the concrete corresponding to the current vibration control node, as well as the node's sequence number on the target guide path. Based on the node sequence number and the preset arc length distance between adjacent nodes, calculate the cumulative path distance between the node and the concrete pouring start point. Considering that the slump gradually decreases as the concrete flows along the path within the mold due to factors such as segregation, aggregate friction, and moisture loss, an exponential attenuation model is constructed: using the initial slump as the baseline value and the cumulative path distance as the attenuation independent variable, the attenuation rate is controlled by an attenuation coefficient to calculate the slump attenuation value at the node. This attenuation coefficient is calibrated using field measurement data and reflects the rate of loss of concrete fluidity along the path; the larger the cumulative path distance, the closer the slump attenuation value is to the initial slump, indicating that the slump loss is approaching saturation.

[0041] Construction of the fused feature vector: The constructed three-dimensional spatial curvature tensor is expanded into a nine-dimensional vector along the row direction. The calculated collapse attenuation value is then used as the tenth feature dimension and concatenated with the expanded tensor vector to form a ten-dimensional fused feature vector. This vector integrates dual information from the mold geometry at the node (including local curvature radius and curvature change rate in each direction) and the concrete flow state (degree of collapse attenuation).

[0042] Regression prediction using the random forest model: A pre-trained random forest regression model is used to predict the fused feature vector. The random forest model consists of multiple decision trees, each splitting its nodes based on different random subsets of the feature vector, using the minimum mean square error criterion. The training data comes from a sample set of measured feature parameters and corresponding optimal vibration parameters for each node under historical pouring conditions. After inputting the fused feature vector into the model, each decision tree independently outputs a set of predicted values. Finally, the arithmetic mean of all decision tree outputs is taken as the regression prediction result, yielding the initial start time and initial vibration frequency of the vibrator corresponding to the vibration control node.

[0043] Frequency compensation correction based on start-up time delay: A preset time threshold is set, which is the latest allowed vibration start time within the initial setting time window of concrete. It is determined whether the predicted initial start time is greater than this threshold: if it is not greater, no compensation is needed, and the initial vibration frequency is the final vibration frequency; if it is greater, the delay difference between the initial start time and the threshold is calculated. A frequency compensation coefficient is calculated based on this delay difference. This coefficient is constructed using a power function, with the ratio of the delay difference to the threshold as the independent variable. The maximum compensation ratio is controlled by a compensation amplitude coefficient, and the nonlinear growth rate of the compensation amount with increasing delay time is controlled by a compensation growth exponent, thus preventing over-compensation for short delays or under-compensation for long delays. The frequency compensation coefficient is multiplied by the initial vibration frequency to obtain the final vibration frequency. This correction mechanism ensures that when vibration start-up is delayed, the vibration frequency is increased to compensate for the decrease in concrete fluidity caused by the delay, thereby ensuring the consistency of the air venting effect. Finally, the initial start time of each vibration control node and the corrected final vibration frequency are output together.

[0044] Step S104: Determine the vibration phase difference based on the difference between the initial start times of the vibrators corresponding to adjacent vibration control nodes and the initial vibration frequency; determine the moving speed of the vibration energy field along the target guidance path based on the ratio of the arc length between adjacent vibration control nodes to the vibration phase difference.

[0045] Furthermore, using the initial start time of the vibrator corresponding to adjacent vibration control nodes, the path distance between nodes, the concrete slump, and the initial vibration frequency as inputs, the vibration phase difference is determined by calculating the difference in the initial start time of the vibrator corresponding to adjacent vibration control nodes and the initial vibration frequency. The ratio of the arc length between nodes to the phase difference is then used as the base value for energy movement velocity. This invention introduces a medium parameter mapping model based on concrete slump. The slump determines the medium density and propagation impedance, and the attenuation coefficient is further calculated. Simultaneously, the vibration wavelength is derived by combining the initial vibration frequency and the base energy movement velocity. Based on this, the dispersion characteristics are extracted to perform nonlinear correction on the energy movement velocity, and finally, the corrected energy field movement velocity is output.

[0046] In a specific implementation of this embodiment, the initial start time of the vibrator corresponding to each of two adjacent vibration control nodes is obtained. The start time of the latter node is subtracted from the start time of the former node, and the difference between the two is calculated. Based on this difference and the initial vibration frequency, the vibration phase difference is determined through conversion. Simultaneously, the path arc length between two adjacent vibration control nodes is obtained. The path arc length is divided by the aforementioned phase difference value, and the quotient is the basic energy velocity, which characterizes the average propagation rate of the vibration energy field along the target guidance path in an ideal homogeneous medium. The concrete slump value corresponding to each vibration control node is obtained. According to the empirical mapping relationship in concrete materials science, the medium density and propagation impedance of the concrete at that node are determined by the slump value. Specifically, slump reflects the fluidity and density of concrete: the larger the slump, the higher the proportion of concrete paste and the slightly lower the medium density; the smaller the slump, the denser the aggregate packing and the correspondingly higher the medium density. Propagation impedance characterizes the resistance of the concrete medium to the propagation of vibration waves. Its value is related to density and elastic modulus, and is calculated from the slump through a pre-calibrated mapping table or empirical function.

[0047] Using the medium density and propagation impedance obtained in the previous step as input parameters, the attenuation coefficient at this node is calculated using an empirical model of concrete vibration wave attenuation. The attenuation coefficient quantitatively describes the rate of energy loss caused by factors such as viscous damping, internal friction, and aggregate interface scattering when the vibration wave propagates in the concrete medium. The larger the value, the more severe the attenuation of vibration energy per unit propagation distance.

[0048] The expression for calculating the attenuation coefficient is: ; In the formula, For viscous intrinsic decay, f The initial vibration frequency, The density of the concrete medium, c The propagation speed of the vibration wave in concrete is given by the propagation impedance Z and the density of the medium. The ratio relationship is calculated. The attenuation factor is the curvature constraint of the mold. l The curvature influence coefficient was determined through on-site vibration experiments, with a value range of 0.05 to 0.15, to control the contribution intensity of curvature factors to attenuation. This is the actual local radius of curvature at the vibration control node. As a reference radius of curvature, the standard radius of curvature given in the mold design is used. The scattering attenuation factor in areas with dense reinforcement. m The scattering coefficient of the reinforcing steel. The average diameter of the reinforcing bars in this area. This represents the density of reinforcing steel in the area where the node is located.

[0049] The initial vibration frequency of the vibrator corresponding to the vibration control node is obtained. Combined with the foundation energy movement velocity, and based on the fundamental physical relationship between wave velocity, frequency, and wavelength, the vibration wavelength at that node is calculated. On this basis, the attenuation coefficient and the calculated vibration wavelength are used as inputs, and a dispersion analysis algorithm is employed to extract dispersion characteristics. Dispersion characteristics reflect the phenomenon that different frequency components of the vibration wave propagate at different speeds in the viscoelastic concrete medium—high-frequency components propagate more slowly due to greater viscous damping, while low-frequency components propagate relatively faster.

[0050] A correction coefficient is constructed using the extracted dispersion characteristics to nonlinearly correct the base energy movement velocity. The specific correction logic is as follows: when the dispersion effect is significant, the single wave velocity represented by the base energy movement velocity cannot accurately describe the actual energy field migration rate; therefore, the velocity value needs to be adjusted according to the strength of the dispersion characteristics. The correction process uses a pre-established dispersion correction function, which takes the dispersion characteristic value as the independent variable and outputs the corresponding velocity correction ratio. Multiplying the base energy movement velocity by this correction ratio yields the corrected movement velocity. This corrected movement velocity fully reflects the influence of the concrete medium's dispersion effect on the propagation speed of the vibration energy field and is more physically reasonable and engineeringly applicable than the base energy movement velocity. Finally, the corrected movement velocity is determined as the speed at which the vibration energy field moves along the target guidance path.

[0051] Step S105: Obtain the spatial distribution data of vibration energy formed inside the mold when the multiple vibrators are working at the initial start time and the initial vibration frequency. If the maximum deviation distance between the motion trajectory of the energy peak point in the spatial distribution data and the target guide path in the time sequence exceeds the preset deviation allowable threshold, it is determined that there is energy scattering feature.

[0052] Vibration energy sampling values ​​of multiple vibrators at various coordinate positions within the mold's inner domain at the initial start time and initial vibration frequency are acquired to construct a vibration energy spatial distribution matrix. Energy peak point coordinates are extracted from the vibration energy spatial distribution matrix time-by-time, and time-series fitting is performed based on these coordinates to obtain the time-series motion trajectory of the energy peak points. Each trajectory point on the time-series motion trajectory of the energy peak points is acquired, and the vertical distance between each trajectory point and the target guide path is calculated. The maximum value is taken as the maximum deviation distance. If the maximum deviation distance exceeds a preset allowable deviation threshold, coordinate regions where the vibration energy sampling values ​​are below a preset lower energy limit are extracted from the vibration energy spatial distribution matrix and marked as energy-deficient regions. Based on these energy-deficient regions, it is determined that energy scattering characteristics exist within the mold's inner domain.

[0053] Furthermore, the vibration energy sampling values ​​formed at each coordinate position within the mold domain when each vibrator operates at its initial start-up time and initial vibration frequency are obtained, and a vibration energy spatial distribution matrix is ​​constructed. Specifically, the data is collected and reconstructed as follows: Several piezoelectric accelerometers are equidistantly installed along the arc length direction on the outer wall of the mold, with a measuring point spacing of no more than 200mm. Local densification is carried out in key locations such as the arc top area, arc bottom area, near the water stop groove, and side ribs, with a densification spacing of no more than 100mm. At least four measuring points are arranged in the arc top area to ensure that the spatial changes of vibration energy can be fully captured. The sensor range is selected according to the working frequency of the vibrator and the estimated vibration amplitude. Its acquisition frequency is set to more than 20 times the working frequency of the vibrator to ensure that the time resolution of the vibration signal waveform meets the analysis requirements.

[0054] All sensors synchronously acquire vibration acceleration signals at various measuring points on the outer wall of the mold. The acquisition time is the complete vibration cycle from the start of the first vibrator to the stop of all vibrators. After filtering and noise reduction preprocessing, the root mean square value of acceleration within each time window is calculated for each channel, which is used as the vibration energy characterization value of that measuring point within that time window.

[0055] Using the three-dimensional spatial coordinates of each measuring point as known data nodes and the corresponding measured vibration energy characterization value as the node value, the radial basis function interpolation method is employed to perform gridded interpolation calculations on the inner space of the mold, obtaining the estimated vibration energy values ​​at each discrete grid node within the mold's inner region, thereby constructing a vibration energy spatial distribution matrix covering the entire inner region of the mold. The radial basis function adopts a multiple quadratic function form, and its shape parameters are adaptively determined based on the spacing between measuring points to balance interpolation accuracy and smoothness. After interpolation, a portion of the measuring point data is reserved as a validation set for cross-validation to ensure that the interpolation reconstruction error is controlled within the allowable range for engineering applications.

[0056] Based on the constructed vibration energy spatial distribution matrix, the energy values ​​in the matrix are scanned time-by-time with a time step of 10 milliseconds. A local maximum search algorithm is used to locate the spatial coordinates of the largest energy value within each time slice as the energy peak point coordinates. The energy peak point coordinates of 20 consecutive time slices are collected, and these coordinate sequences are input into a cubic spline interpolation algorithm for time-series fitting, outputting a smooth three-dimensional spatial curve, i.e., the temporal trajectory of the energy peak point.

[0057] It should be noted that discrete trajectory points on the time-series motion trajectory of the energy peak point are obtained and imported into a pre-defined target guidance path coordinate set. A point-to-space curve shortest distance algorithm is used to calculate the vertical distance from each trajectory point to the target guidance path. All calculated vertical distance values ​​are iterated through, and the largest value is selected as the maximum deviation distance.

[0058] In one possible implementation, a deviation allowable threshold is set at 150 mm. When the calculated maximum deviation reaches 180 mm, exceeding this threshold, the energy deficit region extraction process is triggered. A preset lower energy limit of 200 joules is set, and the vibration energy spatial distribution matrix is ​​traversed to extract continuous coordinate grids with energy sampling values ​​below 200 joules, marking them as energy deficit regions. Based on the spatial distribution pattern of these energy deficit regions, it is determined that the vibration wave within the mold's inner region exhibits energy scattering characteristics during propagation.

[0059] Step S106: If the energy scattering feature is determined to exist, the spatial coordinates of each vibration control node are used as the initial cluster center, and a clustering analysis algorithm is used to group and merge the vibration energy spatial distribution data to obtain the coordinates of the energy aggregation center.

[0060] Obtain the vibration energy spatial distribution matrix corresponding to the energy scattering characteristics, extract the spatial coordinates of each vibrator within the mold's inner domain, and use each spatial coordinate as an initial cluster center. Calculate the Euclidean distance between each coordinate position in the vibration energy spatial distribution matrix and each initial cluster center using a clustering analysis algorithm. Divide each coordinate position into the group corresponding to the initial cluster center with the smallest Euclidean distance, obtaining an initial grouped merge set. Calculate the mean coordinate value of each coordinate position within each group in the initial grouped merge set, and obtain the coordinate offset between the mean coordinate value and the corresponding initial cluster center. If the coordinate offset is less than or equal to a preset offset threshold, then the mean coordinate value is determined as a convergent cluster center. Extract the spatial location information corresponding to each convergent cluster center, and determine the coordinates of the energy aggregation center within the mold's inner domain based on the spatial location information.

[0061] As a specific implementation of this embodiment, the vibration energy spatial distribution matrix corresponding to the determination of energy scattering characteristics in step S105 is obtained. This matrix records the vibration energy sampling values ​​at each discrete coordinate position within the mold's inner domain. Simultaneously, the three-dimensional spatial coordinates of each vibrator actually installed inside the mold are extracted in the mold's inner domain spatial coordinate system. The spatial coordinates of each vibrator are directly used as the initial cluster centers for the clustering analysis algorithm, without any randomization or human intervention. The physical basis of this initialization strategy is that the vibrator is the direct source of vibration energy emission, and its installation position naturally determines the spatial distribution pattern of energy input. Using the actual vibrator position as the starting point for clustering ensures that the algorithm's initial cluster centers correspond one-to-one with the physical excitation sources.

[0062] Clustering analysis algorithms are used to determine the spatial affiliation of each coordinate position in the vibration energy spatial distribution matrix. Specifically, the algorithm iterates through all valid coordinate positions in the matrix, calculating the Euclidean distance between each position and each initial cluster center. The Euclidean distance is calculated based on the geometric distance formula for three-dimensional spatial coordinates, which is the square root of the sum of the squares of the coordinate differences between two points in the horizontal, vertical, and longitudinal directions. According to the principle of minimizing distance, each coordinate position is assigned to the group corresponding to the initial cluster center with the shortest Euclidean distance. Through this partitioning process, all discrete coordinate positions within the entire mold area are grouped into multiple initial group sets, with each group corresponding to the energy radiation influence zone of a vibrator.

[0063] For each initial grouped set obtained in the previous step, calculate the mean coordinates of all coordinate positions within each group. This involves taking the arithmetic mean of the horizontal, vertical, and lateral coordinates of all coordinate positions within the group to obtain the average spatial position of that group. Compare the calculated mean coordinates with the initial cluster centers corresponding to that group, calculating the spatial distance between them; this distance is the coordinate offset. Set a preset offset threshold to determine if the cluster centers are stable. Compare the calculated coordinate offsets with the preset offset threshold: if the coordinate offset is less than or equal to the threshold, the cluster centers of that group have met the convergence condition, and the mean coordinate is determined as the convergent cluster center of that group; if the coordinate offset is greater than the threshold, the mean coordinate is used as the new cluster center, and the spatial assignment and mean calculation are re-executed until the coordinate offsets of all groups meet the convergence condition.

[0064] Once all groups meet the convergence criteria, the three-dimensional spatial location information corresponding to each convergent cluster center is extracted. However, the convergent cluster centers only reflect the spatial geometric center of the coordinate points within each group, without considering the actual vibration energy magnitude at each coordinate position. Therefore, their physical meaning is not equivalent to the actual convergence focus of vibration energy. To address this, energy-weighted refinement is performed on each group: all coordinate positions within the group and their corresponding vibration energy sampling values ​​are obtained. Using the vibration energy value of each coordinate position as a weight, a weighted average is calculated for all coordinate positions. Positions with higher energy values ​​contribute more to the average result, and the coordinates of the energy-weighted center for that group are calculated. Subsequently, the convergent cluster centers obtained from the cluster analysis are weighted and fused with the energy-weighted centers according to a preset fusion weight. The fusion weight coefficient ranges from 0.3 to 0.5, with a default value of 0.4, determined based on field calibration experiments, ensuring that the fusion result balances spatial uniformity and energy concentration. The resulting coordinates are used as the final energy aggregation center for that group. The final energy aggregation centers of all groups are summarized, and the coordinates of the energy aggregation center within the mold's inner domain are determined accordingly.

[0065] Step S107: Obtain the deviation direction and deviation distance of the energy concentration center coordinates relative to the target guide path. Calculate the difference scaling ratio using the deviation distance, and perform differential scaling correction on the initial opening time difference of adjacent vibrators according to the deviation direction. Simultaneously, use the ratio of the deviation distance to the characteristic distances of the arc bottom and arc top regions as differential compensation coefficients to perform asymmetric positive compensation on the vibration frequency of the arc bottom region and the vibration power of the arc top region, respectively. After power compensation, based on the combined effect of power increment and deviation direction, use an analytical fine-tuning function to finely adjust the corrected opening time, and finally output the target opening time and target vibration frequency of each vibrator.

[0066] As a specific implementation of this embodiment, the mapping relationship between the deviation distance and the difference scaling ratio is constructed as follows: ; in, e Scaling ratio for the difference This represents the minimum Euclidean deviation distance from the coordinates of the energy accumulation center to the target guidance path. The mold feature reference dimension is taken as one-tenth of the mold's inner diameter. This mapping relationship exhibits a monotonically increasing characteristic: the larger the deviation distance, the closer the scaling ratio is to 1, and the larger the adjustment range; when the deviation distance approaches zero, the scaling ratio approaches zero, producing almost no adjustment. After obtaining the initial start-up time difference of the vibrators corresponding to adjacent vibration control nodes, the difference is differentially corrected according to the deviation direction: if the deviation direction is above the target guide path, the initial difference is multiplied by 1 and the scaling ratio is subtracted to reduce the difference, thereby accelerating the start-up rhythm of adjacent vibrators and shortening the time interval between the vibration wave front reaching the arc apex; if the deviation direction is below the target guide path, the initial difference is multiplied by 1 and the scaling ratio is added to increase the difference, thereby slowing down the start-up rhythm and allowing more sufficient energy accumulation time for the arc apex region. Using the initial start-up time of the first vibrator as the time reference, the corrected adjacent differences are sequentially added to the reference time to obtain the corrected start-up time sequence of each vibrator.

[0067] The minimum normal distance from the target guide path in the apex region of the mold to the outer wall of the mold, and the minimum normal distance from the target guide path in the apex region to the outer wall of the mold, are extracted. These two distances are directly measured from the 3D design model of the mold. Compensation coefficients for the apex and apex regions are calculated separately, with the sum of the deviation distance and the corresponding normal distance as the denominator and the deviation distance as the numerator. The apex region uses the apex normal distance, and the apex region uses the apex normal distance. The physical meaning of the two compensation coefficients is the relative proportion of the deviation distance at each region's characteristic scale. The larger the deviation distance or the smaller the region's characteristic distance, the closer the compensation coefficient is to 1, and the greater the compensation intensity. During compensation, the target vibration frequency in the apex region is obtained by multiplying the initial vibration frequency by 1 and adding the apex compensation coefficient, achieving a positive frequency increase to enhance the vibration energy injection density in the apex region and overcome the energy deficiency caused by deviation. The target vibration power in the apex region is obtained by multiplying the initial vibration power by 1 and adding the apex compensation coefficient, achieving a positive power enhancement to strengthen the exhaust driving force in the apex region and counteract the energy attenuation due to upward migration caused by deviation.

[0068] After completing power compensation in the arc apex region, the difference between the target vibration power and the initial vibration power is calculated as the power increment. Construct an analytical fine-tuning function with the power increment and the sign of the deviation direction as independent variables. .

[0069] in, For time fine-tuning; To fine-tune the proportional coefficient, the default value is 0.02, which is calibrated by a single power compensation experiment; The reference power increment normalization constant; s A value of +1 indicates a deviation from the target's guiding path above, while a value of -1 indicates a deviation below the target's guiding path. Final target activation time. Complete all corrections. The initial vibration power, The value represents the target vibration power in the apex region.

[0070] Step S108: Control multiple vibrators to operate in coordination according to the target start time and the target vibration frequency, generate a strong vibration area that moves along the target guide path, and if the strong vibration area moves to the arc apex area, stay for a preset dwell time, and make the vibration phase difference of each vibrator return to zero and switch to the same frequency and phase operation mode to generate a directional driving force field.

[0071] The target activation time and target vibration frequency are obtained. Based on these, multiple vibrators are controlled to operate collaboratively, generating a strong vibration zone that moves along the target guidance path. The movement trajectory coordinates of the strong vibration zone are extracted. If the movement trajectory coordinates enter the apex region, apex region dwell is triggered, and a preset dwell time is obtained. Within the dwell time, the vibration phase of multiple vibrators is zeroed to obtain a synchronous operation mode. A directional driving force field is generated based on this synchronous operation mode. The driving force field intensity is obtained and input into a pre-established energy mapping model for calculation to obtain the vibration energy distribution. Spatial node positions are extracted based on the vibration energy distribution. The flow velocity characteristics of the medium are calculated based on the spatial node positions. The rheological state of the medium is identified based on the flow velocity characteristics. The phase transition critical point is determined based on the rheological state of the medium and used as the solid-liquid separation boundary.

[0072] In one possible implementation, after obtaining the target start time and target vibration frequency, the controller sends execution commands to multiple vibrators distributed on the outside of the mold. Each vibrator starts according to the set timing and frequency, and the mechanical waves it generates interfere and superimpose in the medium inside the mold. The point where the wave crests meet generates a strong vibration zone that moves along the target guide path. The coordinates of the movement trajectory of this zone are extracted. When the coordinates enter the pre-marked arc apex region boundary, a dwell mechanism is triggered, and a preset dwell time of 15 seconds is obtained.

[0073] Specifically, during the aforementioned dwell time, vibration phase zeroing processing is performed on multiple vibrators. The real-time operating phase of each vibrator is collected and input into the phase-locked loop control algorithm to calculate the phase difference between each device and the reference phase. Based on this phase difference, a phase compensation signal is output to forcibly adjust the position of the eccentric block, bringing the phase difference between each vibrator close to 0 degrees, thereby achieving a synchronized operation mode. In this mode, the excitation forces of multiple devices are synchronously superimposed in the same direction, generating a directional driving force field perpendicular to the surface of the arc apex.

[0074] Specifically, each vibrator is installed along the normal direction of the mold's arc surface, and its excitation direction is directed towards the center of the mold's inner cavity, i.e., perpendicular to the tangent of the segment's arc surface. The vibrators employ high-frequency eccentric vibration motors, and the direction of their excitation force is determined by the rotational phase of the eccentric block. When multiple vibrators operate at the same frequency and phase, the excitation force vectors are superimposed in the arc apex region along the normal direction, forming a resultant force perpendicular to the arc apex surface and pointing towards the interior of the concrete. This resultant force breaks the original isotropic propagation mode of the vibration wave in the arc apex region, causing the vibration energy to be concentrated and output along the normal direction. When acting on the air bubbles, it generates a directional driving force, driving the air bubbles to detach from the mold's inner wall along the normal direction and migrate upwards. Outside the arc apex, the vibrators are started sequentially according to a predetermined time sequence, and the excitation force progresses in waves over time. The combined effect is that the area of ​​strong vibration migrates along the arc length direction. The vibration gradient at the leading edge of this area generates a pushing effect pointing in the migration direction, causing the air bubbles to gradually advance towards the arc apex along the target guide path during the vibration compaction process.

[0075] For example, the driving force field intensity of the directional driving force field is obtained, such as a peak excitation force of 5000 Newtons. This intensity value and the mold structure parameters are input into a pre-established finite element energy mapping model for iterative calculation, outputting a three-dimensional vibration energy distribution matrix of the mold's inner domain. This distribution matrix is ​​traversed to extract the set of spatial coordinates with energy values ​​below a preset threshold, such as 20 joules, and these regions with extremely low vibration energy are marked as spatial node positions.

[0076] In one possible implementation, the velocity characteristics of the medium are calculated based on the extracted spatial node positions. The particle vibration velocities of the nodes and their neighborhoods are extracted, and the average velocity of the medium is calculated to be less than 10 mm / s. This velocity characteristic is input into a rheological state identification algorithm to determine whether the medium is transitioning from a plastic flow state to a locally stagnant state. The phase transition critical point is determined based on this rheological state transition node, and the spatial surface where this critical point is located is used as the solid-liquid separation boundary to define the spatial distribution range of different physical states within the medium.

[0077] Example 2 like Figure 2 As shown, this embodiment provides a precast concrete segment directional air venting intelligent vibration compaction system based on an arc-directional migration vibration field, including: The target guidance path generation module is used to generate a smooth curve that fits the curvature of the inner wall of the mold based on the mold data, which serves as the target guidance path for bubble migration and discharge. The control node arrangement module is used to obtain the spatial position of the narrow restricted area inside the mold and arrange the control nodes based on the overlapping relationship between the target guide path and the restricted area; The initial parameter analysis module is used to determine the initial start time and initial vibration frequency of the vibrator corresponding to each node based on the parameters of each control node. The energy distribution analysis module is used to obtain the spatial distribution of vibration energy generated by each vibrator, determine the energy scattering characteristics, and identify the coordinates of the energy accumulation center. The parameter correction module is used to correct the initial start time and initial vibration frequency based on the deviation direction and deviation distance of the coordinates of the energy gathering center, so as to obtain the target start time and target vibration frequency of each vibrator. The collaborative operation control module is used to control the collaborative operation of each vibrator based on the target start time and target vibration frequency, and to generate a directional driving force field in the arc apex region.

[0078] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for directional air venting and intelligent vibration compaction of precast concrete segments based on arc-directional migration vibration field, characterized in that, Includes the following steps: A smooth curve that conforms to the curvature of the inner wall of the mold is generated based on the mold data and serves as the target guiding path for bubble migration and discharge. Obtain the spatial location of the narrow restricted area inside the mold, and arrange control nodes based on the overlapping relationship between the target guide path and the restricted area; The initial start time and initial vibration frequency of the vibrator corresponding to each node are determined based on the parameters of each control node. Obtain the spatial distribution of vibration energy generated by each vibrator, determine the energy scattering characteristics, and identify the coordinates of the energy accumulation center. Based on the deviation direction and deviation distance of the energy concentration center coordinates, the initial opening time and initial vibration frequency are corrected to obtain the target opening time and target vibration frequency of each vibrator. Based on the target start time and target vibration frequency, the vibrators are controlled to operate in a coordinated manner, generating a directional driving force field in the arc apex region.

2. The intelligent vibration method for directional air venting of precast concrete segments based on arc-directional migration vibration field according to claim 1, characterized in that, The process of generating a smooth curve that matches the curvature of the mold inner wall based on mold data as the target guiding path for bubble migration and discharge includes: Obtain the three-dimensional structural data and concrete pouring sequence data of the arc-shaped mold, and construct the initial connection line with the bottom pouring start point of the mold as the starting point and the top vent as the ending point; Based on the three-dimensional structural data, the curvature features and arc tangent direction of the inner wall of the mold are extracted to generate a geometric fit curve; A physical constraint model is established based on the pouring sequence data and concrete rheological parameters, and the geometric fit curve is corrected to obtain the physical correction path. The geometric fit curve and the physical correction path are weighted and fused, and then smoothed to obtain the smooth curve.

3. The intelligent vibration method for directional air venting of precast concrete segments based on arc-directional migration vibration field according to claim 2, characterized in that, The expression for the weighted fusion of the geometric fit curve and the physical correction path is as follows: ； In the formula, , To preset the curvature threshold, For geometric fit curves, For physical correction path, This refers to the local curvature of the inner wall of the mold.

4. The intelligent vibration method for directional air venting of precast concrete segments based on arc-directional migration vibration field according to claim 1, characterized in that, The process of arranging control nodes based on the overlapping relationship between the target guidance path and the restricted area includes: Obtain the spatial coordinates of the narrow restricted area inside the mold, compare the spatial coordinates with the target guide path, and extract the intersection line segments of the target guide path that enter each restricted area; Based on the starting and ending arc length positions of each intersecting line segment on the target guidance path, the target guidance path is segmented along the arc length direction to obtain multiple independent path segments with regional attribute labels; From each independent path segment, select the path segments located in the arc top area, water stop groove area, bottom area of ​​embedded part and dense reinforcement area, and arrange control nodes at equal intervals on these path segments according to the preset densification spacing. Arrange control nodes at regular intervals on the remaining path segments to obtain the initial control node set. Extract the target nodes located in the arc apex region from the initial control node set, calculate the rheological resistance based on the rheological parameters of the concrete at each target node, determine the vibration attenuation coefficient based on the rheological resistance, and perform power compensation on the target nodes that exceed the threshold based on the vibration attenuation coefficient to obtain the compensated target nodes. The compensated target node and the uncompensated regular node are merged in order of arc length to obtain multiple vibration control nodes distributed along the target guidance path.

5. The intelligent vibration method for directional air venting of precast concrete segments based on arc-directional migration vibration field according to claim 4, characterized in that, The expression for power compensation of target nodes exceeding the threshold based on the vibration attenuation coefficient is as follows: ； In the formula, This is the reference operating power of the vibrator. μ As a compensation index, To compensate for the power value, δ The vibration attenuation coefficient is... This is the threshold value for the vibration attenuation coefficient.

6. The intelligent vibration method for directional air venting of precast concrete segments based on arc-directional migration vibration field according to claim 1, characterized in that, The process of determining the initial start time and initial vibration frequency of the vibrator corresponding to each node based on the parameters of each control node includes: Obtain the spatial coordinates, local radius of curvature, concrete slump, and node sequence number of each vibration control node; A spatial curvature tensor is constructed based on the spatial coordinates and local curvature radius of each node, and the slump attenuation value is calculated based on the node's arrangement number, cumulative path distance, and concrete slump. The spatial curvature tensor is expanded and then concatenated with the collapse attenuation value to form a fused feature vector. The fused feature vector is predicted using a random forest regression model to obtain the initial start time and initial vibration frequency of the vibrator corresponding to each node.

7. The intelligent vibration method for directional air venting of precast concrete segments based on arc-directional migration vibration field according to claim 1, characterized in that, The process of obtaining the spatial distribution of vibration energy generated by each vibrator, determining the energy scattering characteristics, and identifying the coordinates of the energy accumulation center includes: The vibration energy sampling values ​​of the inner region of the mold are obtained when each vibrator is working at the initial opening time and initial vibration frequency, and the vibration energy spatial distribution matrix is ​​constructed. The coordinates of the energy peak points are extracted from the vibration energy spatial distribution matrix at each time step and time-series fitting is performed to obtain the time-series motion trajectory of the energy peak points; The presence of energy scattering characteristics is determined based on the maximum deviation distance between the temporal motion trajectory of the energy peak point and the target guidance path. When energy scattering characteristics are determined, the spatial coordinates of each vibrator are used as the initial cluster center to perform clustering on the vibration energy spatial distribution matrix to obtain the coordinates of the energy aggregation center.

8. The intelligent vibration method for directional air venting of precast concrete segments based on arc-directional migration vibration field according to claim 1, characterized in that, The process of correcting the initial start-up time and initial vibration frequency based on the deviation direction and deviation distance of the energy accumulation center coordinates includes: The deviation direction is determined based on the spatial orientation of the energy gathering center coordinates relative to the target guidance path, and the deviation distance is determined based on the minimum distance between the energy gathering center coordinates and the target guidance path; Based on the deviation distance, the difference scaling ratio is determined, and the initial opening time difference of adjacent vibrators is scaled according to the deviation direction, and the sum is used to obtain the corrected opening time of each vibrator. Based on the deviation distance, the arc apex compensation coefficient and arc trough compensation coefficient are determined respectively. The arc trough compensation coefficient is used to compensate for the initial vibration frequency, and the arc apex compensation coefficient is used to compensate for the initial vibration power, so as to obtain the target vibration frequency and target vibration power. The target start time is obtained by fine-tuning the corrected start time based on the power compensation amount and the deviation direction.

9. The intelligent vibration method for directional air venting of precast concrete segments based on arc-directional migration vibration field according to claim 1, characterized in that, The process of controlling the coordinated operation of each vibrator based on the target start time and target vibration frequency to generate a directional driving force field in the arc apex region includes: Each vibrator is activated based on its target start time and target vibration frequency. The vibration waves generated by each vibrator are superimposed to form a strong vibration zone that moves along the target guidance path. The movement trajectory of the strong vibration area is monitored. When the movement trajectory enters the arc apex area, the percussion device is controlled to stay for a preset duration. Within the duration, the phase difference of each percussion device is adjusted to zero and switched to the same frequency and phase operation mode. The percussion devices are synchronously superimposed to form a directional driving force field.

10. A precast concrete segment directional venting intelligent vibration system based on an arc-directional migration vibration field, characterized in that, The system for implementing the method of claim 1, wherein the system comprises: The target guidance path generation module is used to generate a smooth curve that fits the curvature of the inner wall of the mold based on the mold data, which serves as the target guidance path for bubble migration and discharge. The control node arrangement module is used to obtain the spatial position of the narrow restricted area inside the mold and arrange the control nodes based on the overlapping relationship between the target guide path and the restricted area; The initial parameter analysis module is used to determine the initial start time and initial vibration frequency of the vibrator corresponding to each node based on the parameters of each control node. The energy distribution analysis module is used to obtain the spatial distribution of vibration energy generated by each vibrator, determine the energy scattering characteristics, and identify the coordinates of the energy accumulation center. The parameter correction module is used to correct the initial start time and initial vibration frequency based on the deviation direction and deviation distance of the coordinates of the energy gathering center, so as to obtain the target start time and target vibration frequency of each vibrator. The collaborative operation control module is used to control the collaborative operation of each vibrator based on the target start time and target vibration frequency, and to generate a directional driving force field in the arc apex region.

Images