Rapid evaluation method and system for bearing capacity of grouting at the bottom of prefabricated road slabs

Abstract

The present application relates to the field of road engineering inspection and maintenance, and discloses a rapid evaluation method and system for the bearing capacity of the bottom grouting of prefabricated road panels, which includes the following steps: scanning the grouting layer at the bottom of the prefabricated road panels to obtain radar image data of the void area; testing the pavement deflection basin data, and combining the geometric information of the void area to obtain the corresponding relationship between the deflection value and the position; analyzing the radar image data based on the void identification model of deep learning, identifying the unfilled area and extracting its spatial geometric features; combining the modulus inversion calculation model, inversely calculating the modulus of the grouting layer, the base modulus and the roadbed modulus; based on the mechanical model, evaluating the bearing capacity reduction rate of the void area and the insufficient bearing capacity of the single board; judging the grouting effect according to the evaluation results, and generating optimized construction suggestions. It can quickly and accurately evaluate the quality of the grouting layer, provide scientific construction guidance, and significantly improve the grouting construction quality and maintenance efficiency of prefabricated road panels.

Description

Technical Field

[0001] The invention relates to the technical field of road engineering detection and maintenance, and in particular to a method and system for quickly evaluating the bearing capacity of bottom grouting of assembled prefabricated road slabs. Background Art

[0002] With the development of prefabricated construction technology, the application of prefabricated road panels in road engineering is becoming more and more widespread. Prefabricated prefabricated panels have the advantages of high construction efficiency, stable quality control, and convenient disassembly and maintenance. Therefore, they have received widespread attention in the field of rapid construction and repair. However, after the construction of prefabricated road panels is completed, the quality of their grouting layer plays a vital role in the overall pavement performance. If the grouting layer is not full or has quality problems, it is very easy to cause defects such as hollowing out of the bottom of the board, cracking of the pavement, and settlement, which not only affects the bearing capacity of the pavement, but also significantly shortens its service life.

[0003] At present, the detection method for the grouting effect of the bottom of the prefabricated road slab mainly relies on a single physical measurement method, such as scanning the internal void area through ground penetrating radar, or testing the pavement deflection data through a heavy-duty drop weight deflectometer to evaluate the pavement structure performance. However, these methods have certain limitations in practical applications. On the one hand, it is difficult to accurately evaluate the bearing capacity of the grouting layer using only ground penetrating radar data, because ground penetrating radar can only provide geometric information of the internal structure and cannot reflect the changes in material properties and mechanical behavior; on the other hand, although relying solely on deflection testing can obtain the mechanical response data of the pavement, it cannot directly correspond to the spatial distribution information of the void area, and it is difficult to accurately locate and quantitatively analyze the impact of the unfilled area on the pavement performance.

[0004] In addition, in the existing technology, human subjective judgment plays an important role in the evaluation of grouting effect, and usually relies on engineering experience to interpret radar maps or deflection test results. This method is not only time-consuming and labor-intensive, but also easily affected by subjective factors, resulting in unstable and inaccurate judgment results. At the same time, traditional methods lack the means of fusion analysis of multi-source data, and cannot achieve a systematic correlation between the geometric characteristics of the grouting layer and the bearing capacity. Therefore, it is difficult to provide scientific and accurate evaluation results when faced with complex pavement structures and uneven void distribution.

[0005] In summary, the existing technology still lacks an efficient, accurate and comprehensive technical means for quality inspection and bearing capacity evaluation of the grouting layer at the bottom of prefabricated road panels. This deficiency has limited the promotion and application of prefabricated road panel technology to a certain extent. Therefore, there is an urgent need for a rapid evaluation method that combines multi-source data fusion and advanced analysis technology, which can not only accurately identify the unfilled areas of the grouting layer, but also quantify its impact on the road bearing capacity, thereby providing a reliable basis for construction quality control and optimization. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention provides a method and system for quickly evaluating the bearing capacity of the bottom grouting of prefabricated road slabs, which solves the problems in the prior art of inaccurate evaluation of the quality and bearing capacity of the bottom grouting layer of prefabricated road slabs, low efficiency and lack of scientific construction guidance.

[0007] To achieve the above purpose, the present invention is implemented by the following technical scheme: A rapid evaluation method for the bearing capacity of the bottom grouting of a prefabricated road slab comprises the following steps:

[0008] The grouting layer at the bottom of the prefabricated slab is scanned by ground penetrating radar to obtain radar image data of the void area;

[0009] The road deflection basin data is tested by a heavy-duty drop-weight deflectometer, and the corresponding relationship between the deflection value and the position is obtained by combining the geometric information of the void area.

[0010] The deep learning-based void recognition model is used to analyze radar images, identify the void areas, and extract their spatial geometric features.

[0011] Combined with the deflection basin data, based on the optimized modulus inversion calculation model, the modulus of the grouting layer, the modulus of the base layer and the subgrade are inverted and calculated;

[0012] Evaluate the bearing capacity reduction rate of the void area and the insufficient bearing capacity of the single plate based on the mechanical model;

[0013] The grouting effect is judged based on the evaluation results and optimized construction suggestions are generated.

[0014] Preferably, the void recognition model is based on a lightweight convolutional neural network algorithm, and the convolutional neural network algorithm includes:

[0015] Extract features from radar images and use a weight-optimized network structure to improve the recognition accuracy of unfilled areas;

[0016] The hybrid attention mechanism is used to extract spatial and channel features, which enhances the model's ability to capture missing features at different locations.

[0017] The pruning optimization algorithm is used to optimize the model and reduce the consumption of computing resources.

[0018] Preferably, the step of inversely calculating the modulus of the grouting layer, the base modulus and the roadbed modulus based on the optimized modulus inversion calculation model in combination with the deflection basin data comprises:

[0019] According to the deflection test data and radar image information, the modulus distribution of the void area is calculated in combination with the modulus inversion model;

[0020] The modulus inversion model is based on an optimized machine learning algorithm. The input includes deflection value and spatial geometric characteristics of the void area. The output includes the modulus of the grouting layer, the base modulus and the subgrade modulus.

[0021] The modulus prediction error in the inversion process is dynamically optimized through reinforcement learning methods to enhance the inversion accuracy.

[0022] Preferably, the modulus inversion model adopts the XGBoost algorithm to improve the parameter search efficiency in the process of reinforcement learning optimization, wherein the input geometric features include the length, width and depth information of the void area, and the deflection test data includes the distribution of deflection values ​​at each measuring point.

[0023] Preferably, the step of evaluating the bearing capacity reduction rate of the hollow area and the insufficient bearing capacity of the single plate based on the mechanical model includes:

[0024] The partial differential equation is used to describe the stress-strain distribution of the grouting layer. Combined with the elastic modulus and thickness of the grouting layer and the geometric shape of the void area, the maximum bearing capacity reduction rate of the grouting layer and the insufficient bearing capacity of the void area are calculated.

[0025] Preferably, the bearing capacity reduction rate is calculated by the ratio of the modulus of the unfilled area to the modulus of the filled area, the modulus of the unfilled area is obtained through a modulus inversion model, and the modulus of the filled area is determined by the material parameters of the grouting layer.

[0026] Preferably, the generating of optimized construction suggestions includes:

[0027] When the grouting degree is greater than 0.7 and the insufficient bearing capacity ratio is less than 20%, local filling of high-fluidity grouting materials is used for the unfilled area;

[0028] When the grouting filling degree is less than 0.7 or the insufficient bearing capacity ratio is greater than 20%, the precast slab shall be replaced and re-grouted.

[0029] The present invention also provides a rapid evaluation system for the bearing capacity of the bottom grouting of a prefabricated road slab, comprising:

[0030] Data acquisition module, used to collect ground penetrating radar scanning data and heavy-duty drop weight deflectometer test data;

[0031] The void identification module is used to identify the unfilled area based on radar data and extract the spatial geometric features;

[0032] Modulus inversion module, used to calculate the modulus of the grouting layer, base modulus and subgrade modulus based on deflection data and geometric characteristics;

[0033] The bearing capacity evaluation module is used to calculate the bearing capacity reduction rate of the void area and the insufficient bearing capacity of the single plate based on the mechanical model;

[0034] The decision-making module is used to generate optimized construction recommendations based on the evaluation results.

[0035] The present invention provides a method and system for quickly evaluating the bearing capacity of the bottom grouting of prefabricated road slabs. It has the following beneficial effects:

[0036] 1. The present invention combines multiple data sources such as ground penetrating radar and deflection test, and uses deep learning models and modulus inversion technology to quickly identify unfilled areas and quantify their impact on the pavement bearing capacity. This multi-source fusion analysis makes the grouting layer quality evaluation more scientific and intuitive, avoiding the errors that may be caused by relying on experience in traditional methods.

[0037] 2. The present invention generates construction suggestions based on the evaluation results, and can provide refined treatment plans for the actual conditions of different areas, such as local re-grouting or re-grouting. This highly targeted construction guidance not only improves construction efficiency, but also avoids unnecessary material waste and repeated construction.

[0038] 3. The present invention combines the geometric features detected by radar with the mechanical data of deflection test to comprehensively judge the grouting effect and the change of pavement performance, which greatly improves the accuracy of the evaluation results. Through data visualization, it helps construction personnel to intuitively understand the location and specific impact of the problem area, and improves the reliability of judgment and decision-making.

[0039] 4. The present invention uses machine learning, partial differential equation mechanics modeling and other technologies to achieve full process automation and intelligence from detection to analysis to construction suggestions. This not only improves the scientific and technological content of road maintenance work, but also provides technical support for the widespread promotion of prefabricated road panel technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0041] Figure 2 It is a schematic diagram of the system structure of the present invention.

[0042] Among them, 10, data acquisition module; 20, void identification module; 30, modulus inversion module; 40, bearing capacity evaluation module; 50, decision-making module. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Please see attached Figure 1 The present invention provides a rapid evaluation method for the bearing capacity of the bottom grouting of prefabricated road slabs, which is used to solve the problem of inaccurate and inefficient evaluation of the bearing capacity of the unfilled grouting area in the prior art. By integrating ground penetrating radar detection technology, deflection test technology, deep learning recognition model and mechanical model calculation technology, the filling effect and bearing capacity of the grouting layer are quickly evaluated to generate optimized construction suggestions. The specific implementation methods of each step are described in detail below.

[0045] like Figure 1 As shown, the rapid evaluation method for the bearing capacity of the bottom grouting of the assembled prefabricated road slab may include the following steps:

[0046] S1, ground penetrating radar scanning to obtain radar image data;

[0047] S2, deflection test to obtain road deflection basin data;

[0048] S3, identification of empty areas and extraction of geometric features;

[0049] S4, modulus inversion calculation based on optimized modulus inversion calculation model;

[0050] S5. Bearing capacity evaluation based on mechanical model;

[0051] S6. Generate optimized construction suggestions.

[0052] For step S1, in this embodiment, the grouting layer at the bottom of the assembled prefabricated board is scanned by a ground penetrating radar to obtain radar image data of the void area. Specifically, this embodiment utilizes the high-resolution characteristics of the ground penetrating radar equipment and combines the slice image analysis technology to identify the void position and distribution inside the grouting layer. The following is a detailed description from three aspects: equipment selection, scanning method and data processing.

[0053] In one possible implementation, a three-dimensional air-pair ground penetrating radar is preferably used, and the detection depth range of the ground penetrating radar is set between 5 cm and 20 cm, which can meet the scanning requirements of the grouting layer under the thin concrete slab. At the same time, the equipment has a fast detection capability, and its detection speed can reach 30 km / h, which is suitable for large-scale road rapid detection tasks.

[0054] As an option, the operating frequency range of the ground penetrating radar in this embodiment can be between 1 GHz and 2 GHz to achieve high-resolution scanning of the pavement structure. Specifically, the selection of the frequency range is based on the electromagnetic properties of different grouting layer thicknesses and base materials. For example, when the grouting layer is thin (such as less than 5 cm), it is preferred to use a higher frequency (close to 2 GHz) for scanning to obtain clearer information on the void area.

[0055] It should be noted that ground penetrating radar records the time delay and intensity of electromagnetic waves propagating inside the medium by emitting high-frequency electromagnetic waves and receiving their reflected signals. There is a significant difference in the dielectric constant between filled areas and unfilled areas, so different reflection characteristics are shown in the radar echo image. Inside the grouting layer, the echo signal of the unfilled area usually has a higher reflection intensity and a shorter time delay.

[0056] In some embodiments, the result of the ground penetrating radar scanning is presented in the form of radar slice images, each slice image representing the electromagnetic reflection signal of the grouting layer at a specific depth. Exemplarily, the slice image can be divided into a number of pixel units, and the gray value of each pixel unit corresponds to the reflection signal intensity at that position.

[0057] In order to improve the recognition accuracy of radar images, the radar data is preprocessed in this embodiment. Specifically, the preprocessing includes the following steps:

[0058] As an option, the radar image is subjected to denoising processing, preferably using a wavelet transform filtering method, to eliminate signal noise caused by equipment noise or external environmental interference.

[0059] In a possible implementation, the radar image is further subjected to contrast enhancement processing. Exemplarily, the contrast between the empty area and the filled area is enhanced by using a histogram equalization technique, thereby facilitating subsequent image segmentation and feature extraction.

[0060] It should be noted that the pre-processed radar slice image can establish a mapping relationship with the specific road surface position through the spatial coordinate information, thereby associating the unfilled area in the image with the actual position of the road surface. This mapping relationship can be achieved through the GPS positioning module of the ground penetrating radar device.

[0061] In some possible implementations, in order to accurately obtain the unfilled area inside the grouting layer, the present embodiment also performs three-dimensional reconstruction of the radar slice image. Specifically, by superimposing slice images of multiple depths, a three-dimensional model of the unfilled area inside the grouting layer can be constructed. Exemplarily, the geometric features of the unfilled area include its length (a), width (b) and depth (c). These geometric features can be extracted by an image segmentation algorithm, such as a segmentation method based on edge detection or region growing.

[0062] It is understandable that the volume of the unfilled area It can be calculated by the following formula:

[0063]

[0064] in, , , are the length, width and depth of the non-filled area, respectively.

[0065] In another implementation, the embodiment further distinguishes the reflection characteristics of different media by performing spectrum analysis on the radar echo signal. For example, by converting the time domain signal to the frequency domain space through Fourier transform and analyzing the energy distribution of different frequency components, the size and shape of the unfilled area can be determined.

[0066] Specifically, the frequency of the reflected signal in the unfilled area is usually biased towards the high frequency part, while the frequency of the reflected signal in the filled area is lower. This frequency characteristic can be used to verify the recognition result of the empty area in the radar slice image, forming a double guarantee.

[0067] It should be noted that the ground penetrating radar scanning method in this embodiment is applicable to both the detection of the grouting layer after the construction is completed and the real-time monitoring of the grouting effect during the construction process. For example, after the grouting is completed, multiple scans are performed by the ground penetrating radar, and the results of each scan are compared to determine the material filling state during the grouting process and whether there is a secondary voiding phenomenon.

[0068] It can be understood that the ground penetrating radar scanning step in this embodiment provides key data support for subsequent void area identification, modulus inversion calculation and bearing capacity evaluation, and is a basic link for realizing a rapid evaluation method.

[0069] For step S2, in this embodiment, the deflection basin data of the bottom pavement of the prefabricated slab is tested by a heavy-duty drop weight deflectometer (HWD), and the corresponding relationship between the deflection value and the position is obtained by combining the geometric information of the void area. Specifically, the vertical deformation of the pavement structure under load is evaluated through loading and response measurement analysis, and key parameter support is provided for subsequent modulus inversion and bearing capacity analysis.

[0070] As an option, the HWD equipment used can simulate actual vehicle loads, and its loading capacity is usually in the range of 30kN to 300kN, and the impact time during loading is less than 30 milliseconds to meet the load requirements of different pavement structures. It should be noted that by adjusting the radius of the loading plate (such as between 150mm and 300mm), the road panel size in different areas can be adapted to ensure the accuracy of the deflection test.

[0071] In a possible implementation, the HWD device arranges several test points at equal intervals on the road surface, obtains the deflection value through point-by-point loading test, and records the spatial coordinate information of each test point in combination with the GPS positioning module. For example, the spacing of the test points can be adjusted according to the size of the road panel and the scale of the void area, for example, the spacing is 1 m or other suitable values ​​to ensure a balance between data coverage and measurement accuracy.

[0072] It is understandable that the core of the HWD deflection test is to apply load to the road surface through a loading plate and measure the vertical displacement (i.e., deflection value) at the loading center and several locations around it. These deflection values ​​can reflect the overall stiffness of the pavement structure and the unevenness of the grouting layer and the base layer. In the case of a void area in the grouting layer, the deflection value at the corresponding location will usually increase significantly, providing additional data support for the precise positioning of the void area.

[0073] Specifically, in this embodiment, obtaining the deflection basin data includes:

[0074] As an option, the magnitude and location of the loading force need to correspond to the void area detected by the ground penetrating radar. For example, when the radar image shows that a certain area may have voids, test points can be arranged in this area first and a larger loading force can be applied to enhance the test sensitivity. It should be noted that this deflection test strategy based on radar detection results can effectively improve the test efficiency and avoid ineffective full-area scanning.

[0075] In some embodiments, the measurement range of the deflection value data is not limited to the loading center point, but also includes several positions within a certain radius from the loading center. For example, sensors can be arranged at radial positions of 0.3 m, 0.6 m, and 0.9 m around the loading center to record multi-point deflection values. This multi-point measurement method can construct a complete deflection basin surface, thereby more intuitively reflecting the mechanical response characteristics of the grouting layer.

[0076] It should be noted that the accuracy of the deflection data has an important impact on the subsequent modulus inversion calculation. In this embodiment, a high-precision displacement sensor is preferably used, and its measurement resolution can reach 0.01 mm, which can accurately capture subtle vertical deformation of the road surface.

[0077] In a possible implementation, in order to enhance the applicability and consistency of the deflection value data, this embodiment performs interpolation processing on the collected data. Specifically, the discrete deflection test data can be expanded into a continuous deflection basin surface using a bicubic spline interpolation method. This processing method can effectively smooth the test error and provide more comprehensive road deformation information for subsequent analysis.

[0078] As an option, the deflection basin surface can be expressed in the form of a function, where the deflection value at the loading center is the largest and gradually decreases with increasing radial distance. For example, the deflection value surface can be expressed as:

[0079]

[0080] in, is the maximum deflection value at the loading center, is the radial distance from the loading center to the test point, It is a parameter that describes the extension range of the surface. By fitting the actual test data, the specific parameter value of the surface can be determined.

[0081] It should be noted that the deflection value data can not only directly reflect the impact of the void area on the stiffness of the pavement structure, but also can be combined with the geometric information detected by the ground penetrating radar to further locate the position and scale of the void area. In this embodiment, the peak point of the deflection value is matched with the high reflection intensity point in the radar image to confirm the center position of the void area. Specifically, by analyzing the spatial overlap between the deflection value surface and the radar spectrum, the specific impact of the void area on the bearing capacity of the grouting layer can be quantified.

[0082] In some embodiments, in order to improve the reliability of the deflection value data, a repeated loading test is also introduced in this embodiment. Exemplarily, the load is applied to the same test point 3 times, and the deflection value of each loading is recorded. When the difference between the deflection values ​​of repeated loading is less than 5%, the test result can be considered stable and used for subsequent analysis. It can be understood that this repeated loading test can effectively eliminate accidental errors caused by loose materials or sensor drift.

[0083] In another possible implementation, in order to meet the needs of large-scale road testing, a multi-axis heavy-duty drop weight deflectometer device is used in this embodiment, which can measure the deflection values ​​of multiple test points simultaneously in one loading. Specifically, the multi-axis HWD device is equipped with multiple loading plates and sensors, which can synchronously apply loads at multiple positions with fixed spacing and record response data. Exemplarily, the deflection value data within a range of 5 m in length and 3 m in width can be collected simultaneously in one loading, thereby significantly improving the test efficiency.

[0084] It should be noted that the deflection basin data obtained in this embodiment provides key input parameters for the subsequent modulus inversion calculation. By combining the geometric information of the void area detected by the ground penetrating radar, this embodiment can establish an accurate correspondence between the deflection value and the specific position of the road surface, thereby laying the foundation for the calculation of modulus distribution and bearing capacity analysis.

[0085] For step S3, in this embodiment, the radar image data acquired by the ground penetrating radar is analyzed using a void identification model based on deep learning to identify the unfilled area of ​​the grouting layer at the bottom of the prefabricated board and extract its spatial geometric features. Specifically, a lightweight convolutional neural network (CNN) model is used to accurately identify the unfilled area from the radar image and calculate the geometric parameters of the area such as length, width and depth through a combination of feature extraction, attention enhancement and pruning optimization.

[0086] As an option, the radar image data used in this embodiment is pre-processed slice image data, which can clearly reflect the reflection characteristics of the grouting layer after denoising and enhancement. Specifically, the radar image characteristics of the unfilled area are usually manifested as a highlighted area with a strong echo signal, whose boundary is relatively clear and has a significant contrast difference with the filled area.

[0087] It should be noted that the void identification in this step mainly relies on deep learning technology to realize the automated analysis of complex radar image data. By training the neural network model, it is possible to effectively overcome the limitations of traditional manual feature extraction methods on the complexity and diversity of radar images, thereby improving the accuracy and efficiency of void area identification.

[0088] In a possible implementation, the gap recognition model in this embodiment is based on an improved lightweight convolutional neural network (CNN), and the model structure includes a feature extraction module, an attention mechanism module, and a pruning optimization module.

[0089] Specifically, the feature extraction module is based on the lightweight structure of the VGGNet network and extracts the spatial features of the radar image through multi-layer convolution and pooling operations. As an option, in order to improve the model's ability to capture the features of the empty area, this embodiment optimizes the feature extraction network and adopts a weighted feature mapping method. For example, for the input radar image , its feature map is expressed as:

[0090]

[0091] in, represents the convolution operation, is the convolution kernel parameter, is the extracted feature map.

[0092] In some embodiments, the attention mechanism module adopts a hybrid attention mechanism, including the joint calculation of channel attention and spatial attention. Specifically, channel attention is used to evaluate the importance of different channels in the feature map, while spatial attention focuses on the local area characteristics of the feature map. It is understandable that this attention mechanism can dynamically adjust the weight of the feature, thereby enhancing the model's ability to recognize key areas (i.e., sparse areas).

[0093] For example, for the input feature map , and its attention weight calculation formula is:

[0094]

[0095] in, is the channel attention map, is the spatial attention map, and is the weight parameter.

[0096] In another implementation, in order to reduce the computational complexity of the model and improve the inference speed, a pruning optimization module is also introduced in this embodiment. It should be noted that pruning optimization can reduce computational overhead without significantly reducing recognition accuracy by deleting redundant parameters in the network. Specifically, pruning optimization uses a velocity-free particle swarm optimization algorithm (PSO), that is, based on the particle swarm optimization algorithm, a linear combination of the optimal position of the particle and the optimal position of the group is combined to update the parameter position of each particle. Exemplarily, the particle update formula is:

[0097]

[0098] in, and Respectively represent The particle in The local best position and the global best position in the iteration, , is the acceleration factor, , is a random weight.

[0099] The model optimized through pruning can process large-scale radar image data more efficiently and is suitable for rapid identification of void areas on site.

[0100] In some embodiments, the geometric features of the identified empty area are extracted and calculated in this embodiment. Specifically, by segmenting the highlighted area in the radar image, the boundary of the unfilled area is determined, and then the length (a), width (b) and depth (c) of the area are calculated, so as to calculate the volume of the unfilled area. The volume calculation formula is shown in the description of step S1. As an option, the area segmentation can adopt a threshold-based segmentation method, and the threshold of the reflection intensity is set to distinguish the unfilled area from the filled area.

[0101] It should be noted that the volume calculation results are not only used to describe the scale of the void area, but also serve as the key input parameter for the subsequent modulus inversion calculation.

[0102] In another possible implementation, this embodiment further performs spatial verification on the void identification results. Specifically, the identified underfilled area is compared and analyzed with the HWD deflection test results. For example, for the test points with significantly higher deflection values, it is verified whether they coincide with the spatial position of the identified void area. This multi-source data fusion verification method can effectively improve the reliability and accuracy of the identification results.

[0103] It can be understood that the void identification and geometric feature extraction in this step provide necessary data support for subsequent modulus inversion calculation and bearing capacity evaluation.

[0104] For step S4, in this embodiment, by combining the deflection basin data and the geometric features of the void area, based on the optimized modulus inversion calculation model, the modulus of the grouting layer, the modulus of the base layer and the roadbed modulus are inverted and calculated. Specifically, the inversion calculation technology based on machine learning is adopted, the deflection test results and geometric features are used as input, and the modulus distribution of the area is output through model training and optimization.

[0105] As an option, the XGBoost algorithm optimized by reinforcement learning is used in this embodiment to perform modulus inversion calculation. It should be noted that XGBoost is an integrated learning method that can achieve modeling of complex data relationships through the integration of multiple decision trees, and the introduction of reinforcement learning further improves the efficiency of model parameter optimization.

[0106] In a possible implementation, the input data for modulus inversion calculation includes:

[0107] Deflection basin data: represents the vertical deformation value of the loading center and surrounding areas, reflecting the mechanical response characteristics of the pavement structure.

[0108] Geometric features: The length, width and depth of the void area obtained from radar recognition results are used to quantify the size and shape of the void area.

[0109] Grouting material parameters: including the initial modulus and material thickness of the grouting layer, which are used to provide the initial conditions of the model.

[0110] It should be noted that the goal of the modulus inversion calculation is to derive the modulus of the grouting layer from the input data ( )、Base modulus( ) and subgrade modulus ( ) to reflect the mechanical properties of each layer of the pavement structure.

[0111] Specifically, the modulus inversion calculation in this embodiment includes:

[0112] In some embodiments, the inversion calculation process can be expressed as a nonlinear optimization problem, the goal of which is to minimize the error between the calculated modulus and the actual modulus. Exemplarily, the objective function can be expressed as:

[0113]

[0114] in, For the The actual deflection value of each test point is is the deflection value calculated based on the modulus inversion result, is the number of test points.

[0115] As an option, in this embodiment, the finite element method (FEM) is used to numerically simulate the deflection basin, and the modulus is iteratively adjusted in combination with the above-mentioned optimization objective function. It should be noted that the FEM simulation process takes into account the distribution of the loading force, the geometric characteristics of the void area, and the material parameters of each layer, so as to ensure the accuracy of the inversion calculation results.

[0116] In a possible implementation, the introduction of reinforcement learning in this embodiment further improves the efficiency and accuracy of modulus inversion calculation. Specifically, reinforcement learning models modulus inversion as a dynamic optimization process, the core of which includes:

[0117] State: includes the current modulus value, deflection basin data, and geometric features.

[0118] Action: Adjust the magnitude and direction of the modulus value to minimize the objective function value.

[0119] Reward: The reward value is defined according to the decrease of the objective function, that is:

[0120]

[0121] in, and Respectively and The objective function value of the iteration.

[0122] It should be noted that through continuous iteration of the reinforcement learning strategy, the model can gradually approach the optimal modulus value, thereby improving the accuracy of the inversion calculation.

[0123] In some embodiments, in order to further improve the efficiency of the inversion calculation, the particle swarm optimization (PSO) algorithm is used in this embodiment to optimize the hyperparameters of XGBoost. Specifically, the PSO algorithm simulates the movement behavior of particles in the search space to find the hyperparameter combination that minimizes the model error. Exemplarily, the update formula of PSO is:

[0124]

[0125] in, For particles In Dimension speed, is the current location, and are the historical optimal position and global optimal position of the particle respectively, , , is the control parameter of the algorithm.

[0126] It can be understood that PSO optimization can effectively reduce the parameter search time of the XGBoost model, thereby significantly improving the efficiency of modulus inversion.

[0127] In another possible implementation, the results of modulus inversion are verified and corrected in this embodiment. Specifically, the modulus value obtained by inversion is substituted into the mechanical model to calculate the corresponding deflection value, and compared with the actual test data. It should be noted that when the deviation between the calculated deflection value and the actual deflection value exceeds a preset threshold (e.g., 5%), the input parameters of the inversion model are further adjusted until the calculated result is basically consistent with the actual data.

[0128] Exemplarily, the corrected modulus value in this embodiment can be used as an input parameter for subsequent bearing capacity analysis to calculate the bearing capacity reduction rate of the grouting layer and the insufficient bearing capacity of the single board.

[0129] In another possible implementation, the present embodiment further combines a multi-layer neural network model to predict and interpolate the modulus distribution. Specifically, by training the multi-layer neural network model, it is possible to predict the modulus distribution of the untested area based on the existing modulus value, thereby completing the spatial data of the modulus inversion. It should be noted that this method can significantly improve the spatial resolution of the modulus distribution and provide more comprehensive data support for subsequent bearing capacity evaluation.

[0130] It can be understood that the modulus inversion calculation in this embodiment provides key basic data for the mechanical performance evaluation of the grouting layer, base layer and roadbed.

[0131] For step S5, in this embodiment, the bearing capacity reduction rate and the insufficient bearing capacity of a single board are evaluated for the hollow area at the bottom of the prefabricated slab based on the mechanical model. Specifically, by combining the modulus distribution obtained by inversion calculation and the geometric characteristics of the hollow area, a mechanical model based on partial differential equations is established to quantify the impact of the unfilled area on the overall bearing capacity of the pavement, and provide a reference for construction optimization.

[0132] As an option, the bearing capacity evaluation process in this embodiment includes the following two core parts: first, the bearing capacity reduction rate of the grouting layer is calculated according to the modulus distribution; second, the single precast panel is divided into several areas to comprehensively evaluate its insufficient bearing capacity.

[0133] It should be noted that this step is based on the evaluation of mechanical principles, which can accurately reflect the physical and mechanical behavior of the void area, and at the same time improves the evaluation accuracy by combining geometric characteristics and modulus distribution.

[0134] In a possible implementation, the calculation of the bearing capacity reduction rate is based on the following formula:

[0135]

[0136] in, is the modulus of the unfilled region, is the modulus of the filled region, Represents the bearing capacity reduction rate.

[0137] It is understandable that when When the unfilled area has a negative impact on the bearing capacity of the road surface, the reduction degree is For example, if , it means that the carrying capacity of the unfilled area is only 80% of that of the filled area.

[0138] As an option, the modulus of the unfilled area The modulus of the filled area is calculated by the modulus inversion model. It can be determined based on the initial performance parameters of the grouting material. For example, for standard cement-based grouting materials, A typical value is 30 GPa.

[0139] In some embodiments, in order to further quantify the overall bearing capacity deficiency of a single prefabricated panel, each prefabricated panel is divided into a number of equal-area regions in this embodiment, and the bearing capacity reduction rate of each region is calculated respectively, and the comprehensive bearing capacity deficiency value of the single panel is calculated according to the following formula:

[0140]

[0141] in, The number of regions divided for the board. For the The carrying capacity reduction rate of the region, Indicates the insufficient value of the comprehensive bearing capacity of the single board.

[0142] It should be noted that, in this embodiment, the single board is preferably divided into 9 areas (i.e. ), the modulus of each region center point is calculated by the modulus inversion model. And by taking weighted average, the overall bearing capacity evaluation result of the single board can be obtained.

[0143] In another possible implementation, this embodiment further analyzes the stress-strain distribution in the unfilled area by establishing a mechanical model based on partial differential equations (PDE). Specifically, assuming that the grouting layer is a thin layer of elastic body, its physical behavior can be described by the following partial differential equation:

[0144]

[0145] in, is the stress tensor, is the external load density. Stress tensor and the strain tensor is related by the generalized Hooke's law:

[0146]

[0147] in, is the modulus of the grouting layer.

[0148] As an option, the above partial differential equations are discretized and solved using the finite difference method (FDM) to obtain the stress and strain distribution in the void area. For example, for each discrete element, the calculation formula is:

[0149]

[0150] in, represents the displacement field, is the external load density.

[0151] It should be noted that the solution results of the partial differential equation can be used to analyze the local stress concentration phenomenon in the void area and further evaluate the bearing capacity of the void area.

[0152] In some embodiments, the present embodiment also performs spatial visualization of the bearing capacity reduction rate of the void area. Specifically, the modulus distribution, bearing capacity reduction rate and stress distribution results in the void area are represented in the form of three-dimensional graphics, which facilitates intuitive analysis of the impact of the unfilled area on the overall pavement structure.

[0153] It is understandable that spatial visualization processing can help engineers quickly locate problem areas and determine their specific impact on pavement performance, providing important reference for subsequent construction decisions.

[0154] In another possible implementation, this embodiment also performs a sensitivity analysis on the bearing capacity reduction in combination with different working conditions. For example, different load intensities and loading positions are simulated to calculate the corresponding changes in the bearing capacity reduction rate. For example, when the loading position is close to the void area, the bearing capacity reduction rate may increase significantly, while when the loading position is far from the void area, the impact may be relatively small.

[0155] It should be noted that this sensitivity analysis can provide a scientific basis for the selection of grouting materials and the improvement of construction technology, and can also evaluate the potential impact of different construction schemes on pavement performance.

[0156] It can be understood that the bearing capacity evaluation in this embodiment provides an accurate quantitative index for analyzing the underfilled area of ​​the grouting layer.

[0157] For step S6, in this embodiment, the grouting effect is judged based on the evaluation result of the bearing capacity of the grouting layer, and an optimized construction suggestion is generated. Specifically, by calculating the grouting filling degree and the insufficient bearing capacity, combined with specific threshold rules, the treatment method of the unfilled area is determined, and construction optimization suggestions are provided to ensure that the grouting effect of the prefabricated road panel meets the engineering requirements.

[0158] As an option, the construction in this embodiment recommends grouting with a filling degree of And the single board load capacity is insufficient It is a key judgment indicator. It should be noted that It indicates the overall filling of the grouting layer and is defined as the ratio of the volume of the unfilled area to the volume of the grouting area. The calculation formula is:

[0159]

[0160] in, is the volume of the unfilled area, is the total volume of the grouting area.

[0161] It is understandable that when When it is close to 1, it means that the filling effect of the grouting layer is better; When it is lower, the grouting construction process needs to be further optimized.

[0162] Specifically, this embodiment uses the following rules to judge the grouting effect and generate construction suggestions:

[0163] In one possible implementation, when the grouting filling degree is And the single board bearing capacity is insufficient When the filling area is insufficient, it is recommended to perform local filling of the insufficient area. For example, the local filling can be repaired by drilling holes and grouting at the center of the insufficient area using high-flow cement-based grouting materials.

[0164] As an option, the drilling location and the scope of recharging can be determined based on the spatial geometric characteristics of the unfilled area. ), single hole grouting is preferred; while for larger volume depleted areas (such as ), then the porous grouting method can be used to ensure uniform distribution of the grouting material.

[0165] In another possible implementation, when the grouting filling degree is Or the single board bearing capacity is insufficient It is recommended to remove the prefabricated panels and re-grout. It should be noted that such situations usually indicate that the structural performance of the grouting layer cannot meet the design requirements, and the grouting effect must be ensured through re-construction.

[0166] It is understandable that the material selection and construction process should be optimized during the re-grouting process. For example, for areas with repeated underfilling problems, it is preferred to use grouting materials with better resistance to separation, and at the same time introduce real-time monitoring technology during grouting construction to ensure the uniformity and fullness of the grouting process.

[0167] In some embodiments, this embodiment further combines a visualization display module of construction suggestions to provide graphical feedback to users on grouting effect evaluation results and optimization suggestions. Specifically, color coding is used to identify grouting effects in different areas, for example:

[0168] Grouting fullness The area is marked in green, indicating that the grouting effect is excellent;

[0169] The area is marked in yellow, indicating that local re-irrigation is needed;

[0170] or Areas are marked in red where removal and re-grouting are recommended.

[0171] For example, this graphical display can be viewed in real time on a tablet computer or other terminal device at the construction site, thereby helping engineers to quickly develop construction plans.

[0172] It should be noted that the construction suggestions in this embodiment are not only based on the current grouting effect evaluation results, but can also be optimized in combination with historical construction data and working conditions. In a possible implementation, by analyzing the impact of different construction processes on the grouting effect, a correlation model between construction parameters and grouting quality is established to guide subsequent construction optimization.

[0173] For example, for areas with lower modulus of the base layer (such as ), grouting materials with stronger fluidity can be used preferentially; for areas with frequent loading and higher risk of voiding, materials with better durability can be selected and the frequency of construction quality monitoring can be increased.

[0174] In another possible implementation, this embodiment further combines multi-source data fusion technology to optimize construction suggestions. Specifically, the location of the empty area detected by the ground penetrating radar is fused and analyzed with the mechanical response data of the deflection test to comprehensively judge the impact of the unfilled area on the overall pavement performance. For example, for areas with significantly larger deflection values, priority is given to arranging refilling points to reduce construction risks.

[0175] It is understandable that this multi-source data fusion technology can effectively improve the pertinence and reliability of construction recommendations while reducing unnecessary construction investment.

[0176] In some embodiments, in order to verify the actual effect of the construction suggestions, this embodiment also introduces a post-construction evaluation process. Specifically, after the completion of the supplementary grouting or re-grouting, the ground penetrating radar detection and deflection test are re-performed, and the detection results are compared with the original evaluation results. It should be noted that when the post-construction grouting filling degree is And the bearing capacity is insufficient The grouting effect can be regarded as qualified.

[0177] It is understandable that this post-construction evaluation process can form a closed-loop management mechanism and provide data support for the continuous optimization of grouting construction quality.

[0178] It can be understood that the construction suggestion generation and optimization process in this embodiment fully considers the fullness of the grouting layer, the bearing capacity evaluation result and the construction feasibility.

[0179] In general, the present invention obtains the data of the void area of ​​the grouting layer through ground penetrating radar scanning, combines the deflection basin data collected by the heavy-duty drop weight deflectometer, uses the void identification model based on deep learning to extract the geometric features of the unfilled area, and calculates the modulus distribution of the grouting layer and the pavement structure through the optimized modulus inversion model. Based on the partial differential equation mechanical model, the bearing capacity reduction rate of the grouting layer and the insufficient bearing capacity of a single plate are quantified, and optimized construction suggestions are generated according to the filling degree and the insufficient bearing capacity ratio, such as local supplementary grouting or re-grouting. Rapid and accurate evaluation of the grouting effect is achieved, providing a scientific basis for improving the construction quality and optimizing maintenance of prefabricated road panels.

[0180] The rapid evaluation system for the bearing capacity of the bottom grouting of prefabricated road slabs described below and the rapid evaluation method for the bearing capacity of the bottom grouting of prefabricated road slabs described above can be referred to each other.

[0181] Please see attached Figure 2 The present invention also provides a rapid evaluation system for the bearing capacity of the bottom grouting of prefabricated road slabs, comprising:

[0182] A data acquisition module 10 is used to collect ground penetrating radar scanning data and heavy-duty drop weight deflectometer test data;

[0183] A void identification module 20, for identifying the unfilled area based on radar data and extracting spatial geometric features;

[0184] A modulus inversion module 30, for calculating the modulus of the grouting layer, the base modulus and the roadbed modulus based on the deflection data and geometric characteristics;

[0185] The bearing capacity evaluation module 40 is used to calculate the bearing capacity reduction rate of the hollow area and the insufficient bearing capacity of the single plate based on the mechanical model;

[0186] The decision module 50 is used to generate optimized construction suggestions according to the evaluation results.

[0187] The system of this embodiment can be used to execute the above method embodiments, and its principles and technical effects are similar, which will not be repeated here.

[0188] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A rapid evaluation method for the bearing capacity of the bottom grouting of prefabricated road slabs, characterized in that: The following steps are involved: The grouting layer at the bottom of the prefabricated slab is scanned by ground penetrating radar to obtain radar image data of the void area; The road deflection basin data is tested by a heavy-duty drop-weight deflectometer, and the corresponding relationship between the deflection value and the position is obtained by combining the geometric information of the void area. The deep learning-based void recognition model is used to analyze radar images, identify the void areas, and extract their spatial geometric features. Combined with the deflection basin data, based on the optimized modulus inversion calculation model, the modulus of the grouting layer, the modulus of the base layer and the subgrade are inverted and calculated; Evaluate the bearing capacity reduction rate of the void area and the insufficient bearing capacity of the single plate based on the mechanical model; Judge the grouting effect based on the evaluation results and generate optimized construction suggestions; The step of inversely calculating the modulus of the grouting layer, the base modulus and the roadbed modulus based on the optimized modulus inverse calculation model in combination with the deflection basin data includes: According to the deflection test data and radar image information, the modulus distribution of the void area is calculated in combination with the modulus inversion model; The modulus inversion model is based on an optimized machine learning algorithm. Its input includes deflection values ​​and spatial geometric characteristics of the void area, and its output includes the modulus of the grouting layer, the base modulus, and the subgrade modulus.

2. The rapid evaluation method for the bearing capacity of the bottom grouting of prefabricated road slabs according to claim 1 is characterized in that: The gap recognition model is based on a lightweight convolutional neural network algorithm, which includes: Extract features from radar images and use a weight-optimized network structure to improve the recognition accuracy of unfilled areas; The hybrid attention mechanism is used to extract spatial and channel features, which enhances the model's ability to capture missing features at different locations. The pruning optimization algorithm is used to optimize the model and reduce the consumption of computing resources.

3. The rapid evaluation method for the bearing capacity of the bottom grouting of the prefabricated road slab according to claim 1 is characterized in that: The step of inversely calculating the modulus of the grouting layer, the base modulus and the roadbed modulus based on the optimized modulus inverse calculation model in combination with the deflection basin data also includes: The modulus prediction error in the inversion process is dynamically optimized through reinforcement learning methods to enhance the inversion accuracy.

4. The rapid evaluation method for the bearing capacity of the bottom grouting of the prefabricated road slab according to claim 3 is characterized in that: The modulus inversion model adopts the XGBoost algorithm to optimize the parameter search efficiency in the process of reinforcement learning, wherein the input geometric features include the length, width and depth information of the void area, and the deflection test data includes the distribution of deflection values ​​at each measuring point.

5. The rapid evaluation method for the bearing capacity of the bottom grouting of prefabricated road slabs according to claim 1 is characterized in that: The step of evaluating the bearing capacity reduction rate of the hollow area and the insufficient bearing capacity of the single plate based on the mechanical model includes: The partial differential equation is used to describe the stress-strain distribution of the grouting layer. Combined with the elastic modulus and thickness of the grouting layer and the geometric shape of the void area, the maximum bearing capacity reduction rate of the grouting layer and the insufficient bearing capacity of the void area are calculated.

6. The rapid evaluation method for the bearing capacity of the bottom grouting of prefabricated road slabs according to claim 5 is characterized in that: The bearing capacity reduction rate is calculated by the ratio of the modulus of the unfilled area to the modulus of the filled area. The modulus of the unfilled area is obtained through a modulus inversion model, and the modulus of the filled area is determined by the material parameters of the grouting layer.

7. The rapid evaluation method for the bearing capacity of the bottom grouting of prefabricated road slabs according to claim 1 is characterized in that: The generated optimized construction suggestions include: When the grouting degree is greater than 0.7 and the insufficient bearing capacity ratio is less than 20%, local filling of high-fluidity grouting materials is used for the unfilled area; When the grouting filling degree is less than 0.7 or the insufficient bearing capacity ratio is greater than 20%, the precast slab shall be replaced and re-grouted.

8. A rapid evaluation system for the bearing capacity of the bottom grouting of a prefabricated road slab, based on the rapid evaluation method for the bearing capacity of the bottom grouting of a prefabricated road slab as claimed in any one of claims 1 to 7, characterized in that: include: Data acquisition module, used to collect ground penetrating radar scanning data and heavy-duty drop weight deflectometer test data; The void identification module is used to identify the unfilled area based on radar data and extract the spatial geometric features; Modulus inversion module, used to calculate the modulus of the grouting layer, base modulus and subgrade modulus based on deflection data and geometric characteristics; The bearing capacity evaluation module is used to calculate the bearing capacity reduction rate of the void area and the insufficient bearing capacity of the single plate based on the mechanical model; The decision-making module is used to generate optimized construction recommendations based on the evaluation results.

Images