A highway pavement compressive capacity detection system and detection method

Abstract

The present application relates to the technical field of nondestructive testing of highway engineering, in particular to a highway pavement compressive capacity detection system and method, comprising: setting a detection grid on the pavement to be detected and applying a preloading signal, obtaining stress wave propagation spectrum and deformation relaxation time series of each point. According to the wave peak attenuation law, the design load sequence is constructed, and according to the deformation fluctuation trend, the detection scanning path is constructed. Synchronously running both and collecting dynamic response data, and iteratively comparing with the response mode library to identify the target response mode. Based on the rules and conditions of the mode, load reconstruction instructions and path optimization instructions are generated, and executed to correct the subsequent detection process. The method realizes the dynamic generation of detection parameters according to the real-time response of the pavement, and the online adaptive optimization of the detection process based on the intermediate results, which improves the accuracy and efficiency of the evaluation of the compressive capacity of the pavement.

Description

Technical Field

[0001] This invention relates to the field of non-destructive testing technology for highway engineering, specifically a method for testing the compressive strength of highway pavement. Background Technology

[0002] Existing methods for testing the compressive strength of highway pavements generally employ standardized static loading or fixed-mode dynamic loading. Before implementation, the loading sequence, amplitude, and movement path of the testing equipment are typically pre-set based on general specifications or historical experience. This pre-defined testing mode struggles to adapt to the real-time response of pavements with non-uniform materials and varying internal damage, resulting in insufficient matching between the testing load and the actual mechanical properties of the pavement, and the testing path may not effectively cover critical response areas.

[0003] Another drawback of conventional technical solutions lies in the linear and open-loop characteristics of the detection process. Data obtained from a single detection process is primarily used for final condition assessment and compliance determination; the entire detection operation itself remains fixed during execution. Even if an anomaly is detected, it's impossible to adjust subsequent loading and scanning strategies in real time for deeper investigation. This results in insufficient sensitivity to capturing changes in the internal state of complex pavement structures, and a lack of flexibility and deep self-optimization capabilities in the detection process. Therefore, a detection method is needed that can dynamically generate detection parameters based on real-time pavement feedback and achieve self-correction during the detection process. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and to propose a road pavement compressive strength testing system and method.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for testing the compressive strength of highway pavement, comprising:

[0006] A detection area grid is set on the road surface to be tested, and a controllable preload signal is applied to each sub-grid point in the detection area grid;

[0007] In response to the preload signal, the initial road surface features fed back by each sub-grid point in the detection area grid are obtained, and the initial road surface features include stress wave propagation spectrum and deformation relaxation time series;

[0008] Based on the attenuation law of the peaks in the stress wave propagation spectrum, a design load sequence is constructed, and based on the fluctuation trend of the deformation relaxation time sequence, a detection scanning path is constructed.

[0009] The design load sequence and the detection scanning path are run synchronously to obtain the road surface dynamic response data corresponding to the synchronous operation process;

[0010] The dynamic response data of the road surface is iteratively compared with a preset response pattern library, and the target response pattern that matches the current road surface structure is identified in the response pattern library.

[0011] Based on the load distribution rules and deformation constraints in the target response mode, load reconstruction instructions for adjusting the design load sequence and path optimization instructions for adjusting the detection scanning path are generated.

[0012] The load reconstruction command and the path optimization command are executed to correct the compressive strength testing process of the detection area mesh.

[0013] As a further aspect of the present invention, a distributed pressure pulse generator is used to sequentially apply a set of pressure pulses with progressively increasing values ​​to each sub-grid point in the detection area grid.

[0014] Using a multi-axis high-resolution sensor array, the full-frequency vibration response signal of each sub-grid point after receiving the pressure pulse is captured, and the vibration response signal is encoded into a vibration energy spectrum.

[0015] The vibration energy spectrum is subjected to modal decomposition to extract the frequency components and phase information corresponding to the dominant modes, and the frequency components and phase information are combined to generate the stress wave propagation spectrum.

[0016] Simultaneously, using a laser interferometric ranging device, the time interval sequence required for the surface deformation of each subgrid point to recover to its initial state after the pressure pulse stops is measured, and the time interval sequence is recorded as the deformation relaxation time sequence;

[0017] A spatiotemporal correlation mapping is established between the stress wave propagation spectrum and the deformation relaxation time series, so that the wave propagation characteristics and deformation recovery characteristics of each sub-grid point form a one-to-one corresponding feature pair within the grid of the detection area.

[0018] As a further aspect of the present invention, the stress wave propagation spectrum is divided into regions to identify stable regions where the peak attenuation rate is lower than a standard threshold, and sensitive regions where the peak attenuation rate is higher than a standard threshold.

[0019] A stepped increasing load application scheme is configured for the stable region, wherein the amplitude increment of each load step in the load application scheme is calculated proportionally based on the energy retention rate of the wave peak within the stable region.

[0020] An oscillating test load application scheme is configured for the sensitive area. The load amplitude in the load application scheme fluctuates slightly around a reference value. The reference value is reverse-calibrated based on the average attenuation rate of the wave peaks in the sensitive area.

[0021] The stepped incremental load application scheme for the stable region and the oscillating trial load application scheme for the sensitive region are sorted and connected according to their spatial positions in the grid of the detection area, and integrated to generate the design load sequence.

[0022] Cluster analysis was performed on the deformation relaxation time series to divide it into fast relaxation clusters and slow relaxation clusters based on the density of time intervals.

[0023] A continuous scanning path is planned for the fast relaxation cluster, and the moving speed of the continuous scanning path is positively correlated with the average speed of deformation recovery within the cluster;

[0024] Plan an intermittent pause scan path for the slow relaxation cluster, wherein the duration of the intermittent pause scan path at each sampling point within the cluster is proportional to the deformation relaxation time of the sampling point;

[0025] The continuous scanning path of the fast relaxation cluster and the intermittent pause scanning path of the slow relaxation cluster are spatially fused and temporally coordinated to generate the detection scanning path covering the entire detection area grid.

[0026] As a further aspect of the present invention, the load execution unit is activated, and loads are applied sequentially at the corresponding sub-grid points according to the load application scheme specified in the design load sequence;

[0027] Simultaneously, the path scanning unit is activated, driving the detection probe to move strictly following the detection scanning path, and high-frequency data acquisition is performed at each position point or dwell time period specified by the path.

[0028] During each collaborative working cycle of the load execution unit and the path scanning unit, three sets of key data are recorded: first, the peak value of transient compressive stress at the loaded subgrid point at the moment of load application; second, the strain distribution cloud map of adjacent points with conductive correlation to the load application position captured by the detection probe on the scanning path; and third, the energy dissipation curve of the entire process from the start of load application to the strain distribution stabilizing.

[0029] The transient compressive stress peak, strain distribution cloud map and energy dissipation curve recorded in each collaborative work cycle are timestamped and encapsulated to form an independent dynamic response data package.

[0030] According to the execution order of the design load sequence and the detection scanning path, the dynamic response data packets generated by all collaborative work cycles are concatenated to form the complete road dynamic response data.

[0031] As a further aspect of the present invention, a standard feature template for each candidate response mode is extracted from a plurality of candidate response modes pre-stored in the response mode library. The standard feature template includes a standard stress peak range, a standard strain cloud shape, and a standard energy curve envelope.

[0032] From the currently acquired pavement dynamic response data stream, the measured transient compressive stress peak sequence, the measured strain distribution cloud map sequence, and the measured energy dissipation curve sequence are extracted.

[0033] Perform the first round of feature comparison: calculate the overall distribution of the measured transient compressive stress peak sequence and the degree of overlap with the standard stress peak interval of each candidate response mode;

[0034] Only candidate response patterns with an overlap exceeding a preset overlap threshold are retained for the next round of comparison;

[0035] A second round of morphological comparison is conducted: the spatiotemporal evolution characteristics of the measured strain distribution cloud map sequence are compared with the standard strain cloud map morphology of the candidate response modes retained in the previous round to perform structural similarity analysis.

[0036] Only candidate response patterns with structural similarity exceeding a preset similarity threshold are retained for the final round of comparison;

[0037] Perform the final round of envelope comparison: match the overall trend of the measured energy dissipation curve sequence with the standard energy curve envelope of the candidate response modes retained in the previous round to ensure trend consistency.

[0038] From the final round of comparison, the candidate response mode with the highest trend consistency matching degree is selected and determined as the target response mode that matches the current road structure;

[0039] If, after multiple rounds of comparison, no candidate response pattern simultaneously meets all threshold conditions, the pattern learning mechanism is triggered to summarize the current road dynamic response data stream and its characteristics into a new response pattern and store it in the response pattern library.

[0040] As a further aspect of the present invention, the load distribution rules encapsulated in the target response mode are analyzed, and the load distribution rules define the ideal load spectrum that different types of road surface areas should bear;

[0041] The load spectrum actually generated during the execution of the design load sequence is compared point by point with the ideal load spectrum defined by the target response mode to identify the load application points with significant deviations, as well as the amount and direction of the deviations.

[0042] For each load application point with significant deviation, a specific load correction item is generated. The load correction item includes the location identifier of the load application point, the suggested adjusted load amplitude, and the suggested load application duration.

[0043] All load correction items are compiled in the order of their execution in the design load sequence to generate the load reconfiguration instruction;

[0044] Simultaneously, the deformation constraint conditions encapsulated in the target response mode are analyzed. These deformation constraint conditions specify the allowable gradient range of road surface deformation in space and the evolution rhythm in time under ideal response.

[0045] The measured deformation data collected during the execution of the detection scanning path is compared with the deformation constraint conditions in two dimensions: spatial gradient and temporal evolution, to locate the path segment or sampling point in the scanning path that causes the measured deformation data to exceed the allowable range.

[0046] For each identified problematic path segment or sampling point, a specific path adjustment item is generated. The path adjustment item includes adjusting the scanning speed, extending or shortening the dwell time at a specific point, or fine-tuning the scanning trajectory to avoid abnormal deformation areas.

[0047] All path adjustment items are integrated according to their logical order in the detection scan path to generate the path optimization instruction.

[0048] As a further aspect of the present invention, the load execution unit receives and parses the load reconstruction instruction, and extracts all load correction items therefrom;

[0049] For each load correction item, the load execution unit updates the target load parameters of the corresponding load application point in its internal state table. The target load parameters include amplitude and duration.

[0050] In subsequent detection cycles, when the load execution unit needs to apply a load again at the location point specified by the corresponding load correction item, it will call the updated target load parameters instead of the parameters in the original design load sequence, thereby realizing real-time load correction.

[0051] The path scanning unit receives and parses the path optimization instruction, and extracts all path adjustment items from it;

[0052] For each path adjustment item, the path scanning unit modifies the motion parameters of the corresponding path segment or sampling point in its internal path planner. The motion parameters include speed, dwell time, or coordinates.

[0053] In subsequent detection cycles, the path scanning unit drives the detection probe to operate according to the modified motion parameters, thereby achieving dynamic optimization of the scanning path;

[0054] After the load execution unit and the path scanning unit complete the instruction-based correction, the system re-enters the data acquisition state to acquire a new round of dynamic road surface response data after process correction, for subsequent analysis and evaluation.

[0055] As a further aspect of the present invention, a vehicle-mounted wide-area survey device is used to cruise at low speed along the road section to be measured, and macroscopic image data of the road surface and shallow ground-penetrating radar reflection data are collected simultaneously.

[0056] The macroscopic image data is subjected to texture recognition and crack segmentation to calculate the area ratio and distribution density of visible defects on the road surface, and a road surface integrity index is generated.

[0057] The geological radar reflection data is analyzed for stratigraphic position and dielectric constant is inverted to identify the clarity of interlayer interfaces and material homogeneity of the subgrade structure, and to generate a structural layer health index.

[0058] The road table integrity index and the structural layer health index are input into a pre-trained classification decision tree model;

[0059] The classification decision tree model outputs a recommended detection grid density level and preloaded signal strength range for the current road segment under test based on the input double exponential combination.

[0060] Based on the recommended detection grid density level, the system automatically divides the spacing and number of sub-grid points in the detection area grid;

[0061] The system sets the initial amplitude range of the controllable preload signal based on the recommended preload signal strength range.

[0062] As a further aspect of the present invention, during the model building stage, a large amount of historical road segment survey data is collected as training samples. Each training sample includes the road surface integrity index, structural layer health index, and the optimal detection grid density and optimal preloading intensity calibrated by expert experience for the corresponding road segment.

[0063] Using the road surface integrity index and structural layer health index as decision features, the optimal detection grid density and optimal preloading intensity as decision objectives, and information gain or Gini impurity as splitting criteria, the classification decision tree model is recursively constructed.

[0064] During the model working phase, the road surface integrity index and structural layer health index calculated for the current road segment are input starting from the root node of the decision tree;

[0065] At each decision node, the input data is distributed to different child nodes according to the feature judgment rules of the current node;

[0066] The decision is passed down the decision path until it reaches the leaf node, where the recommended detection grid density level and the preloaded signal strength range are stored.

[0067] The system has a built-in feedback calibration mechanism: after each full-process detection based on the recommended detection grid density level and the preloaded signal strength range, the key response features found in the actual detection process, including the number of abnormal sensitive areas and the frequency of load adjustment, are used as a post-hoc evaluation index.

[0068] The correlation analysis is performed between the post-abstract evaluation indicators and the initial recommended parameters of the decision tree model. If a systematic bias is found, the model parameters are fine-tuned, and the judgment thresholds of the relevant decision nodes are updated so that the model recommendations are more in line with the actual detection needs.

[0069] As a further aspect of the present invention, the present invention also includes a road pavement compressive strength testing system, the system including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements the steps of the road pavement compressive strength testing method described above.

[0070] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0071] The design load sequence is constructed based on the attenuation law of the peaks in the stress wave propagation spectrum, and the detection scanning path is constructed based on the fluctuation trend of the deformation relaxation time series. This ensures that the load excitation mode applied to the road surface and the scanning trajectory of the sensor are directly derived from the physical response characteristics of the road surface to the initial excitation. The construction of the load sequence takes into account the dissipation characteristics of stress waves in actual road surface materials, and the construction of the scanning path follows the time law of local deformation recovery of the road surface. The detection parameters are no longer fixed values ​​detached from the detection object, but variables generated by the real-time characteristics of the object, significantly improving the fit between the detection action and the current state of the specific road surface, and enhancing the targeting of the detection.

[0072] The dynamic response data of the road surface obtained during synchronous operation is iteratively compared with a preset response mode library. After identifying the matching target response mode, load reconstruction instructions and path optimization instructions are generated and executed based on the load distribution rules and deformation constraints in the mode. This process constitutes a real-time decision-making and control closed loop embedded in the detection process. Detection is no longer a fixed sequence of procedures executed only once, but an adaptive process that can self-feedback and self-adjust based on intermediate results. By correcting the load sequence and scanning path online through instructions, the detection can automatically focus on more sensitive parameter ranges and more critical spatial locations, improving the entire process's ability to distinguish and diagnose complex road conditions. Attached Figure Description

[0073] Figure 1 This is a flowchart of a method for testing the compressive strength of highway pavement according to the present invention;

[0074] Figure 2 This is a comparison chart of the transient compressive stress peak sequence and the standard interval.

[0075] Figure 3 A schematic diagram illustrating the working principle of the method for generating load reconfiguration instructions and path optimization instructions;

[0076] Figure 4 This is a multi-dimensional characteristic curve diagram of the road surface dynamic response;

[0077] Figure 5 Analysis diagram of the design load sequence and detection scanning path. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0079] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0080] See Figure 1A rectangular detection area grid is defined on the road surface to be tested, consisting of multiple uniformly distributed sub-grid points. A central control system sends commands to a multi-axis loading device deployed above the detection area grid. This device sequentially applies a set of preload signals, the amplitude, frequency, and duration of which can be independently controlled, to each sub-grid point. The sensing system integrated on the loading device responds to these preload signals, capturing feedback information from each sub-grid point in real time and processing it to obtain initial road surface characteristics including stress wave propagation patterns and deformation relaxation time series. Based on the energy attenuation patterns of each wave peak in the stress wave propagation pattern from the excitation point to the receiving point, the data processing unit designs a series of design load sequences with specific spatial and temporal distributions. Simultaneously, based on the clustering or trend fluctuations exhibited by each time value in the deformation relaxation time series, a detection scanning path adaptable to the deformation recovery rate of different areas is planned. The control system synchronously drives the loading device to apply loads according to the design load sequence and drives the mobile detection platform to scan along the detection scanning path. During this coordinated process, dynamic road surface response data, including transient stress, dynamic strain, and energy dissipation, are collected. The dynamic response data is transmitted to the analysis module, where it undergoes multiple rounds of iterative comparison with a pre-stored response pattern library containing the response characteristics of various typical pavement structures under standard loads. By calculating feature overlap, morphological similarity, and trend consistency, the module selects the target response pattern from the library that best matches the current measured data. The analysis module further analyzes the load distribution rules and deformation constraints embedded in the target response pattern, comparing them with the currently executed design load sequence and detection scanning path to generate load reconstruction instructions and path optimization instructions containing specific parameter adjustment items. The control system receives and executes these instructions, dynamically adjusting the load application parameters and scanning motion parameters in subsequent detection cycles, thereby achieving online correction and optimization of the grid compressive strength testing process in the detection area.

[0081] In one embodiment of the present invention, initial road surface features are acquired and a design load sequence and detection scanning path are constructed. A distributed pressure pulse generator sequentially applies a set of pressure pulses with increasing increments to each sub-grid point in the detection area grid. A multi-axis high-resolution sensor array captures the full-frequency vibration response signal of each sub-grid point after receiving the pressure pulse. The vibration response signal is encoded into a vibration energy spectrum after fast Fourier transform processing. The vibration energy spectrum is modally decomposed to extract the frequency components and phase information corresponding to the dominant mode. The frequency components and phase information are combined to generate a stress wave propagation map. A laser interferometric ranging device measures the time interval sequence required for the surface deformation of each sub-grid point to recover to the initial state after the pressure pulse stops. The time interval sequence is recorded as a deformation relaxation time series. A spatiotemporal correlation mapping between the stress wave propagation map and the deformation relaxation time series is established so that the wave propagation characteristics and deformation recovery characteristics of each sub-grid point form a one-to-one corresponding feature pair within the detection area grid. In practical implementation, the stress wave propagation spectrum is segmented to identify stable regions where the peak attenuation rate is below a standard threshold and sensitive regions where the peak attenuation rate is above a standard threshold. A stepped-increasing load application scheme is then configured for the stable regions. The amplitude increment of each load step in the stepped-increasing load application scheme is calculated proportionally based on the energy retention rate of the peak within the stable region. The proportional calculation formula is as follows:

[0082] ;

[0083] in: Indicates the increment of load amplitude. This represents the proportionality coefficient. This represents the energy retention rate. An oscillating test load application scheme is configured for the sensitive area. The load amplitude in the oscillating test load application scheme fluctuates slightly around a reference value. The reference value is reverse-calibrated based on the average attenuation rate of the wave peak in the sensitive area. The average attenuation rate is the average value of the wave peak energy attenuation rate in the sensitive area. Reverse calibration means that the reference value is inversely proportional to the average attenuation rate. The stepped increasing load application scheme in the stable area and the oscillating test load application scheme in the sensitive area are sorted and integrated according to their spatial location to generate the design load sequence.

[0084] In some embodiments, cluster analysis is performed on the deformation relaxation time series to divide it into fast relaxation clusters and slow relaxation clusters based on the density of time intervals. Fast relaxation clusters contain points with short deformation relaxation times, while slow relaxation clusters contain points with longer deformation relaxation times. A continuous scanning path is planned for the fast relaxation clusters, with the movement speed of the continuous scanning path positively correlated with the average deformation recovery speed within the fast relaxation cluster. An intermittent pause scanning path is planned for the slow relaxation clusters, with the dwell time of the intermittent pause scanning path at each sampling point within the slow relaxation cluster proportional to the deformation relaxation time of the sampling point. The continuous scanning paths of the fast relaxation clusters and the intermittent pause scanning paths of the slow relaxation clusters are spatially fused and temporally coordinated to generate a detection scanning path covering the entire detection area grid. In some embodiments, spatial fusion uses a linear interpolation algorithm to ensure path continuity, and temporal coordination is based on a global timetable to synchronize load application and path scanning. Optionally, when constructing the design load sequence, a standard threshold is derived from historical data statistics to distinguish between stable and sensitive regions. Optionally, the cluster analysis uses a density-based clustering algorithm such as the DBSCAN algorithm to distinguish between fast and slow relaxation clusters. It is understandable that the stress wave propagation spectrum reflects the stiffness distribution of the pavement, while the deformation relaxation time series reflects the viscoelastic properties of the pavement. The spatiotemporal correlation mapping combines wave propagation characteristics with deformation recovery characteristics to provide a comprehensive view of pavement characteristics, guiding the construction of design load sequences and detection scan paths.

[0085] In one embodiment of the present invention, the synchronous operation process is matched with the response mode. The load execution unit applies loads sequentially to the corresponding sub-grid points according to the load application scheme specified in the design load sequence. The path scanning unit synchronously drives the detection probe to move strictly following the detection scanning path and performs high-frequency data acquisition at each position point or dwell time period specified in the path. In each collaborative working cycle of the load execution unit and the path scanning unit, the system records three sets of key data. The key data include the transient compressive stress peak value of the loaded sub-grid point at the moment of load application, the strain distribution cloud map of adjacent points with transmission correlation with the load application position captured by the detection probe on the scanning path, and the energy dissipation curve of the entire process from the start of load application to the strain distribution stabilizing. The transient compressive stress peak value, strain distribution cloud map and energy dissipation curve recorded in each collaborative working cycle are timestamped and encapsulated to form an independent dynamic response data packet. The dynamic response data packets generated by all collaborative working cycles are concatenated according to the execution order of the design load sequence and the detection scanning path to form a complete road dynamic response data stream. In practice, the response mode library stores multiple candidate response modes. Each candidate response mode has a standard feature template, which includes a standard stress peak range, a standard strain cloud map shape, and a standard energy curve envelope. The measured transient compressive stress peak sequence, the measured strain distribution cloud map sequence, and the measured energy dissipation curve sequence are parsed from the currently acquired pavement dynamic response data stream. The first round of feature comparison is performed to calculate the overlap between the overall distribution of the measured transient compressive stress peak sequence and the standard stress peak range of each candidate response mode. The overlap calculation involves the percentage evaluation of the statistical distribution overlap area. Only candidate response modes with an overlap exceeding a preset overlap threshold are retained for the next round of comparison.

[0086] In some embodiments, a second round of morphological comparison is performed by analyzing the spatiotemporal evolution characteristics of the measured strain distribution cloud map sequence against the standard strain cloud map morphology of the candidate response modes retained from the previous round. The structural similarity analysis employs a multi-scale structural similarity index algorithm, retaining only candidate response modes with structural similarity exceeding a preset similarity threshold for the final round of comparison. In some embodiments, a final round of envelope comparison is performed by matching the overall trend of the measured energy dissipation curve sequence with the standard energy curve envelope of the candidate response modes retained from the previous round to ensure trend consistency. Trend consistency matching is achieved by calculating the dynamic time warping distance.

[0087] ;

[0088] in: Indicates the dynamic time-warped distance. Indicates the length of the regular path. Indicates path weight, Represents the measured curve sequence points With standard envelope sequence points The local distance between the two is used to select the candidate response pattern with the highest trend consistency matching degree from the final round of comparisons, which is then determined as the target response pattern that matches the current pavement structure. Optionally, the preset overlap threshold and preset similarity threshold are set based on the statistical results of historical matching experiments. If no candidate response pattern meets all the threshold conditions after multiple rounds of comparisons, the pattern learning mechanism is triggered. It can be understood that the pattern learning mechanism summarizes the current pavement dynamic response data stream and its characteristics into a new response pattern and stores it in the response pattern library.

[0089] See Figure 2 This is a comparison chart of the transient compressive stress peak sequence and the standard range. This chart is the core visualization result of the first round of feature comparison in highway pavement compressive strength testing, showing the variation pattern of the measured transient compressive stress peak over 50 collaborative working cycles. The blue broken line represents the measured transient compressive stress peak, with data points clearly marked. The overall trend exhibits high-frequency fluctuations, with peaks mainly concentrated in the 70–100 MPa range, and occasional spikes exceeding 100 MPa. The rapid fluctuations in stress peaks over a short period reflect the dynamic response of the pavement under different loads, which is consistent with the step-by-step increase and oscillating trial load scheme in the design load sequence. Most measured values ​​fall within the 70–100 MPa range, which highly overlaps with the standard ranges of candidate mode B (80–100 MPa) and candidate mode A (70–90 MPa). A few spikes exceeding 100 MPa may correspond to load application in sensitive areas, and their matching degree with the standard mode needs to be closely monitored in subsequent comparisons.

[0090] See Figure 3 In one embodiment of the present invention, load reconstruction instructions and path optimization instructions are generated. The load distribution rules encapsulated in the target response mode are parsed. These load distribution rules define the ideal load spectrum that different types of pavement areas should bear in the form of functions or data tables. The load spectrum actually generated during the execution of the design load sequence is compared point-by-point with the ideal load spectrum defined by the target response mode. The absolute difference between the actual load amplitude and the ideal load amplitude at each corresponding application point, as well as the phase difference between the actual load timing and the ideal load timing, are calculated point-by-point. Load application points with significant deviations, along with their deviation amounts and directions, are identified. The criteria for determining significant deviations are that both the absolute difference and the phase difference simultaneously exceed their respective set thresholds. For each load application point with a significant deviation, a specific load correction item is generated. This load correction item includes the location identifier of the load application point, the suggested adjusted load amplitude, and the suggested load application duration. The suggested adjusted load amplitude is calculated using the formula:

[0091] ;

[0092] in: This indicates the suggested adjusted load amplitude. Indicates the ideal load amplitude. This indicates the actual load amplitude. This represents a convergence coefficient between 0 and 1 used to control the adjustment range. All load correction items are assembled in the order of their execution in the design load sequence to generate load reconfiguration instructions.

[0093] In some embodiments, deformation constraints encapsulated in the target response mode are parsed synchronously. These constraints, in the form of a system of inequalities, define the allowable gradient range of pavement deformation in space and its temporal evolution rhythm under ideal response. The measured deformation data collected during the detection scanning path execution is compared with the deformation constraints in both spatial gradient and temporal evolution dimensions. Spatial gradient comparison calculates the difference in deformation between adjacent sampling points, while temporal evolution comparison calculates the change in deformation recovery rate at the same location over time. Path segments or sampling points in the scanning path that cause the measured deformation data to exceed the allowable range are located. For each located problematic path segment or sampling point, a specific path adjustment item is generated. This item includes adjusting the scanning speed, extending or shortening the dwell time at specific points, or fine-tuning the scanning trajectory to avoid abnormal deformation zones. All path adjustment items are integrated according to their logical order in the detection scanning path to generate path optimization instructions. Optionally, the deviation direction indicates whether the actual load is higher or lower than the ideal load. The allowable gradient range in space defines the maximum value of the deformation difference between adjacent points, and the temporal evolution rhythm defines the minimum value of the deformation recovery rate. It is understandable that the load reconfiguration instruction is used to guide the load execution unit to correct the subsequent loading strategy.

[0094] In one embodiment of the present invention, the load execution unit receives and parses a load reconstruction instruction from the superior processing unit. The load reconstruction instruction is a structured data list. The data parsing module of the load execution unit extracts all load correction items from the load reconstruction instruction. For each load correction item contained in the load reconstruction instruction, the load execution unit finds the corresponding status record in its internally maintained load parameter status table by querying the record index that matches the position identifier in the load correction item. It then updates the target load parameters stored in the status record using the suggested adjusted load amplitude and suggested load application duration from the load correction item. The target load parameters include amplitude and duration. In subsequent detection loops, when the load execution unit needs to apply load at a certain load application point again, the control logic of the load execution unit first queries the internally updated load parameter status table, calls the updated target load parameters corresponding to the position point in the table to drive the pressure actuator, thereby replacing the preset parameters in the original design load sequence and realizing real-time correction of the load application strategy at that point. Refer to Table 1 for example data of load correction items.

[0095] Table 1 Load Correction Items:

[0096]

[0097] In some embodiments, the path scanning unit receives and parses path optimization instructions from the same upstream processing unit. The instruction parsing module of the path scanning unit extracts all path adjustment items from the path optimization instructions. For each path adjustment item contained in the path optimization instructions, the path scanning unit, within its internal path planner, locates the path segment or sampling point that needs adjustment based on the path segment identifier or sampling point coordinates specified in the path adjustment item. The path planner modifies the motion parameters of the corresponding path segment or sampling point according to the specific content of the path adjustment item. The motion parameters include speed, dwell time, or coordinates. In some embodiments, the modification of motion parameters follows a specific update logic, such as a new planned speed. From the original speed Combined with adjustment factor and reference speed Through the relation:

[0098] ;

[0099] in: This indicates the revised planning speed. Indicates the original velocity. This represents an adjustment factor between 0 and 1. This indicates the target speed reference value derived from the path adjustment terms. In subsequent detection cycles, when the path scanning unit drives the detection probe, the path planner will call the modified motion parameter sequence to control the probe to run according to the optimized speed, time, and trajectory, achieving dynamic optimization of the scanning path. Optionally, the load parameter status table is stored in key-value pair format, where the key is the position identifier and the value is a structure containing amplitude and duration parameters. Optionally, when the path adjustment terms include "adjust scanning speed," the adjustment factor... A value of 0.5 is used to achieve a smooth transition. This means that after both the load execution unit and the path scanning unit complete the instruction-based parameter correction and behavior adjustment, the system control logic automatically triggers a state switch to re-enter the data acquisition state. It can be understood that after the system re-enters the data acquisition state, it acquires a new round of road surface dynamic response data after process correction. This new round of road surface dynamic response data is used for subsequent analysis and evaluation.

[0100] See Figure 4 This is a multi-dimensional characteristic curve of pavement dynamic response. This graph shows the changes in the three core response indicators over time during a complete collaborative working cycle in highway pavement compressive strength testing, and is a key visualization result of the dynamic response data package. The transient compressive stress peak rises rapidly from 0–5 seconds, reaching a peak of approximately 150 kPa around 5 seconds and briefly plateauing. It then decreases linearly from 5–10 seconds, falling back to 100 kPa at 10 seconds, reflecting the complete process of load application and release. Energy dissipation increases linearly from 0 to 80% from 0–5 seconds, rising synchronously with the stress peak, reflecting the energy conversion of the pavement under load. It continues to slowly increase to 100% from 5–10 seconds, indicating that even as stress decreases, energy dissipation continues until the system stabilizes. The maximum strain distribution value rises rapidly to 80% from 0–4.5 seconds, highly synchronized with stress and energy dissipation, reflecting the development process of pavement deformation. It continuously decreases to 40% from 4.5–10 seconds, reflecting the trend of deformation recovery, matching the stress release process.

[0101] In one embodiment of the present invention, before setting up a detection area grid on the road surface to be tested and applying a controllable preload signal to each sub-grid point in the detection area grid, a road surface condition pre-assessment stage is performed. During the road surface condition pre-assessment stage, a vehicle-mounted wide-area survey device is used to cruise along the road section to be tested at low speed. The vehicle-mounted wide-area survey device simultaneously collects macroscopic image data and shallow ground-penetrating radar reflection data of the road surface. Texture recognition and crack segmentation are performed on the macroscopic image data to calculate the area ratio and distribution density of visible defects on the road surface. The area ratio is the ratio of the total defect area to the assessment area, and the distribution density is the number of defects per unit area. Based on the area ratio and distribution density, a road surface integrity index is generated. The calculation of the road surface integrity index can be expressed as:

[0102] ;

[0103] in: Indicates the road table integrity index. Indicates the area percentage. Indicates the distribution density. and These are predefined weighting coefficients. Layer analysis and dielectric constant inversion are performed on shallow ground-penetrating radar reflection data to identify the clarity of interlayer interfaces and material homogeneity of the roadbed structure. Based on this clarity and homogeneity, a structural layer health index is generated. The calculated road surface integrity index and structural layer health index are input into a pre-trained classification decision tree model. The model outputs a recommended detection grid density level and preloaded signal intensity range for the current road segment under test. The system automatically divides the spacing and number of sub-grid points in the detection area grid according to the recommended detection grid density level, and sets a controllable initial amplitude range for the preloaded signal based on the recommended preloaded signal intensity range.

[0104] In some embodiments, the construction and workflow of the pre-trained classification decision tree model includes specific operations. During the model construction phase, a large amount of historical road segment survey data is collected as training samples. Each training sample contains the road surface integrity index, structural layer health index, and the optimal detection grid density and optimal preloading intensity calibrated by expert experience for the corresponding road segment. Using the road surface integrity index and structural layer health index as decision features, and the optimal detection grid density and optimal preloading intensity as decision objectives, the classification decision tree model is recursively constructed using information gain or Gini impurity as the splitting criterion. In some embodiments, during the model operation phase, the road surface integrity index and structural layer health index calculated for the current road segment are input from the root node of the decision tree. At each decision node, the input data is allocated to different child nodes according to the feature judgment rules of the current node, and the process is passed down the decision path until the leaf node is reached. The leaf node stores the recommended detection grid density level and preloading signal intensity range. Optionally, the system has a built-in feedback calibration mechanism that, after each full-process detection based on recommended parameters, uses key response features discovered during the actual detection process, including the number of abnormal sensitive areas and the frequency of load adjustments, as a posterior evaluation index. Optionally, the feedback calibration mechanism performs a correlation analysis between the posterior evaluation indicators and the initial recommended parameters of the decision tree model. If a systematic bias is found, it triggers fine-tuning of the model parameters. This fine-tuning updates the judgment thresholds of relevant decision nodes to make the model recommendations more aligned with actual detection needs. It can be understood that the road surface condition pre-assessment stage provides preliminary parameter guidance for subsequent detailed detection.

[0105] See Figure 5This is an analysis diagram of the design load sequence and the detection scan path. This diagram illustrates the synchronous operation of the design load sequence and the detection scan path during highway pavement compressive strength testing, visually presenting the load application strategies and scan speed control logic for different areas. In the stable region, the load is maintained at a base load of 100kN for 0–2.5 seconds, preparing for a step-by-step increase. From 2.5–5 seconds, it linearly increases to 120kN and is briefly held, reflecting the "step-by-step increase" load scheme in the stable region. From 5–7.5 seconds, it increases again to 140kN, and then stabilizes at 160kN from 7.5–9 seconds, completing the step-by-step load application. In the sensitive region, the load oscillates at a high frequency within the 140–160kN range throughout, reflecting the "oscillating probing" load strategy used to evaluate the pavement response to small load fluctuations. The rapid relaxation cluster scan speed fluctuates between 10–14cm / s, maintaining a high level overall, consistent with the "continuous scan path" design, and is positively correlated with the average deformation recovery speed. The slow relaxation cluster scan speed fluctuated significantly between 3 and 7 cm / s and decreased significantly at multiple time points, reflecting the characteristics of the "intermittent pause scan path" and staying longer at sampling points with slow deformation recovery.

[0106] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for testing the compressive strength of highway pavement, characterized in that, The method includes: A detection area grid is set on the road surface to be tested, and a controllable preload signal is applied to each sub-grid point in the detection area grid; In response to the preload signal, the initial road surface features fed back by each sub-grid point in the detection area grid are obtained, and the initial road surface features include stress wave propagation spectrum and deformation relaxation time series; Based on the attenuation law of the peaks in the stress wave propagation spectrum, a design load sequence is constructed, and based on the fluctuation trend of the deformation relaxation time sequence, a detection scanning path is constructed. The design load sequence and the detection scanning path are run synchronously to obtain the road surface dynamic response data corresponding to the synchronous operation process; The dynamic response data of the road surface is iteratively compared with a preset response pattern library, and the target response pattern that matches the current road surface structure is identified in the response pattern library. Based on the load distribution rules and deformation constraints in the target response mode, load reconstruction instructions for adjusting the design load sequence and path optimization instructions for adjusting the detection scanning path are generated. The load reconstruction command and the path optimization command are executed to correct the compressive strength testing process of the detection area mesh.

2. The method for testing the compressive strength of highway pavement according to claim 1, characterized in that, Obtain the initial road surface features fed back by each sub-grid point in the detection area grid. Includes the following operations: A distributed pressure pulse generator is used to sequentially apply a set of pressure pulses with progressively increasing values ​​to each sub-grid point in the detection area grid. Using a multi-axis high-resolution sensor array, the full-frequency vibration response signal of each sub-grid point after receiving the pressure pulse is captured, and the vibration response signal is encoded into a vibration energy spectrum. The vibration energy spectrum is subjected to modal decomposition to extract the frequency components and phase information corresponding to the dominant modes, and the frequency components and phase information are combined to generate the stress wave propagation spectrum. Simultaneously, using a laser interferometric ranging device, the time interval sequence required for the surface deformation of each subgrid point to recover to its initial state after the pressure pulse stops is measured, and the time interval sequence is recorded as the deformation relaxation time sequence; A spatiotemporal correlation mapping is established between the stress wave propagation spectrum and the deformation relaxation time series, so that the wave propagation characteristics and deformation recovery characteristics of each sub-grid point form a one-to-one corresponding feature pair within the grid of the detection area.

3. The method for testing the compressive strength of highway pavement according to claim 2, characterized in that, Constructing the design load sequence and detection scan path includes the following operations: The stress wave propagation spectrum is segmented into regions to identify stable regions where the peak attenuation rate is below a standard threshold, and sensitive regions where the peak attenuation rate is above a standard threshold. A stepped increasing load application scheme is configured for the stable region, wherein the amplitude increment of each load step in the load application scheme is calculated proportionally based on the energy retention rate of the wave peak within the stable region. An oscillating test load application scheme is configured for the sensitive area. The load amplitude in the load application scheme fluctuates slightly around a reference value. The reference value is reverse-calibrated based on the average attenuation rate of the wave peaks in the sensitive area. The stepped incremental load application scheme for the stable region and the oscillating trial load application scheme for the sensitive region are sorted and connected according to their spatial positions in the grid of the detection area, and integrated to generate the design load sequence. Cluster analysis was performed on the deformation relaxation time series to divide it into fast relaxation clusters and slow relaxation clusters based on the density of time intervals. A continuous scanning path is planned for the fast relaxation cluster, and the moving speed of the continuous scanning path is positively correlated with the average speed of deformation recovery within the cluster; Plan an intermittent pause scan path for the slow relaxation cluster, wherein the duration of the intermittent pause scan path at each sampling point within the cluster is proportional to the deformation relaxation time of the sampling point; The continuous scanning path of the fast relaxation cluster and the intermittent pause scanning path of the slow relaxation cluster are spatially fused and temporally coordinated to generate the detection scanning path covering the entire detection area grid.

4. The method for testing the compressive strength of highway pavement according to claim 3, characterized in that, The design load sequence and the detection scanning path are run synchronously to obtain the road surface dynamic response data corresponding to the synchronous operation process, including the following operations: The load execution unit is activated, and loads are applied sequentially to the corresponding sub-grid points according to the load application scheme specified in the design load sequence. Simultaneously, the path scanning unit is activated, driving the detection probe to move strictly following the detection scanning path, and high-frequency data acquisition is performed at each position point or dwell time period specified by the path. During each collaborative working cycle of the load execution unit and the path scanning unit, three sets of key data are recorded: first, the peak value of transient compressive stress at the loaded subgrid point at the moment of load application; second, the strain distribution cloud map of adjacent points with conductive correlation to the load application position captured by the detection probe on the scanning path; and third, the energy dissipation curve of the entire process from the start of load application to the strain distribution stabilizing. The transient compressive stress peak, strain distribution cloud map and energy dissipation curve recorded in each collaborative work cycle are timestamped and encapsulated to form an independent dynamic response data package. According to the execution order of the design load sequence and the detection scanning path, the dynamic response data packets generated by all collaborative work cycles are concatenated to form the complete road dynamic response data.

5. The method for testing the compressive strength of highway pavement according to claim 4, characterized in that, The process involves iteratively comparing the road surface dynamic response data with a preset response pattern library to identify target response patterns that match the current road surface structure. Includes the following operations: From the multiple candidate response modes pre-stored in the response mode library, a standard feature template for each candidate response mode is extracted. The standard feature template includes a standard stress peak range, a standard strain cloud shape, and a standard energy curve envelope. From the currently acquired pavement dynamic response data stream, the measured transient compressive stress peak sequence, the measured strain distribution cloud map sequence, and the measured energy dissipation curve sequence are extracted. Perform the first round of feature comparison: calculate the overall distribution of the measured transient compressive stress peak sequence and the degree of overlap with the standard stress peak interval of each candidate response mode; Only candidate response patterns with an overlap exceeding a preset overlap threshold are retained for the next round of comparison; A second round of morphological comparison is conducted: the spatiotemporal evolution characteristics of the measured strain distribution cloud map sequence are compared with the standard strain cloud map morphology of the candidate response modes retained in the previous round to perform structural similarity analysis. Only candidate response patterns with structural similarity exceeding a preset similarity threshold are retained for the final round of comparison; Perform the final round of envelope comparison: match the overall trend of the measured energy dissipation curve sequence with the standard energy curve envelope of the candidate response modes retained in the previous round to ensure trend consistency. From the final round of comparison, the candidate response mode with the highest trend consistency matching degree is selected and determined as the target response mode that matches the current road structure; If, after multiple rounds of comparison, no candidate response pattern simultaneously meets all threshold conditions, the pattern learning mechanism is triggered to summarize the current road dynamic response data stream and its characteristics into a new response pattern and store it in the response pattern library.

6. The method for testing the compressive strength of highway pavement according to claim 5, characterized in that, Based on the load distribution rules and deformation constraints in the target response mode, load reconstruction instructions for adjusting the design load sequence and path optimization instructions for adjusting the detection scanning path are generated, including the following operations: The load distribution rules encapsulated in the target response mode are analyzed, and the load distribution rules define the ideal load spectrum that different types of road surface areas should bear; The load spectrum actually generated during the execution of the design load sequence is compared point by point with the ideal load spectrum defined by the target response mode to identify the load application points with significant deviations, as well as the amount and direction of the deviations. For each load application point with significant deviation, a specific load correction item is generated. The load correction item includes the location identifier of the load application point, the suggested adjusted load amplitude, and the suggested load application duration. All load correction items are compiled in the order of their execution in the design load sequence to generate the load reconfiguration instruction; Simultaneously, the deformation constraint conditions encapsulated in the target response mode are analyzed. These deformation constraint conditions specify the allowable gradient range of road surface deformation in space and the evolution rhythm in time under ideal response. The measured deformation data collected during the execution of the detection scanning path is compared with the deformation constraint conditions in two dimensions: spatial gradient and temporal evolution, to locate the path segment or sampling point in the scanning path that causes the measured deformation data to exceed the allowable range. For each identified problematic path segment or sampling point, a specific path adjustment item is generated. The path adjustment item includes adjusting the scanning speed, extending or shortening the dwell time at a specific point, or fine-tuning the scanning trajectory to avoid abnormal deformation areas. All path adjustment items are integrated according to their logical order in the detection scan path to generate the path optimization instruction.

7. The method for testing the compressive strength of highway pavement according to claim 6, characterized in that, Executing the load reconstruction command and the path optimization command to correct the compressive strength testing process of the detection area mesh includes the following operations: The load execution unit receives and parses the load reconstruction instruction, and extracts all load correction items from it; For each load correction item, the load execution unit updates the target load parameters of the corresponding load application point in its internal state table. The target load parameters include amplitude and duration. In subsequent detection cycles, when the load execution unit needs to apply a load again at the location point specified by the corresponding load correction item, it will call the updated target load parameters instead of the parameters in the original design load sequence, thereby realizing real-time load correction. The path scanning unit receives and parses the path optimization instruction, and extracts all path adjustment items from it; For each path adjustment item, the path scanning unit modifies the motion parameters of the corresponding path segment or sampling point in its internal path planner. The motion parameters include speed, dwell time, or coordinates. In subsequent detection cycles, the path scanning unit drives the detection probe to operate according to the modified motion parameters, thereby achieving dynamic optimization of the scanning path; After the load execution unit and the path scanning unit complete the instruction-based correction, the system re-enters the data acquisition state to acquire a new round of dynamic road surface response data after process correction, for subsequent analysis and evaluation.

8. The method for testing the compressive strength of highway pavement according to claim 7, characterized in that, Before setting up a detection area grid on the road surface to be tested and applying a controllable preload signal to each sub-grid point in the detection area grid, a road surface condition pre-assessment stage is also included, which includes the following steps: Using vehicle-mounted wide-area survey equipment, the system cruises at low speed along the road section to be surveyed, simultaneously collecting macroscopic image data of the road surface and shallow ground-penetrating radar reflection data. The macroscopic image data is subjected to texture recognition and crack segmentation to calculate the area ratio and distribution density of visible defects on the road surface, and a road surface integrity index is generated. The geological radar reflection data is analyzed for stratigraphic position and dielectric constant is inverted to identify the clarity of interlayer interfaces and material homogeneity of the subgrade structure, and to generate a structural layer health index. The road table integrity index and the structural layer health index are input into a pre-trained classification decision tree model; The classification decision tree model outputs a recommended detection grid density level and preloaded signal strength range for the current road segment under test based on the input double exponential combination. Based on the recommended detection grid density level, the system automatically divides the spacing and number of sub-grid points in the detection area grid; The system sets the initial amplitude range of the controllable preload signal based on the recommended preload signal strength range.

9. The method for testing the compressive strength of highway pavement according to claim 8, characterized in that, The construction and workflow of the pre-trained classification decision tree model. Includes the following operations: During the model building phase, a large amount of historical road survey data was collected as training samples. Each training sample includes the road surface integrity index, structural layer health index, and the optimal detection grid density and optimal preloading intensity calibrated by expert experience for the corresponding road segment. Using the road surface integrity index and structural layer health index as decision features, the optimal detection grid density and optimal preloading intensity as decision objectives, and information gain or Gini impurity as splitting criteria, the classification decision tree model is recursively constructed. During the model working phase, the road surface integrity index and structural layer health index calculated for the current road segment are input starting from the root node of the decision tree; At each decision node, the input data is distributed to different child nodes according to the feature judgment rules of the current node; The decision is passed down the decision path until it reaches the leaf node, where the recommended detection grid density level and the preloaded signal strength range are stored. The system has a built-in feedback calibration mechanism: after each full-process detection based on the recommended detection grid density level and the preloaded signal strength range, the key response features found in the actual detection process, including the number of abnormal sensitive areas and the frequency of load adjustment, are used as a post-hoc evaluation index. The correlation analysis is performed between the post-abstract evaluation indicators and the initial recommended parameters of the decision tree model. If a systematic bias is found, the model parameters are fine-tuned, and the judgment thresholds of the relevant decision nodes are updated so that the model recommendations are more in line with the actual detection needs.

10. A road pavement compressive strength testing system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for testing the compressive strength of a highway pavement as described in any one of claims 1 to 9.

Images