Survey and prediction method and system based on data processing in highway construction

Abstract

The invention relates to the technical field of data processing, in particular to a survey and prediction method and system based on data processing in highway construction, a total station and an electronic level are used for surveying a site, a soil sample is collected, the moisture content, density and grain composition of the soil sample are analyzed, and a soil characteristic database is established. In the construction process, a land leveler with an automatic leveling function is selected. When the gravel or pebble water storage cushion layer is laid, the volume is calculated according to design and material characteristics, and the thickness is determined. And after the permeable pavement slab is laid, hole cleaning, gap treatment and other operations are carried out. After construction, microseismic sensors are arranged on the periphery of the road surface, micro cracks are positioned, and the development trend of the micro cracks is judged; constructing a big data monitoring system, burying a sensor in a pavement structure layer, and collecting performance data under traffic load and environment change; and a dynamic correlation model is established, detection data at different time points and actual pavement performance are accurately predicted, and a basis is provided for pavement maintenance.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a survey and prediction method and system based on data processing in highway construction. Background Art

[0002] With the acceleration of the urbanization process and the continuous growth of transportation demands, highways, as an important part of infrastructure, have increasingly higher requirements for construction scale and quality. However, a series of intractable problems have emerged in the actual application of current highway construction technologies.

[0003] In large-scale highway construction projects, it is difficult to always maintain the operation accuracy of construction equipment at an ideal state. Coupled with the large natural differences in the softness of the subgrade soil at the construction site, it is difficult to effectively control the pavement smoothness, and local over-standard problems are common. Poor pavement smoothness not only reduces the driving stability and comfort of vehicles, but also causes uneven pavement stress, accelerates pavement damage, and significantly increases subsequent maintenance and repair costs.

[0004] In the quality inspection stage after construction, the shortcomings of existing technical means become more prominent. The currently widely used ground-penetrating radar and ultrasonic non-destructive testing technologies have inherent defects in resolution and are difficult to accurately detect tiny structural cracks inside the pavement and subtle property changes inside the materials. Such insufficient detection capabilities are likely to lead to misjudgments of the true structural performance of the pavement, posing potential safety hazards for subsequent road use. In addition, the data obtained from current detections only reflect the pavement structural performance at a specific time point. It is extremely difficult to accurately correlate these discrete data with the actual performance changes and potential risks of the pavement caused by various factors such as traffic loads and environmental changes during long-term use. This undoubtedly poses a huge challenge to scientifically and efficiently carrying out pavement maintenance and management work, making it impossible to timely and accurately formulate targeted maintenance strategies to ensure the long-term stable operation of highways.

[0005] Facing the urgent need for high-quality and long-life highways in modern society, traditional highway construction survey and prediction methods have become inadequate. Summary of the Invention

[0006] To solve the problems mentioned in the above background art, the present invention provides a survey and prediction method and system based on data processing in highway construction.

[0007] The survey and prediction method and system based on data processing in highway construction provided by the present invention adopt the following technical solutions: A survey and prediction method based on data processing in highway construction, comprising the steps: Survey the construction site and establish a soil property database; Conduct performance surveys on the materials selected for the permeable road; Calibrate and debug the grader, carry out layered leveling operations on the base course, move forward at a constant speed during the leveling process, and at the same time use the laser flatness monitoring system installed on the grader to monitor the flatness situation in real time, and conduct manual fine leveling on local areas with excessive standards; Ram the base course, determine the number of compaction passes, the traveling speed of the roller, and the vibration frequency according to the soil type, equip the roller with a real-time compaction degree monitor, continuously collect compaction degree data, and when the compaction degree does not meet the standard, adjust the compaction parameters in time and conduct compaction operations again; Calculate the laying volume of the required crushed stones or pebbles and determine the laying thickness; Lay and vibrate the concrete densely. When laying the concrete bonding layer above the crushed stone or pebble water storage cushion layer, use a paver equipped with a laser positioning and leveling system to lay it evenly; use a vibrator to vibrate the concrete densely; Lay the prefabricated permeable road surface panels on the concrete bonding layer according to the designed arrangement method; Carry out hole cleaning operations by the operator; After the concrete bonding layer has finally set, first wash the gaps between adjacent permeable road surface panels, then blow dry and clean them again; pour fine aggregate concrete into the gaps; After pouring is completed, carry out jointing treatment on the gaps; Monitor in real time the microseismic signals generated by the occurrence of microcracks and internal material changes during the use of the road surface, locate the positions of the microcracks and judge their development trends; Build a big data monitoring system for the structural performance of the road surface, and collect the structural performance data of the road surface under the influence of various factors such as traffic load and environmental changes in real time; Establish a dynamic correlation model for the structural performance of the road surface changing with multiple factors such as time, load, and environment, and predict the detection data at different time points and the actual performance data during the use of the road surface through this model.

[0008] Furthermore, conducting performance surveys on the materials selected for the permeable road includes: for the permeable road surface panels, use an electron microscope to analyze their microstructures and detect pore connectivity; for crushed stones or pebbles, use a laser particle size analyzer to accurately measure the particle size distribution, and for the concrete bonding layer and fine aggregate concrete, use a rheometer to monitor their rheological properties. According to the test results, determine the best use parameters and mixing ratios of each material and build a material property database.

[0009] Furthermore, calculating the required laying volume of crushed stones or pebbles includes determining the design thickness requirements of the crushed stone or pebble water storage cushion corresponding to different functional areas, accurately measuring the porosity and particle size distribution of the selected crushed stones or pebbles; calculating the volume based on the determined design thickness of each area and the corresponding laying area, and laying according to the calculated volume, while monitoring in real time and determining the final laying thickness.

[0010] Furthermore, during the hole cleaning process, use an electronic depth measuring instrument to monitor the hole cleaning depth in real time to ensure that the permeable holes penetrate the concrete bonding layer and extend to the crushed stone or pebble water storage cushion; for each completed hole cleaning operation, enter the detailed information of the hole cleaning depth, hole diameter size, perpendicularity of the permeable hole, and the corresponding permeable road surface panel number into the hole cleaning quality file.

[0011] Furthermore, the specific steps for locating the position of microcracks and judging their development trend are as follows: Based on the hyperbola positioning principle, locate the position of microcracks. Establish an equation set through the elastic wave propagation velocity and arrival time difference, and use the Newton-Raphson iteration method. Set the initial estimated value to reduce the number of iterations, and calculate the arrival time difference to establish an equation set to locate the coordinates of the crack source; The specific method for judging the development trend is to extract various characteristic parameters from the microseismic signals, construct a prediction model based on support vector regression, model the extracted characteristic parameters, select the radial basis kernel function, and optimize the kernel parameters through the cross-validation method combined with grid search. Input the characteristic parameters of the real-time collected microseismic signals into the model to predict the crack development.

[0012] Furthermore, establish a dynamic correlation model for the change of pavement structure performance with time, load, and environment. Specifically, based on the unified fusion of LSTM and random forest, establish a dynamic correlation model for pavement structure performance. Specifically: Construct the LSTM branch: Input the time stamps collected at fixed time intervals, as well as the pavement performance and environmental data of strain, temperature, and humidity changing with time, and the time series data of traffic loads with different traffic volumes at different time periods. Organize them into a sequence and input them in chronological order; Establish input gates, forget gates, and output gates to control the information flow; the input gate determines the input of new information, the forget gate controls the retention or discard of past memories, and the output gate determines the final output; Establish an output layer to integrate and output the learned time series features. The output feature vector reflects the change trend and characteristics of pavement performance in the time dimension, providing time-related features for fusion with random forest; Construct the random forest branch: Input the static features that are not time series and the time features output by LSTM; the static features include pavement structure parameters, once-determined environmental features, and statistical features extracted from time series; Construct multiple decision trees of the random forest. When constructing, randomly select some features from the input features and sample data, so that each decision tree learns the relationship between different feature combinations and pavement performance, improving the diversity of the model; Concatenate the time feature vector output by the LSTM with the static and statistical features input to the random forest to form a comprehensive feature vector, comprehensively reflecting the factors affecting pavement performance; Input the fused feature vector into the comprehensive model for training. By adjusting the model parameters, using the gradient descent algorithm, update the parameters according to the error between the predicted value and the true value; The training data is respectively input into the LSTM and random forest branches. The LSTM calculates the time features, the random forest processes the static and statistical features, and then fuses the features for forward propagation to obtain the prediction result; According to the prediction result and the true pavement performance data, calculate the value of the loss function; Calculate the gradient of the loss function with respect to the model parameters through backpropagation. Use the optimizer to update the parameters according to the gradient, reduce the value of the loss function, and repeat the steps of forward propagation, calculating the loss, and backpropagation to update the parameters until the loss function converges or reaches the preset number of training epochs.

[0013] A survey and prediction system based on data processing in highway construction includes a pre-construction data collection and analysis unit, a construction process monitoring and control unit, and a post-construction quality inspection and maintenance management unit. The post-construction quality inspection and maintenance management unit includes: Microseismic sensors, wireless transmission modules, and data acquisition systems. The microseismic sensors, wireless transmission modules, and data acquisition systems are used to monitor the microseismic signals of the pavement, locate the crack positions, and predict the crack development trends.

[0014] The post-construction quality inspection and maintenance management unit includes: strain gauges, temperature sensors, humidity sensors, wireless transmission networks, and servers. The strain gauges, temperature sensors, humidity sensors, wireless transmission networks, and servers are used to collect the pavement structure performance data and provide data for performance analysis.

[0015] The beneficial technical effects of the present invention: Through the calibration of total stations and levels, record the topography, geomorphology, and soil conditions in detail, carry out multi-point and multi-level soil sampling and analysis, accurately measure soil moisture content, density, particle size distribution and other indicators, and construct a comprehensive soil property database. At the same time, use electron microscopes, laser particle size analyzers, and rheometers to conduct performance surveys on permeable pavement slabs, gravel or pebbles, concrete bonding layer materials, fine aggregate concrete, etc., clarify parameters such as microstructure, particle size distribution, and rheological properties, and build a material performance database. These preliminary works lay a solid foundation for the accurate formulation of construction plans, effectively avoid construction problems caused by missing or inaccurate basic data, and improve the project quality from the source.

[0016] For graders, comprehensively calibrate and debug their blade angle adjustment devices, traveling systems, and hydraulic systems. During the base layer leveling operation in layers, with the help of a laser flatness monitoring system, conduct high-density monitoring at a monitoring point every 0.2 meters to ensure precise control of the base layer flatness. For rollers, accurately determine the number of compaction passes, traveling speed, and vibration frequency based on the soil type, and continuously collect compaction degree data through the equipped real-time compaction degree monitor to promptly adjust the compaction parameters.

[0017] Optimize the design and construction of the drainage system: According to different functional areas, such as vehicle lanes, bicycle lanes, and leisure squares, clarify the design thickness requirements of the gravel or pebble water storage cushion layer, and accurately calculate the laying volume in combination with the porosity and particle size distribution of the material. During the hole cleaning operation, use an electronic depth measuring instrument to monitor the hole cleaning depth in real time to ensure that the permeable holes penetrate the concrete bonding layer and extend to the gravel or pebble water storage cushion layer. The drainage system constructed in this way is efficient and reliable, can quickly drain the road surface water, reduce the driving safety risks caused by water accumulation, and at the same time protect the roadbed from water damage and maintain the long-term stability of the road structure.

[0018] Scientifically arrange microseismic sensors around the road surface according to the importance and stress characteristics of different areas, and focus on densifying the sensors at stress concentration parts such as curves and intersections, as well as areas at the edge that are greatly affected by the outside. Through the hyperbola positioning principle and the Newton-Raphson iteration method, quickly locate the positions of microcracks, and at the same time extract various characteristic parameters from the microseismic signals, and use the support vector regression model to predict the crack development trend.

[0019] Improve the intelligent level of road surface performance management: Build a big data monitoring system for road surface structure performance, bury various sensors such as strain gauges, temperature sensors, and humidity sensors in different structural layers such as the base layer, concrete bonding layer, and permeable road surface slab, and collect road surface structure performance data affected by multiple factors such as traffic load and environmental changes at a high frequency of once every 5 minutes. Use the LSTM and random forest fusion model to accurately correlate and predict the detection data and actual performance data at different time points, provide a scientific and systematic decision-making basis for the road management department, help formulate intelligent and refined maintenance plans and traffic management strategies, and improve the overall efficiency and scientific nature of highway operation management. Specific implementation manners

[0020] An embodiment of the present invention discloses a surveying and predicting method based on data processing during highway construction, including steps: data collection and processing before construction, surveying the construction site, and establishing a soil property database; Work is carried out using calibrated total station and electronic level measuring instruments. During the exploration process, in addition to paying attention to the overall macro-topography of the site (such as whether there are slopes, low-lying areas, etc.), the geomorphic features should also be carefully recorded, such as whether there are rock outcrops and special geological structures within the site. For sloping areas, the slope angles at different positions should be accurately measured, and measurements and records should be made at regular intervals (such as every 5 meters). At the same time, by drawing a simple topographic sketch, the positions of key measurement points should be marked to provide rich basic data for subsequent drawing of accurate construction site plans and profiles. For soil conditions, multi-point and multi-level soil sampling and analysis are carried out. In the layout of sampling points, the entire construction site should be covered, and sampling points should be distributed in a grid pattern. For example, one sampling point is set every 100 square meters, and soil samples should be collected from different depths (such as the surface layer of 0-20 cm, the middle layer of 20-40 cm, the deep layer of 40-60 cm, etc.) at each sampling point. The drying method is used to determine the water content of the soil, the density bottle method is used to accurately measure the soil density, and the sieve analysis method and hydrometer method are combined to determine the particle size distribution index of the soil. All analysis data should be detailedly recorded in a special table to prepare for establishing a soil property database, and a special person should be arranged to review the data to prevent input errors.

[0021] A survey and prediction method based on data processing in highway construction, including the steps of: carrying out performance surveys on the selected materials of permeable road surface slabs, gravel or pebbles, concrete bonding layer materials, and fine aggregate concrete. For permeable road surface slabs, an electron microscope is used to analyze their microscopic structure and detect pore connectivity; for gravel or pebbles, a laser particle size analyzer is used to accurately measure the particle size distribution. For the concrete bonding layer and fine aggregate concrete, a rheometer is used to monitor their rheological properties. According to the test results, the best usage parameters and mixing ratios of each material are determined to construct a material performance database. Performance survey of permeable road panels. Reasonably determine the sample quantity and extraction method according to the project scale and panel production batches. For example, when there are multiple production batches, 2 batches can be extracted for every 5 batches, and 6 - 8 panels can be extracted from each batch. At the same time, products under different production times and process conditions should be taken into account to ensure the representativeness of the samples. When using cutting equipment to cut samples, cutting parameters need to be accurately set according to factors such as panel thickness. After cutting, use special cleaning tools and mild cleaning agents to remove surface impurities, oil stains, etc. Then, use fine grinding equipment to polish the sample surface to an appropriate roughness in multiple steps to facilitate observation under an electron microscope. After placing the sample on the electron microscope stage, first preset a rough range of microscope parameters according to the characteristics of the sample, and then gradually fine-tune to the optimal observation state. The observation should not only focus on pore connectivity but also pay attention to details such as pore shape and distribution pattern, and classify and record the microscopic structure characteristics of different regions. Systematically compare the microscopic structure data with the actual performance indicators in multiple aspects such as the water permeability rate, abrasion resistance, and weather resistance of the panel. For example, analyze the changes in a panel with good pore connectivity but irregular pore shape during long-term water permeation. Through a large number of comparative analyses, accurately lock the range of microscopic structure parameters that make each performance reach an optimal state, and then determine the best parameters in terms of material selection, production process, etc., and enter them into the database.

[0022] Performance survey of gravel or pebbles. Develop a sampling plan considering factors such as stone source, batch, stacking form, etc. Set different sampling point densities for regions with obvious differences in particle size and quality. For example, increase the sampling points appropriately in regions with large quality fluctuations. Make marks and records during collection to ensure the traceability of the samples. Manually sort out the parts that obviously do not meet the requirements from the collected samples first. Then, remove impurities thoroughly with an appropriate water flow rate, flushing duration, and stirring method. After washing, place them in a standard environment with temperature and humidity control and good ventilation to dry, avoiding interference from the external environment on the stone characteristics. When pouring the dried samples into the sample cell of the laser particle size analyzer, prevent local accumulation from affecting the test results. After completing the calibration according to the standard operation procedure of the instrument, reasonably select the built-in test mode and parameters of the instrument according to factors such as the estimated particle size range of the sample, and take the average value after multiple measurements. Record in detail the proportion of particles in each particle size range and other data, and draw a clear and intuitive particle size distribution chart for subsequent analysis. Combine the particle size distribution data with other performance indicators such as the hardness and abrasion rate of the stone, and comprehensively consider the requirements for functions such as cushion drainage and anti-deformation in different construction scenarios. For example, in areas with heavy traffic and a lot of rain, stones with moderate particle size, uniform distribution, and high hardness need to be selected. Through the analysis and summary of the feedback from multiple actual engineering applications, determine the specific usage parameters and enter them into the database.

[0023] When mixing concrete, according to the performance characteristics of the mixer and the requirements of the concrete mix ratio, accurately set parameters such as the mixing time and rotation speed at each stage. During the mixing process, regularly check the mixing state of the concrete to ensure that there are no problems such as caking and unevenness. During the testing process, strictly operate according to the preset parameters such as the time interval of shear rate change and the stabilization duration, observe and record the details of the rheological state change of the specimen in real time, draw curves based on the rheological data and conduct in-depth analysis to judge whether the construction workability of the concrete meets the requirements of on-site paving, vibration and other processes. In addition to rheological properties, comprehensively carry out tests on performance indicators such as the air content and impermeability of the concrete. Compare and analyze the mutual relationship between various performance indicators under different mix ratios and their influence on the overall performance of the road surface.

[0024] A survey and prediction method based on data processing in highway construction, including the steps of: calibrating and debugging a grader based on data processing construction, carrying out layered leveling operations on the base course, moving forward at a constant speed during the leveling process, and at the same time using a laser flatness monitoring system installed on the grader to set a monitoring point every 0.2 meters to monitor the flatness situation in real time, manually fine-leveling local areas with excessive standards, and using a scraper to process uneven places to meet the requirements. A survey and prediction method based on data processing in highway construction, including the steps of: compacting the base course, determining the number of compaction passes, the traveling speed of the roller, and the vibration frequency according to the soil type, equipping the roller with a real-time compaction degree monitor to continuously collect compaction degree data, and when the compaction degree does not meet the standard, adjusting the compaction parameters in time and performing the compaction operation again; before construction, arranging geological exploration personnel to conduct a comprehensive inspection and analysis of the soil at the construction site to clarify the soil type, such as clay, sandy soil or silty soil, etc. Different soil types have different required compaction passes due to their different characteristics such as particle composition and cohesion. For clay, due to its fine particles and large cohesion, the number of compaction passes is usually set to 9-10 times; while sandy soil has coarser particles and less cohesion, and generally 8-9 compaction passes are sufficient, so as to accurately determine the number of compaction passes and ensure the compaction effect. According to factors such as the model of the roller, the scale of the construction site, and the thickness of the base course, reasonably control the traveling speed of the roller. If the speed is too fast, the compaction time of the roller on the base course will be short, affecting the compaction quality; if the speed is too slow, the construction efficiency will be reduced. At the same time, according to the soil type and compaction requirements, set the vibration frequency at 30-35Hz to ensure that the vibration energy can be effectively transmitted to the base course soil, so that the soil particles can be better arranged and compacted. Equip the roller with a real-time compaction degree monitor (monitoring accuracy is ±0.5%). During the construction process, the operator closely monitors the data displayed on the monitor to ensure its normal operation and stable data transmission. This instrument can real-time feedback the compaction degree of different positions of the base course. Once every certain area (such as every 10 square meters) of the rolling area is passed, the compaction degree data is recorded to form a compaction degree data record ledger. Once it is detected that the compaction degree does not meet the standard, the operator immediately stops the current rolling operation and analyzes the reasons for non-compliance. If it is due to high soil moisture content, the construction can be suspended first, wait for the soil to dry naturally or take appropriate drainage measures to reduce the moisture content, and then re-perform the compaction operation according to the established traveling speed, vibration frequency and other parameters; if the number of compaction passes is insufficient, the number of compaction passes is increased accordingly until the compaction degree reaches more than 96%, ensuring that the compaction quality of the base course is uniform and stable, and laying a solid foundation for the subsequent processes.

[0025] A survey and prediction method based on data processing in highway construction, including the steps of: laying a gravel or pebble water storage cushion layer, accurately calculating the required laying volume of gravel or pebbles and determining the laying thickness according to the design thickness requirements and the porosity and particle size distribution data of the materials. First, clarify the design thickness requirements of the gravel or pebble water storage cushion corresponding to different functional areas (such as vehicle lanes, bicycle lanes, leisure squares, etc.). For example, due to the long-term heavy pressure of vehicles on the vehicle lane, the design thickness may be 20 - 25 cm; the load on the bicycle lane is relatively smaller, and the thickness may be 15 - 20 cm. At the same time, considering the drainage direction, slope, etc., the thickness of each part may vary slightly, and these specific values need to be carefully recorded as the basis for subsequent calculations. For the selected gravel or pebbles, accurately measure their porosity and particle size distribution.

[0026] In terms of porosity measurement, select an appropriate amount of representative material samples and use the common volume replacement method. For example, put the samples into a standard container filled with water and measure the volume of the drained water, and then calculate the porosity. To ensure accuracy, measure multiple times and take the average value. For particle size distribution detection, when using a laser particle size analyzer, first stir the samples evenly to remove large particle impurities, etc. Then, pour the samples steadily into the analyzer to obtain the proportion of particles in different particle size ranges (such as 3 - 6 mm, 6 - 10 mm, etc.), and draw a clear particle size distribution histogram. Calculate the theoretical volume based on the determined design thickness and corresponding laying area of each region. For example, if the area of a certain region is 80 square meters and the design thickness is 18 cm (converted to 0.18 m), then the theoretical volume is 80×0.18 = 14.4 cubic meters. Then adjust it in combination with the material porosity. Assuming the porosity is 20%, due to the existence of pores, the actual filling volume will increase, and the actual required volume is 14.4÷(1 - 20%) = 18 cubic meters. Calculate the actual required laying volume of each region according to this method. During actual laying, operate according to the calculated volume, and at the same time, monitor in real time and determine the final laying thickness.

[0027] A survey and prediction method based on data processing in highway construction, including the steps: laying and vibrating the concrete densely. When laying the concrete bonding layer above the gravel or pebble water storage cushion, use a paver equipped with a laser positioning and leveling system to lay it evenly; use a vibrator to vibrate the concrete densely; Preparation for laying the concrete bonding layer, select a paver equipped with a high-precision laser positioning and leveling system, and before construction, arrange technicians to conduct a full range of debugging on it. Focus on checking the accuracy of the laser positioning system so that it can accurately feedback the paving position and elevation information, and strictly control the error range to minimize it. At the same time, carefully check the sensitivity of the leveling system to ensure that it can flexibly and timely adjust the paving thickness according to the preset flatness requirements.

[0028] A survey and prediction method based on data processing in highway construction, including the steps: laying the precast permeable road surface panels on the concrete bonding layer according to the designed arrangement method; during laying, use an aluminum alloy level and a steel wire tensioning tool to ensure the flatness and uniform gap width of the permeable road surface panels; The hole cleaning operation is carried out by the operator. During the hole cleaning process, an electronic depth measuring instrument is used to monitor the hole cleaning depth in real time to ensure that the permeable hole penetrates the concrete bonding layer and extends to the gravel or pebble water storage cushion layer. For each completed hole cleaning operation, the detailed information such as the hole cleaning depth, hole diameter size, perpendicularity of the permeable hole, and the corresponding permeable road panel number is entered into the hole cleaning quality file. After the concrete bonding layer has finally set, first flush the gaps between adjacent permeable road panels, then dry and clean them once. Pour fine aggregate concrete into the gaps. After the pouring is completed, carry out caulking treatment on the gaps. Drill drainage holes at the caulking positions. The diameter, spacing, and depth of the drainage holes meet the requirements, and at the same time, record the position, diameter, and depth information of the drainage holes in detail. A survey and prediction method based on data processing in highway construction, including steps: quality inspection and data evaluation after construction, real-time monitoring of microseismic signals generated by the occurrence of microcracks and internal material changes during the use of the road surface, positioning the positions of microcracks and judging their development trends. Sensor layout: Layout microseismic sensors according to the importance and stress characteristics of different areas of the road surface. For example, at key parts such as road bends and intersections where stress concentration is likely to occur, appropriately increase the density of sensor layout to ensure that subtle changes can be captured sensitively. For the edge parts of the road, considering that they are relatively more affected by external factors, also reasonably increase the number of sensors. At the same time, combine past similar project experiences and the estimated traffic flow and load conditions of this section of the road, etc., to scientifically plan the overall distribution of sensors to ensure that there are no dead spots in monitoring and the key points are prominent. Before installation, strictly calibrate and test the performance of the sensors to ensure that their initial state is good and the measurement accuracy meets the standards. When installing, choose to operate on a sunny day with a dry road surface. Use an adhesive to firmly paste the sensors at the designated positions on the road surface that have been pre-cleaned and polished flat, and then use a sealant to seal the surrounding areas to prevent rainwater, dust, etc. from infiltrating and affecting the performance of the sensors. During the installation process, use tools such as a level to ensure that the installation angle of the sensors meets the requirements to ensure that they can accurately receive and transmit signals.

[0029] Positioning the positions of microcracks and judging their development trends: After arranging microseismic sensors around the road surface, when microcracks generate microseismic signals, the sensors convert the received elastic wave signals into electrical signals and transmit them to the data acquisition system. When positioning the positions of microcracks, based on the hyperbola positioning principle, use the time difference of arrival of signals received by multiple sensors to construct equations. Taking three sensors as an example, establish a system of equations through the elastic wave propagation speed and the time difference of arrival. Since it is complex to directly solve this system of equations, the Newton-Raphson iteration method is adopted, and the initial estimated values are set in combination with the actual situation of the road surface, such as referring to the positions where cracks have occurred on past similar road surfaces, to reduce the number of iterations and quickly and accurately locate the coordinates of the crack source.

[0030] To determine the development trend of microcracks specifically, multiple types of characteristic parameters are extracted from microseismic signals. When calculating the signal energy, time weighting is considered, and higher weights are assigned to recent signals to more accurately reflect the current energy change of the cracks. When determining the center frequency, time-frequency analysis methods such as the short-time Fourier transform are used to observe the change of the center frequency in different time segments. For the signal duration, the dynamic threshold method is adopted, and the threshold is calculated in real time based on the historical data and statistical characteristics of the signal to determine the start and end points of the significant signal fluctuations, improving the accuracy of judging the complexity of crack propagation.

[0031] A prediction model is constructed based on support vector regression to model the extracted characteristic parameters. A kernel function such as the radial basis kernel function is selected, and the kernel parameters are optimized by combining the cross-validation method with the grid search technique. The characteristic parameters of the microseismic signals collected in real time are input into the model to predict the development state of the cracks, such as the change amount of crack length, the change rate of width, etc., and the pavement maintenance strategy is adjusted in a timely manner according to the predicted values.

[0032] A survey and prediction method based on data processing in highway construction includes the steps of: constructing a big data monitoring system for pavement structural performance, arranging and burying sensors (such as strain gauges, high-sensitivity temperature sensors, humidity sensors) in different structural layers (base course, concrete bonding layer, permeable pavement slab) inside the pavement, and collecting the structural performance data (such as strain, temperature, humidity changes) of the pavement under the influence of various factors such as traffic load and environmental changes (such as temperature changes, humidity changes, groundwater level changes) in real time. The collection frequency is set to once every 5 minutes; Collect the structural performance data of each structural layer sensor collected at a frequency of once every 5 minutes under different traffic loads and environmental changes (such as seasonal alternation, rainfall and snowfall), and perform standardization processing on the data to unify the format, unit, etc., and remove obvious outliers to ensure the data quality and prepare for subsequent analysis. According to the pavement structure characteristics and performance influencing factors, key performance indicators such as the strain difference between different structural layers and the change rate of temperature and humidity are selected. These indicators can effectively reflect the health status of the pavement structure and serve as an important basis for constructing the model.

[0033] Establish a dynamic correlation model of pavement structural performance changing with multiple factors such as time, load, and environment, and predict the detection data at different time points and the actual performance data during the pavement use process through this model; To establish a dynamic correlation model of pavement structural performance changing with multiple factors such as time, load, and environment, specifically, a dynamic correlation model of pavement structural performance is established based on the unified fusion of LSTM and random forest. Specifically: Construct the LSTM branch: Input the time stamps collected at fixed time intervals (such as every 5 minutes), as well as the pavement performance and environmental data of strain, temperature, and humidity changing over time, and the traffic load time series data of traffic flow in different periods, and organize them into a sequence in chronological order for input; Establish input gates, forget gates, and output gates to control the flow of information; the input gate determines the input of new information, the forget gate controls the retention or discard of past memories, and the output gate determines the final output; for example, when analyzing the impact of pavement temperature on performance, remember the long-term changes in temperature and key impact points, thereby capturing the long-term dependencies of the time series and mastering the pattern of pavement performance changing over time; Establish an output layer to integrate and output the learned time series features. The output feature vector reflects the changing trends and characteristics of pavement performance in the time dimension, providing time-related features for fusion with random forests; Construct the branches of the random forest: Input the static features that are not time series and the time features output by the LSTM; the static features include pavement structure parameters (such as base thickness, material type), once-determined environmental features (such as geological conditions, initial groundwater level), and statistical features extracted from the time series (such as daily average traffic flow, monthly standard deviation of temperature); Construct multiple decision trees for the random forest. When constructing, randomly select some features and sample data from the input features so that each decision tree learns the relationship between different feature combinations and pavement performance, enhancing the diversity of the model; Concatenate the time feature vector output by the LSTM with the static and statistical features input to the random forest to form a comprehensive feature vector, comprehensively reflecting the factors affecting pavement performance; Input the fused feature vector into the comprehensive model for training. By adjusting the model parameters, using the gradient descent algorithm, and based on the error between the predicted value and the true value (mean squared error for regression, cross-entropy loss for classification), update the parameters to reduce the error and optimize the model performance; Model training and optimization: Set the initial parameters of the number of units in the LSTM hidden layer, the number of decision trees in the random forest, and the learning rate. For example, set 128 units in the LSTM hidden layer, 50 decision trees in the random forest, and the learning rate to 0.001; Input the training data into the LSTM and the random forest branches respectively. The LSTM calculates the time features, the random forest processes the static and statistical features, and then fuses the features for forward propagation to obtain the prediction results; According to the prediction results and the true pavement performance data, calculate the value of the loss function, such as the mean squared error loss function; Calculate the gradient of the loss function with respect to the model parameters through backpropagation. Use an optimizer (such as the Adam optimizer) to update the parameters according to the gradient, reduce the value of the loss function, and repeat the steps of forward propagation, calculating the loss, and backpropagation to update the parameters until the loss function converges or reaches the preset number of training epochs; Data involved in the data processing process: Time series data: Time stamps: Starting from the time when the road surface is built and put into use, accurately record the time points of each data collection at set intervals (such as every 5 minutes, every hour, every day); Related to road surface performance: Real-time collect performance indicators such as road surface strain, displacement, and crack width through sensors, as well as environmental factors such as temperature, humidity, and rainfall, and data on the change of traffic flow, vehicle type distribution, and axle load over time at different time periods; Static and statistical data: Road surface structure parameters: Collect parameters such as the thickness, material type, and porosity of each structural layer (base course, cushion course, surface course) of the road surface that basically remain unchanged after completion; Environmental static characteristics: Measure and record geological conditions, soil types, and initial groundwater level environmental information at the location of the road surface; Statistical characteristics: Calculate statistical characteristics from time series data, such as the daily average traffic flow, the monthly maximum temperature, and the quarterly change rate of humidity, to reflect the statistical laws of time series; Data division: Training set: Select 60%-70% of the data as the training set to train the fusion model and let the model learn the relationship between input features and road surface performance; Validation set: Divide 15%-20% of the data as the validation set to adjust hyperparameters during model training and evaluate the model performance under different hyperparameter combinations; Test set: The remaining 15%-20% of the data is used as the test set to evaluate the performance of the trained model on unseen data, and calculate accuracy, mean square error and other indicators to verify the generalization ability of the model.

[0034] This application also discloses a survey and prediction system based on data processing in highway construction, including a pre-construction data collection and analysis unit, a construction process monitoring and control unit, and a post-construction quality inspection and maintenance management unit. The pre-construction data collection and analysis unit includes: Total station, level and supporting computers, used to measure the topography and geomorphology of the site, obtain data such as slope and elevation, and draw topographic sketches and 3D models to provide a basis for construction planning.

[0035] Soil sampler, moisture content measuring instrument, densitometer, particle size distribution analyzer and data terminal, used to collect soil samples, measure moisture content, density, and particle size distribution, and establish a soil property database to provide a basis for construction.

[0036] Electron microscopes, laser particle size analyzers, and rheometers are used to detect the material properties of permeable road panels, crushed stones, etc., determine parameters such as microstructure, particle size distribution, and rheological properties, and construct a material property database.

[0037] The construction process monitoring and control unit includes: A grader with an automatic leveling function, a laser flatness monitoring system, a roller, a compaction degree monitor, and a central control platform are used to debug the parameters of the grader, real-time monitor the flatness of the base layer and the compaction degree, speed, and vibration frequency of the roller, and adjust automatically or manually to ensure that the quality of the base layer meets the standards.

[0038] The cushion laying system (including laying equipment and sensors), the concrete construction monitoring system (paver, vibrator, vibration sensor, mixing station, and weighing equipment) are used to calculate and control the laying volume and thickness of the cushion, accurately pave and vibrate the concrete bonding layer, and ensure the construction quality of each layer.

[0039] The laying positioning system (spirit level, wire-pulling tool, measuring equipment), the hole cleaning monitoring system (hole cleaning equipment, depth measuring instrument), the gap treatment and drainage hole construction system (high-pressure water gun, perfusion equipment, caulking tool, drilling equipment, and measuring sensor) are used to accurately lay the permeable road panel, control the hole cleaning depth, clean and treat the panel gap, and construct the drainage hole according to the standard.

[0040] The post-construction quality inspection and maintenance management unit includes: Microseismic sensors, wireless transmission modules, and data acquisition systems are used to monitor the microseismic signals of the road surface, locate the crack positions, and predict the crack development trend.

[0041] Strain gauges, temperature sensors, humidity sensors, wireless transmission networks, and servers are used to collect the road surface structure performance data, screen key indicators, and provide data for performance analysis.

[0042] The above are all the preferred embodiments of the present invention, and the protection scope of the present invention is not limited accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A survey and prediction method based on data processing in highway construction, characterized in that, Including the steps: Survey the construction site and establish a soil property database; Conduct performance surveys on the materials selected for the permeable road surface; Calibrate and debug the grader and carry out layered leveling operations on the base course; Compact the base course; Calculate the laying volume of the required gravel or pebbles and determine the laying thickness; Use a vibrator to vibrate and compact the concrete; Lay the precast permeable road surface panels on the concrete bonding layer according to the designed arrangement; The operator conducts hole cleaning operations; After the concrete bonding layer has finally set, first wash the gaps between adjacent permeable road surface panels, then dry and sweep them once; Pour fine aggregate concrete into the gaps; After pouring is completed, carry out joint sealing treatment on the gaps; Real-time monitor the microseismic signals generated by microcracks and internal material changes during the use of the road surface, locate the positions of the microcracks and judge their development trends; Construct a big data monitoring system for the structural performance of the road surface and collect the structural performance data of the road surface under the influence of various factors such as traffic load and environmental changes in real time; Establish a dynamic correlation model for the structural performance of the road surface changing with multiple factors of time, load, and environment. Through this model, predict the detection data at different time points and the actual performance data during the use of the road surface.

2. The survey and prediction method based on data processing during highway construction according to claim 1, wherein, Conducting performance surveys on the materials selected for the permeable road surface includes: for the permeable road surface panels, use an electron microscope to analyze their microstructures and detect pore connectivity; for gravel or pebbles, use a laser particle size analyzer to accurately measure the particle size distribution. For the concrete bonding layer and fine aggregate concrete, use a rheometer to monitor their rheological properties. According to the test results, determine the best usage parameters and mixing ratios of each material and construct a material performance database.

3. A survey and prediction method based on data processing in highway construction according to claim 1, characterized in that, Calculating the laying volume of the required gravel or pebbles includes determining the design thickness requirements of the gravel or pebble water storage cushions corresponding to different functional areas, accurately measuring the porosity and particle size distribution of the selected gravel or pebbles; calculating the volume based on the determined design thickness of each area and the corresponding laying area, lay according to the calculated volume, and at the same time, real-time monitor and determine the final laying thickness.

4. A survey and prediction method based on data processing in highway construction according to claim 1, characterized in that During the hole cleaning process, use an electronic depth measuring instrument to real-time monitor the hole cleaning depth to ensure that the permeable holes penetrate the concrete bonding layer and extend to the gravel or pebble water storage cushion; After each hole cleaning operation is completed, enter the detailed information of the hole cleaning depth, hole diameter size, perpendicularity of the permeable holes, and the corresponding permeable road surface panel numbers into the hole cleaning quality file.

5. A survey and prediction method based on data processing during highway construction according to claim 1, characterized in that, Locating the positions of the microcracks includes: locating the positions of the microcracks based on the hyperbola location principle, establishing a system of equations through the elastic wave propagation velocity and arrival time difference, using the Newton-Raphson iteration method, setting the initial estimate value, reducing the number of iterations, and calculating the arrival time difference to establish a system of equations to locate the coordinates of the crack source.

6. The survey and prediction method based on data processing during highway construction according to claim 5, characterized in that Judging the development trend specifically means extracting various types of characteristic parameters from the microseismic signals, constructing a prediction model based on support vector regression, modeling the extracted characteristic parameters, selecting a radial basis kernel function, and optimizing the kernel parameters through cross-validation combined with grid search. Input the characteristic parameters of the real-time collected microseismic signals into the model to predict the crack development.

7. A survey and prediction method based on data processing in highway construction according to claim 1, characterized in that Establishing a dynamic correlation model for the structural performance of the road surface changing with multiple factors of time, load, and environment is specifically based on the unified fusion of LSTM and random forest to establish a dynamic correlation model for the structural performance of the road surface. Specifically: Construct the LSTM branch: Input the time stamps collected at fixed time intervals, as well as the pavement performance and environmental data of strain, temperature, and humidity changing over time, and the traffic load time series data of vehicle flow in different periods. Organize them into a sequence in chronological order and input the sequence; Establish input gates, forget gates, and output gates to control the information flow; the input gate determines the input of new information, the forget gate controls the retention or discard of past memories, and the output gate determines the final output; Establish an output layer to integrate and output the learned time series features. The output feature vector reflects the change trend and characteristics of pavement performance in the time dimension, providing time-related features for fusion with the random forest; Construct the random forest branch: Input the static features that are not time series and the time features output by the LSTM; the static features include pavement structure parameters, once-determined environmental features, and statistical features extracted from the time series; Construct multiple decision trees of the random forest. When constructing, randomly select some features and sample data from the input features so that each decision tree learns the relationship between different feature combinations and pavement performance, improving the diversity of the model; Concatenate the time feature vector output by the LSTM with the static and statistical features input to the random forest to form a comprehensive feature vector, comprehensively reflecting the factors affecting pavement performance; Input the fused feature vector into the comprehensive model for training. By adjusting the model parameters, use the gradient descent algorithm to update the parameters according to the error between the predicted value and the true value; Input the training data into the LSTM and random forest branches respectively. The LSTM calculates the time features, and the random forest processes the static and statistical features, and then fuses the features for forward propagation to obtain the prediction result; According to the prediction result and the real pavement performance data, calculate the value of the loss function; Calculate the gradient of the loss function with respect to the model parameters through backpropagation, and use the optimizer to update the parameters according to the gradient to reduce the value of the loss function. Repeat the steps of forward propagation, calculating the loss, and backpropagation to update the parameters until the loss function converges or reaches the preset number of training epochs.

8. A survey and prediction system based on data processing in highway construction corresponding to the method of surveying and predicting based on data processing in highway construction according to claim 1, characterized in that, It includes a pre-construction data collection and analysis unit, a construction process monitoring and control unit, and a post-construction quality inspection and maintenance management unit. The post-construction quality inspection and maintenance management unit includes: Microseismic sensors, wireless transmission modules, and data acquisition systems. The microseismic sensors, wireless transmission modules, and data acquisition systems are used to monitor the microseismic signals of the pavement, locate the crack positions, and predict the crack development trends.

9. The survey and prediction system based on data processing in highway construction according to claim 8, characterized in that, The post-construction quality inspection and maintenance management unit includes: strain gauges, temperature sensors, humidity sensors, wireless transmission networks, and servers. The strain gauges, temperature sensors, humidity sensors, wireless transmission networks, and servers are used to collect the pavement structure performance data and provide data for performance analysis.