Asphalt performance testing method and asphalt performance testing system

Abstract

The present invention discloses an asphalt performance detection method and an asphalt performance detection system, which relate to the technical field of asphalt performance detection. The acquired pavement defect data is summarized as state monitoring data to generate an asphalt defect data set; defect points are screened out in the application area by the acquired asphalt defect data, and defect values ​​are generated from the defect state data of the defect points. If the defect value exceeds the preset defect threshold, a corresponding improvement plan is matched from the defect path of the asphalt performance, and the optimized asphalt is applied in the application area, and the asphalt defect data is re-detected and acquired. If the improvement degree of the optimized asphalt is lower than expected, the asphalt is detected after the detection interval that meets the constraint conditions, and the maintenance plan library matches the corresponding maintenance plan for the asphalt defect according to the acquired detection data. Targeted inspection and maintenance are carried out on the defect points, and when the asphalt defects need to be processed, the processing and maintenance efficiency can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of asphalt performance detection, and in particular to an asphalt performance detection method and an asphalt performance detection system. Background Art

[0002] Asphalt performance testing is a comprehensive evaluation work, which covers the comprehensive analysis of asphalt materials from basic physical properties to complex chemical properties, aiming to ensure that asphalt materials meet specific industrial and engineering standards, thereby ensuring the quality, safety and durability of road construction. This testing process not only involves the needle penetration test to evaluate the hardness of asphalt, that is, the depth of penetration by a standard needle under specified conditions, thereby reflecting its consistency and softness; it also includes the softening point test, which heats the asphalt to the temperature point where it begins to flow to determine the temperature at which the asphalt changes from a hard state to a soft state, which is crucial for evaluating the stability of asphalt at high temperatures.

[0003] Ductility testing measures asphalt's ability to resist crack formation by measuring its ability to deform when stretched. Viscosity testing measures asphalt's fluidity, or its ability to resist shear deformation, which has an important impact on asphalt's performance during paving and compaction. Asphalt performance testing also includes density testing to understand its volume mass, flash point and fire point testing to assess its safety and flammability, solubility testing to check its compatibility with other materials, and stability testing to ensure that asphalt maintains its performance over time.

[0004] In the Chinese invention patent with application publication number CN118731034A, an environmentally friendly asphalt pavement crack detection system is disclosed, including: data acquisition platform, positioning module, data processing module, detection module, carrier vehicle and auxiliary devices; the data acquisition platform is used to collect pavement image information; the positioning module is used to mark the location information of cracks in the pavement image information; the data processing module is used to identify and judge the pavement image information with cracks marked to obtain the pavement damage condition; the detection module is used to obtain the type of pavement cracks based on the pavement damage condition; the carrier vehicle and auxiliary devices are used to carry the data acquisition platform, data processing module and positioning module; the auxiliary devices include: light detector, light shield, lifting mechanism. The road base crack detection model using a multi-scale strategy extracts image features to realize crack detection of complex asphalt pavements and directly locate the crack position and type in the input image.

[0005] Combined with the above application and the contents of the prior art:

[0006] After asphalt is applied, although the chemical properties of asphalt are very stable, it is limited by the temperature and load in the application area, and the finalized asphalt is prone to defects, such as large deformation, cracks and falling off. Therefore, it is necessary to regularly inspect the state of asphalt so that it can be dealt with in time when abnormalities occur.

[0007] In the existing asphalt performance testing methods, various performance parameters of asphalt are usually tested and collected, and then the asphalt performance is judged based on the collected performance data. However, there are many application scenarios for asphalt, and there are large differences between different application scenarios. Therefore, only collecting asphalt performance data will result in the asphalt performance being qualified but still not adapting to the actual application scenarios, especially in industrial parks where the asphalt application area is an industrial park. When asphalt is used on the main transportation roads in the industrial park, the asphalt will be subjected to a large load due to the high transportation pressure, and the temperature difference between day and night in the industrial park is large. For example, when the temperature is high during the day and low at night, asphalt is more likely to have defects during application. Although the existing asphalt testing methods can obtain test data, such test data is difficult to adapt to the application scenarios of asphalt. Therefore, when the asphalt actually has defects, the obtained asphalt test data is difficult to actually improve the performance defects of the asphalt.

[0008] To this end, the present invention provides an asphalt performance detection method and an asphalt performance detection system. Summary of the invention

[0009] 1. Technical issues to be resolved

[0010] In view of the deficiencies of the prior art, the present invention provides an asphalt performance detection method and an asphalt performance detection system, which generates an asphalt defect data set by aggregating the acquired pavement defect data with the status monitoring data; the asphalt defect data acquired by the detection is used to screen out defect points in the application area, and the defect value is generated from the defect status data of the defect point. If the defect value exceeds the preset defect threshold, the defect path of the asphalt performance is matched with a corresponding improvement plan, and the optimized asphalt is applied in the application area, and the asphalt defect data is re-detected and acquired. If the improvement degree of the optimized asphalt is lower than expected, the asphalt is detected after the detection interval that meets the constraint conditions, and the maintenance plan library matches the corresponding maintenance plan for the asphalt defect based on the acquired detection data. Targeted inspection and maintenance are carried out on the defect points, and when the asphalt defects need to be processed, the processing and maintenance efficiency can be improved; thereby solving the technical problems recorded in the background technology.

[0011] (II) Technical solution

[0012] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0013] The asphalt performance detection method includes collecting application scenario data of asphalt in the application area, and if there is an abnormality in the application scenario, generating a scenario abnormal value from the abnormal state data of the abnormal scenario , if the scene is outlier If the abnormal threshold is exceeded, a collection test instruction is issued to the outside;

[0014] Collect asphalt status data in the application area. If there is any abnormality in the asphalt status in the application area, the drone will collect asphalt image data in the application area, identify and classify road surface defects, and summarize the acquired road surface defect data into status monitoring data to generate an asphalt defect data set;

[0015] The defect points are screened out in the application area using the asphalt defect data obtained by the detection, and the defect value is generated from the defect status data of the defect point. If the defect value exceeds the preset defect threshold, an alarm instruction is issued to the outside;

[0016] After matching the corresponding improvement plan from the defect path of asphalt performance, simulate and test the improvement plan and obtain test data. The performance value is generated from the simulation test data. If the performance value does not exceed the expectation, the obtained improvement plan is iteratively optimized.

[0017] Apply the optimized asphalt in the application area and retest to obtain asphalt defect data. If the improvement of the optimized asphalt is lower than expected, the asphalt is tested after a test interval that meets the constraints. Based on the obtained test data, the maintenance plan library matches the asphalt defects with corresponding maintenance plans.

[0018] Further, several detection points are selected in the application area of ​​the asphalt, and sensor groups are arranged at the detection points;

[0019] The sensor group collects application scenario data at the detection point, including day and night temperature and the corresponding temperature difference data, traffic volume data and asphalt load data, and summarizes them to generate a corresponding scene detection data set.

[0020] Furthermore, the scene detection data in the scene detection data set is used as input, and the trained abnormal scene recognition model is used to perform abnormal scene recognition, output the corresponding scene abnormality degree, obtain the scene abnormality degree of the abnormal scene and the time node when the abnormal scene is generated, and summarize to generate the corresponding scene abnormal data set.

[0021] Furthermore, the scene abnormality value is generated by the scene abnormality degree and abnormal time node of the abnormal scene. , as follows:

[0022] ;

[0023] in, is the time interval between the i-th and j-th abnormal time nodes, is the mean of the time interval, is the abnormality of the i-th scene, is the corresponding mean, , is the number of abnormal nodes, weight coefficient: , ,and .

[0024] Furthermore, after receiving the collection monitoring instruction, the sensor network in the application area collects data, including deformation data, pressure data and viscosity data, and the acquired detection data are aggregated to obtain the asphalt state monitoring data set;

[0025] Taking asphalt state monitoring data as input, the trained state detection model is used to perform abnormal state detection. If there is any abnormality in the asphalt state in the application area, an image acquisition instruction is issued to the outside.

[0026] Furthermore, each abnormal point is marked on the electronic map, and the pre-trained path planning algorithm plans the corresponding inspection path for the inspection drone. After receiving the image acquisition instruction, the drone collects road surface image data in the application area according to the inspection path and obtains the corresponding road surface image data;

[0027] Taking the collected road image data as input, the pre-trained YOLOv4 algorithm is used to identify and classify road defects, and the defect data is marked with time and location information.

[0028] Furthermore, taking the asphalt defect data as input, the trained defect evaluation model is used to evaluate the defect degree at each pavement defect location in the application area to obtain the corresponding defect score;

[0029] If the defect type at the pavement defect location is consistent with the expected defect and the defect score exceeds expectations, the corresponding location will be regarded as a defect point and an alarm command will be issued to the outside.

[0030] Furthermore, the defect points are marked on the electronic map, the detection and identification data are verified, the identification results are checked and secondary verification and annotation are performed; the defect point location, defect degree and defect generation time are summarized to generate a defect status data set; the defect value is generated from the defect status data in the defect status data set , as follows:

[0031] ;

[0032] Where: is the distance between the i-th and j-th defect point positions, is the mean distance, is a very small number greater than 0, is the threshold value of defect points, is the number of defect points; is the length of the detection cycle, is the time attenuation coefficient; is the defect degree at the defect point at time t, weight coefficient, , ,and .

[0033] Furthermore, principal component analysis was performed on asphalt application scenario data to determine the key factors affecting defect values, and asphalt defect identification was used as the target word to pre-build a knowledge graph for defect path identification;

[0034] After receiving the alarm command, the defect path of asphalt performance is output from the defect path identification knowledge graph based on the correspondence between key factors and asphalt composition data and the defect path.

[0035] Furthermore, according to the performance defect path matching improvement plan, the improvement plan is simulated and tested in various test scenarios, and the acquired test data is aggregated to generate an asphalt simulation test data set;

[0036] The simulated test data in the asphalt simulation test data set is used as input, and the trained performance evaluation model is used to evaluate the asphalt performance and generate a performance value, so as to evaluate the current asphalt performance with the obtained performance value.

[0037] Further, the optimized improvement plan is executed to optimize the asphalt components to obtain the optimized asphalt; after the performance value is reconstructed from the asphalt defect data, the performance value before and after optimization is obtained at each defect point;

[0038] Performance value at defect point Generate improvement , if the performance value is improved If the improvement level is lower than the preset threshold, several maintenance plans can be formulated in advance and a corresponding maintenance plan library can be generated.

[0039] Furthermore, feature extraction is performed on the asphalt defect data obtained by the detection to obtain the corresponding defect features. According to the correspondence between the defect features and the maintenance plan, the maintenance plan library selects the corresponding maintenance plan for each asphalt defect; the performance value at the defect point is used to determine the maintenance plan. Generate improvement The way is as follows:

[0040] ;

[0041] in, and Before and after intervention i The performance value at the defect point, and is the corresponding mean, i is the defect point number, , n is the number of defect points, For the i The median improvement value of each defect point is is the corresponding mean.

[0042] Furthermore, after a detection interval that meets the constraint conditions, detection is performed on the corresponding detection node. The method of constraining the detection interval is as follows:

[0043] ;

[0044] Where: The number of asphalt tests in the application area is: is the j-th scene outlier, is the corresponding mean, For time node and The difference between the defect values, is the expected value of the defect value difference, is the corresponding difference threshold; is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0. The weight coefficient is , ,and .

[0045] The asphalt performance detection system includes a scene abnormality analysis unit, which collects asphalt application scene data in the application area. If there is an abnormality in the application scene, the scene abnormality value is generated from the abnormal state data of the abnormal scene. , if the scene is outlier If the abnormal threshold is exceeded, a collection test instruction is issued to the outside;

[0046] The anomaly detection unit collects asphalt status data in the application area. If there is an anomaly in the asphalt status in the application area, the drone collects asphalt image data in the application area, identifies and classifies road surface defects, and summarizes the acquired road surface defect data into status monitoring data to generate an asphalt defect data set;

[0047] The defect alarm unit uses the asphalt defect data obtained by detection to screen out defect points in the application area, and generates defect values ​​based on the defect status data of the defect points. If the defect value exceeds the preset defect threshold, an alarm instruction is issued to the outside;

[0048] The test optimization unit matches the corresponding improvement plan from the defect path of asphalt performance, performs simulation test on the improvement plan and obtains test data, generates performance value from the simulation test data, and iterates and optimizes the obtained improvement plan if the performance value does not exceed the expectation;

[0049] Maintenance plan: Apply the optimized asphalt in the application area and re-test to obtain asphalt defect data. If the improvement of the optimized asphalt is lower than expected, the asphalt will be tested after a test interval that meets the constraints. Based on the obtained test data, the maintenance plan library will match the corresponding maintenance plan for the asphalt defects.

[0050] (III) Beneficial effects

[0051] The present invention provides an asphalt performance detection method and an asphalt performance detection system, which have the following beneficial effects:

[0052] 1. When there are persistent anomalies in the current application scenario, it can predict possible anomalies in asphalt. Through data recognition, it can judge whether there are anomalies in the current application scenario. When there are anomalies in the current scenario, it can be perceived in time and corresponding detection and protection measures can be taken for the asphalt in time.

[0053] 2. Generate scene abnormal value from scene abnormality degree and abnormal node of abnormal scene , a comprehensive evaluation of the abnormality of the current application scenario can be made to determine whether it will affect the performance of asphalt; the various data of asphalt are detected in real time through the deployed sensor network, and whether there is an abnormality in the current state of asphalt is judged based on the detection data, and whether there are abnormal defects and damage in the asphalt in the current application scenario is verified.

[0054] 3. The road surface defects can be identified through image analysis and recognition, and the asphalt at the abnormal point can be verified to see if there is any abnormality. Compared with single detection and recognition, the detection reliability is higher. The corresponding defect points can be screened out in the application area, and the defect points can be inspected and maintained in a targeted manner. When the asphalt defects need to be handled, the handling and maintenance efficiency can be improved.

[0055] 4. Through dual defect detection and identification, the defect value is used to comprehensively evaluate the utilization status of asphalt in the defect area, completing the comprehensive detection process of asphalt in the application area, and timely adjustments can be made when the performance of asphalt is poor; the defect path of asphalt is output by the knowledge graph, and the asphalt performance can be improved based on the defect path and the detection of asphalt.

[0056] 5. Carry out simulation testing and iterative optimization on the improvement plan, simulate and test and predict the improved asphalt performance and defects, and obtain the optimized improvement plan based on actual operation and adjustment. Compared with all manual adjustments and inspections, the improvement efficiency is higher when adjusting and testing the asphalt components.

[0057] 6. Apply the optimized asphalt in the current application scenario and retest it. Based on the performance value and improvement obtained, judge the degree of improvement in asphalt performance. If the performance value and improvement do not exceed expectations, the road surface in the application area can be maintained.

[0058] 7. Carry out targeted maintenance on asphalt regularly to improve the efficiency and targeting of maintenance; extend the service life of asphalt through targeted maintenance; constrain the maintenance intervals to make the maintenance frequency adapt to the current application scenarios, so that the application and maintenance of asphalt are suitable. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a schematic diagram of the process of the asphalt performance detection method of the present invention;

[0060] Figure 2 It is a schematic diagram of the structure of the asphalt performance detection system of the present invention. DETAILED DESCRIPTION

[0061] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] See also Figure 1 The present invention provides a method for detecting asphalt performance, comprising:

[0063] Step 1: Collect asphalt application scenario data in the application area. If there is an abnormality in the application scenario, generate a scenario abnormal value based on the abnormal state data of the abnormal scenario. , if the scene is outlier If the abnormal threshold is exceeded, a collection test instruction is issued to the outside;

[0064] The step 1 includes the following contents:

[0065] Step 101: When asphalt is used on the road surface of the main transportation road in the industrial area, determine the corresponding application area, evenly select a number of detection points in the application area of ​​the asphalt, and arrange sensor groups at the detection points, including temperature and humidity sensors, load sensors, and quantity sensors;

[0066] The sensor group collects application scenario data at the detection point, including day and night temperature and the corresponding temperature difference data, traffic volume data and asphalt load data, etc., and summarizes them to generate the corresponding scenario detection data set;

[0067] During use, considering that the performance and use status of asphalt are greatly affected by external environmental conditions, when asphalt enters the application state, by testing its application scenario, when there are persistent abnormalities in the current application scenario, it can predict possible abnormalities of asphalt;

[0068] Step 102: training a machine learning algorithm with the labeled sample data to obtain a trained abnormal scene recognition model;

[0069] Taking the scene detection data in the scene detection data set as input, the trained abnormal scene recognition model is used to recognize abnormal scenes, and the recognized abnormal scenes are compared with the reference scenes to output the corresponding scene abnormality degree;

[0070] Describe the abnormality of the current scene with the scene abnormality degree, obtain the scene abnormality degree of the abnormal scene and the time node when the abnormal scene occurs, and summarize and generate the corresponding scene abnormal data set;

[0071] As a further content, based on the collection of abnormal scene data, through data recognition, it is possible to judge whether there is an abnormality in the current application scene, and when there is an abnormality in the current scene, it can be perceived in time, and when there is an abnormality in the current application scene, corresponding detection and protection measures can be taken for the asphalt in time;

[0072] Step 103: Under dimensionless conditions, a scene abnormality value is generated from the scene abnormality degree and abnormal time node of the abnormal scene. , using abnormal values ​​to comprehensively evaluate the current usage scenarios and usage conditions of asphalt, as follows:

[0073] ;

[0074] in, is the time interval between the i-th and j-th abnormal time nodes, is the mean of the time interval, is the abnormality of the i-th scene, is the corresponding mean, , is the number of abnormal nodes, weight coefficient: , ,and ;

[0075] Pre-set abnormal thresholds based on historical data and management expectations for asphalt usage scenarios and usage environments;

[0076] If the scene is outlier If the abnormal threshold is exceeded, it means that the performance index of the applied asphalt may not meet the expected use effect in the current usage scenario, and it needs to be optimized or maintained on the basis of detection and testing. At this time, a collection test instruction is issued to the outside;

[0077] When using, combine the contents in steps 101 to 103:

[0078] As a further content, the scene abnormality value is generated by the scene abnormality degree and abnormal node of the abnormal scene. ,When the current asphalt utilization scenario has been continuously abnormal, a comprehensive evaluation of the abnormality level of the current application scenario is conducted to determine whether the current continuous abnormality will affect the asphalt performance.

[0079] In the existing asphalt performance testing methods, various performance parameters of asphalt are usually tested and collected, and then the asphalt performance is judged based on the collected performance data. However, there are many application scenarios for asphalt, and there are large differences between different application scenarios. Therefore, only collecting asphalt performance data will result in the asphalt performance being qualified but still not adapting to the actual application scenarios, especially in industrial parks where the asphalt application area is an industrial park. When asphalt is used on the main transportation roads in the industrial park, the asphalt will be subjected to a large load due to the high transportation pressure, and the temperature difference between day and night in the industrial park is large. For example, when the temperature is high during the day and low at night, asphalt is more likely to have defects during application. Although the existing asphalt testing methods can obtain test data, such test data is difficult to adapt to the application scenarios of asphalt. Therefore, when the asphalt actually has defects, the obtained asphalt test data is difficult to actually improve the performance defects of the asphalt.

[0080] Step 2: Collect asphalt status data in the application area. If the asphalt status in the application area is abnormal, the drone will collect asphalt image data in the application area, identify and classify road surface defects, and summarize the acquired road surface defect data into status monitoring data to generate an asphalt defect data set;

[0081] The step 2 includes the following contents:

[0082] Step 201: after receiving the collection and monitoring instruction, a sensor network is arranged in the application area, including deformation sensors, pressure sensors, viscosity sensors and displacement sensors, etc.; the sensor network collects data, including deformation data, pressure data and viscosity data, etc., and the acquired detection data is aggregated to obtain a set of asphalt state monitoring data;

[0083] Train the machine learning algorithm with the labeled sample data to obtain the trained state detection model;

[0084] Taking asphalt state monitoring data as input, the trained state detection model is used to detect abnormal states. If the asphalt state in the application area is abnormal, the corresponding abnormal points are marked and image acquisition instructions are issued to the outside.

[0085] When in use, if there are persistent anomalies in the application scenario within the application area, the sensor network deployed will be used as the corresponding feedback content to detect various data of the asphalt in real time. Based on the detection data, it can be judged whether there are any abnormalities in the current state of the asphalt, and whether there are any abnormal defects and damage to the asphalt in the current application scenario;

[0086] Step 202: construct an electronic map covering the application area, mark each abnormal point on the electronic map, use the pre-trained path planning algorithm to plan a corresponding inspection path for the inspection drone, and mark the inspection path on the electronic map;

[0087] After receiving the image acquisition instruction, the UAV uses high-resolution camera equipment to collect road image data in the application area according to the inspection path and obtains the corresponding road image data;

[0088] Step 203: using the collected road surface image data as input, using the pre-trained YOLOv4 algorithm to identify and classify road surface defects, and marking the defect data with time and location information to obtain corresponding road surface defect data; after aggregating the road surface defect data with the state monitoring data, an asphalt defect data set is generated;

[0089] When using, combine the contents in steps 201 to 203:

[0090] As a further content, after the sensor network identifies the corresponding abnormal point in the application area, image data is collected for the abnormal point and its adjacent area, and the road surface defects are determined by analyzing and identifying the image. It can be verified whether the asphalt at the abnormal point has abnormalities. Compared with single detection and identification, the detection reliability is higher;

[0091] There are many reasons for asphalt defects. The following is a detailed analysis of these reasons:

[0092] Asphalt material: The wax content and aging properties of asphalt directly affect its quality. For example, asphalt with a high wax content is prone to aging, thus affecting the durability of the asphalt pavement.

[0093] Base material: The shrinkage, strength and other characteristics of the base material have a significant impact on the crack rate of the asphalt surface layer. The smaller the shrinkage of the base material, the fewer cracks in the surface layer; otherwise, there will be more cracks.

[0094] Aggregate: The properties of aggregates, such as acidity, alkalinity, shape, size, etc., also affect the adhesion and stability of the asphalt mixture. For example, acidic aggregates have a low adhesion to asphalt, which may lead to pavement spalling.

[0095] Insufficient compaction: Low compaction temperature, mixture gradation that does not meet requirements, and few compaction passes may all lead to a lack of density in asphalt concrete, thus affecting the strength and durability of the road surface.

[0096] Irregular construction: such as inaccurate asphalt-to-stone ratio, improper use of penetration oil or tack oil, etc., may cause problems such as oil spills, oil pockets, and rutting on the road surface.

[0097] Paving and rolling: Outdated paving equipment, too low paving temperature, untimely rolling, etc. may all lead to poor road surface flatness or even cracks.

[0098] Unreasonable structural design: The original structural design did not fully take into account various unfavorable factors, such as improper foundation treatment, uneven roadbed compaction, etc., which may lead to uneven roadbed settlement and cracks.

[0099] Improper material selection: Failure to select appropriate asphalt and aggregate based on factors such as local climatic conditions and traffic volume may lead to premature fatigue failure of the pavement.

[0100] Climate conditions: Climate factors such as temperature, humidity, and rainfall can affect the performance of asphalt pavement. For example, low temperatures increase the shrinkage of asphalt, which may cause pavement cracking; rainfall may accelerate water damage to the pavement.

[0101] Traffic volume: With the rapid development of transportation, the existing road surface strength may become increasingly insufficient and unable to meet the growing traffic demand.

[0102] Step 3: Use the asphalt defect data obtained through the detection to screen out defect points of asphalt application in the application area, generate defect values ​​based on the defect status data of the defect points, and send an alarm command to the outside if the defect value exceeds a preset defect threshold;

[0103] The step three includes the following contents:

[0104] Step 301: train a machine learning algorithm with the labeled sample data to obtain a trained defect evaluation model; use the asphalt defect data as input, use the trained defect evaluation model to evaluate the defect degree at each pavement defect position in the application area, and obtain a corresponding defect score;

[0105] If the defect type at the pavement defect location is consistent with the expected defect, and the defect score exceeds expectations, it means that the current pavement defect is of high severity and needs attention and treatment. At this time, the corresponding location is regarded as a defect point and a first-level alarm instruction is issued to the outside;

[0106] When in use, as a further content, after completing the asphalt detection in the application area and obtaining the corresponding defect detection data, the corresponding defect points in the application area can be screened out, and the defect points can be repaired and maintained in a targeted manner. When the asphalt defects need to be processed, the processing and maintenance efficiency can be improved;

[0107] Step 302: After receiving the first-level alarm instruction, the defect point is marked on the electronic map, and the inspection personnel manually verify the inspection identification data, check the identification results and perform secondary verification and annotation; summarize the defect point location, defect degree and defect generation time to generate a defect status data set;

[0108] Step 303: Under dimensionless conditions, generate defect values ​​from defect state data in the defect state data set , as follows:

[0109] ;

[0110] Where: is the distance between the i-th and j-th defect point positions, is the mean distance, is a very small number greater than 0, used to avoid the denominator being 0. is the threshold value of defect points, is the number of defect points; is the length of the detection cycle, is the time attenuation coefficient, the value is between 0 and 1, and the value of e is 2.718; is the defect degree at the defect point at time t, weight coefficient, , ,and ;

[0111] Defect thresholds are set in advance based on historical data and management expectations for asphalt pavement defects; if the defect value obtained is If the preset defect threshold is exceeded, it means that the current asphalt performance cannot achieve the expected performance and use effect. At this time, a secondary alarm command is issued to the outside;

[0112] When using, combine the contents in steps 301 to 303:

[0113] As a further content, defect values ​​are constructed based on the defect-related data of the defect points. When there are many defect points in the application area, in the current application scenario, dual defect detection and identification are carried out, and the utilization status of the asphalt in the defect area is comprehensively evaluated by the defect value. The comprehensive detection process of asphalt in the application area is completed, and the comprehensive detection and testing of asphalt performance is realized. When the performance of the asphalt is not good, timely adjustments can be made.

[0114] Step 4: After matching the corresponding improvement plan with the defect path of asphalt performance, simulate and test the improvement plan and obtain test data, generate performance values ​​from the simulation test data, and iterate and optimize the obtained improvement plan if the performance value does not exceed expectations;

[0115] The step 4 includes the following contents:

[0116] Step 401: Perform principal component analysis on asphalt application scenario data to determine key factors that affect defect values, use asphalt defect recognition as the target word, and pre-build a defect path recognition knowledge graph after deep retrieval and entity relationship building; after receiving the secondary alarm instruction, output the defect path of asphalt performance from the defect path recognition knowledge graph based on the correspondence between the key factors and the component data of asphalt and the defect path;

[0117] When in use, after obtaining the key factors that affect the performance of asphalt, the knowledge graph outputs the defect path of asphalt. Through analysis and comparison, the causes of defects in asphalt are determined, and the defect path that causes defects in asphalt performance is determined. Based on the defect path, the asphalt performance can be improved on the basis of asphalt testing;

[0118] Step 402: According to the performance defect path matching improvement plan, pre-set test scenarios, including combinations of different seasons, different time periods, and different traffic flows, simulate and test the improvement plan in each test scenario, and obtain corresponding test data, including deformation data, pressure data, viscosity data, and defect data generated in different scenarios, and aggregate the obtained test data to generate an asphalt simulation test data set;

[0119] Step 403: train a machine learning algorithm with the labeled sample data to obtain a trained performance evaluation model; use the simulation test data in the asphalt simulation test data set as input, use the trained performance evaluation model to evaluate the asphalt performance and generate a performance value, and use the obtained performance value to evaluate the current asphalt performance;

[0120] If the obtained performance value does not exceed the expectation, it means that the current improvement plan is not reliable enough and the improvement degree of asphalt performance after execution does not exceed the expectation. At this time, the improvement plan is iteratively optimized and tested using the pre-trained genetic algorithm to obtain the optimized improvement plan.

[0121] When used, combine the contents in steps 401 to 403:

[0122] As a further content, after matching the corresponding improvement plan according to the defect path, in order to improve the subsequent improvement efficiency, the improvement plan is simulated and tested and iteratively optimized, and the improved asphalt performance and defects are simulated and tested and predicted, so as to obtain the optimized improvement plan based on actual operation and adjustment. Compared with all manual adjustments and inspections, the improvement efficiency is higher when adjusting and testing the asphalt components.

[0123] Step 5: Apply the optimized asphalt in the application area and retest to obtain asphalt defect data. If the improvement of the optimized asphalt is lower than expected, the asphalt is tested after a test interval that meets the constraint conditions, and the maintenance program library matches the corresponding maintenance program for the asphalt defect based on the obtained test data;

[0124] The step five includes the following contents:

[0125] Step 501: To improve the crack resistance, wear resistance and temperature change resistance of asphalt, the optimized improvement plan is executed to optimize the asphalt components to obtain the optimized asphalt; the optimized asphalt is applied in the application area and then retested to obtain asphalt defect data;

[0126] After reconstructing the performance value from the asphalt defect data, the performance values ​​before and after optimization are obtained at each defect point. Generate improvement , as follows:

[0127] ;

[0128] in, and Before and after intervention i The performance value at the defect point, and is the corresponding mean, i is the defect point number, , n is the number of defect points, For the i The median improvement value of each defect point is is the corresponding mean;

[0129] If the performance value is improved If the improvement is lower than the preset improvement threshold, it means that the adjustment and optimization of the asphalt components may not have the expected effect. As a corresponding improvement measure, in order to extend the application time of asphalt, several maintenance plans can be formulated in advance and the corresponding maintenance plan library can be generated;

[0130] When in use, based on the testing and optimization of the asphalt, the optimized asphalt is obtained, and the optimized asphalt is applied and re-tested in the current application scenario. The performance values ​​and improvements obtained on this basis are used to judge the degree of improvement in the asphalt performance. If the performance values ​​and improvements do not exceed expectations, the road surface in the application area needs to be maintained to improve the application status of the asphalt.

[0131] Step 502: constrain the detection interval. After the detection interval that meets the constraint conditions, perform detection on the corresponding detection node. The constraint method is as follows:

[0132] ;

[0133] Where: The number of asphalt tests in the application area is: is the j-th scene outlier, is the corresponding mean, For time node and The difference between the defect values, is the expected value of the defect value difference, is the corresponding difference threshold; is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0. The weight coefficient is , ,and , the value is consistent with the previous value;

[0134] Extract features from the asphalt defect data obtained through detection to obtain corresponding defect features. Based on the correspondence between the defect features and the maintenance plans, the maintenance plan library selects the corresponding maintenance plan for each asphalt defect.

[0135] When using, combine the contents in steps 501 and 502:

[0136] After testing and optimizing the asphalt, if the optimized asphalt still cannot adapt to the current application scenario, targeted maintenance can be carried out on the asphalt regularly to improve the efficiency and targeting of maintenance. When there are deficiencies in the asphalt performance, the service life of the asphalt can be extended through targeted maintenance. As a further content, by constraining the maintenance interval, the maintenance frequency is adapted to the current application scenario, making the application and maintenance of the asphalt appropriate.

[0137] The following is a detailed method for constructing a knowledge graph for identifying asphalt performance defect paths:

[0138] The knowledge graph for asphalt performance defect path identification aims to integrate knowledge on the identification, causes, detection methods and repair measures of asphalt pavement defects to provide intelligent support for road maintenance and management.

[0139] Determine the goal and scope of the knowledge graph - Goal: Build a knowledge graph that can identify asphalt performance defects, analyze causes, recommend detection methods, and provide repair measures. Scope: Cover common defect types, causes, detection methods, and repair measures of asphalt pavement.

[0140] Data collection and preprocessing - Defect types: Collect common defect types of asphalt pavement, such as cracks (longitudinal cracks, transverse cracks, cracks, etc.), potholes, looseness, oil spills, segregation, etc. Cause analysis: For each defect type, collect and analyze its possible causes, such as improper construction, poor material quality, environmental factors (such as temperature changes, water infiltration), etc. Detection methods: Collect and organize various methods for detecting asphalt performance defects, such as observation method, coring method, nuclear density meter method, infrared photography method, etc. Repair measures: For each defect type, collect and organize its corresponding repair measures. Data preprocessing: Clean and organize the collected data to ensure the accuracy and completeness of the data.

[0141] Defining entities and relationships - Entity definition: Defect types: such as cracks, potholes, etc. Causes: such as improper construction, poor material quality, etc. Detection methods: such as observation method, nuclear density meter method, etc. Repair measures: such as grouting, re-laying, etc. Relationship definition: Defects and causes: Each defect type may be caused by multiple causes. Defects and detection methods: Each defect type can be detected by a specific detection method. Defects and repair measures: Each defect type requires specific repair measures to be repaired.

[0142] Build a knowledge graph model - Select a knowledge graph tool: Select a suitable knowledge graph building tool according to actual needs, such as Neo4j, Dgraph, etc. Create entities and relationships: Create entities and relationships in the knowledge graph tool, connect defect types, causes, detection methods, and repair measures, etc. to form a complete knowledge graph model. Add attributes and descriptions: Add attributes to entities, such as the shape, size, and location of defect types; add descriptions to relationships, such as a detailed description of the cause, the operating steps of the detection method, etc.

[0143] Verification and optimization of knowledge graphs - Data verification: Verify the imported data to ensure the accuracy and completeness of the data. Relationship verification: Verify the created relationships to ensure the correctness and rationality of the relationships. Graph optimization: Optimize the knowledge graph based on the verification results, such as adjusting the attributes of entities and relationships, adding new entities and relationships, etc.

[0144] Application and maintenance of knowledge graphs—Application: Apply the constructed knowledge graphs to the identification, detection, and repair of asphalt performance defects to improve work efficiency and accuracy. Maintenance: Regularly update and maintain the knowledge graphs, such as adding new defect types, causes, detection methods, and repair measures, to maintain the timeliness and accuracy of the graphs.

[0145] Case study of knowledge graph construction - Taking cracks as an example, the fragments of constructing the knowledge graph are as follows: entity: crack, attributes: shape (linear, mesh), size (length, width), position (center of the road surface, edge), etc.

[0146] Relationship: Causes: aging of the road surface, heavy vehicle loads, water intrusion, etc. Detection methods: observation, infrared photography, radar detection, etc. Repair measures: grouting, surface sealing, etc.

[0147] Building a knowledge graph for identifying asphalt performance defect paths requires steps such as clarifying goals and scope, collecting and preprocessing data, defining entities and relationships, building a knowledge graph model, validating and optimizing the graph, and applying and maintaining the graph. By building a knowledge graph, we can integrate knowledge on the identification, causes, detection methods, and repair measures of asphalt pavement defects, providing intelligent support for road maintenance and management.

[0148] See also Figure 2 The present invention provides an asphalt performance detection system, comprising:

[0149] The scene abnormality analysis unit collects the application scene data of asphalt in the application area. If there is an abnormality in the application scene, the scene abnormality value is generated from the abnormal state data of the abnormal scene. , if the scene is outlier If the abnormal threshold is exceeded, a collection test instruction is issued to the outside;

[0150] The anomaly detection unit collects asphalt status data in the application area. If there is an anomaly in the asphalt status in the application area, the drone collects asphalt image data in the application area, identifies and classifies road surface defects, and summarizes the acquired road surface defect data into status monitoring data to generate an asphalt defect data set;

[0151] The defect alarm unit uses the asphalt defect data obtained by detection to screen out defect points in the application area, and generates defect values ​​based on the defect status data of the defect points. If the defect value exceeds the preset defect threshold, an alarm instruction is issued to the outside;

[0152] The test optimization unit matches the corresponding improvement plan from the defect path of asphalt performance, performs simulation test on the improvement plan and obtains test data, generates performance value from the simulation test data, and iterates and optimizes the obtained improvement plan if the performance value does not exceed the expectation;

[0153] Maintenance plan: Apply the optimized asphalt in the application area and re-test to obtain asphalt defect data. If the improvement of the optimized asphalt is lower than expected, the asphalt will be tested after a test interval that meets the constraints. Based on the obtained test data, the maintenance plan library will match the corresponding maintenance plan for the asphalt defects.

[0154] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0155] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0156] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0157] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0158] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. Asphalt performance testing method, characterized in that: include, Collect asphalt application scenario data in the application area. If there is an abnormality in the application scenario, the abnormal state data of the abnormal scenario will generate a scenario abnormal value. , if the scene is outlier If the abnormal threshold is exceeded, a collection test instruction is issued to the outside; Collect asphalt status data in the application area. If there is any abnormality in the asphalt status in the application area, the drone will collect asphalt image data in the application area, identify and classify road surface defects, and summarize the acquired road surface defect data into status monitoring data to generate an asphalt defect data set; The defect points are screened out in the application area using the asphalt defect data obtained by the detection, and the defect value is generated from the defect status data of the defect point. If the defect value exceeds the preset defect threshold, an alarm instruction is issued to the outside; After matching the corresponding improvement plan from the defect path of asphalt performance, simulate and test the improvement plan and obtain test data. The performance value is generated from the simulation test data. If the performance value does not exceed the expectation, the obtained improvement plan is iteratively optimized. Apply the optimized asphalt in the application area and retest to obtain asphalt defect data. If the improvement of the optimized asphalt is lower than expected, test the asphalt after a test interval that meets the constraint conditions. Based on the test data obtained, the maintenance program library matches the corresponding maintenance program for the asphalt defect. Select several detection points in the application area of ​​asphalt and arrange sensor groups at the detection points; The sensor group collects application scenario data at the detection point, including day and night temperature and the corresponding temperature difference data, traffic volume data and asphalt load data, and summarizes them to generate a corresponding scenario detection data set; Taking the scene detection data in the scene detection data set as input, use the trained abnormal scene recognition model to perform abnormal scene recognition, output the corresponding scene abnormality degree, obtain the scene abnormality degree of the abnormal scene and the time node when the abnormal scene was generated, and summarize and generate the corresponding scene abnormal data set; Generate scene abnormal value based on scene abnormality degree and abnormal time node of abnormal scene , as follows: ； in, is the time interval between the i-th and j-th abnormal time nodes, is the mean of the time interval, is the abnormality of the i-th scene, is the corresponding mean, , is the number of abnormal nodes; weight coefficient: , ,and .

2. The asphalt performance detection method according to claim 1, characterized in that: After receiving the collection and monitoring instruction, the sensor network in the application area collects data, including deformation data, pressure data and viscosity data, and aggregates the acquired detection data to generate an asphalt state monitoring data set; Taking asphalt state monitoring data as input, the trained state detection model is used to perform abnormal state detection. If there is any abnormality in the asphalt state in the application area, an image acquisition instruction is issued to the outside.

3. The asphalt performance detection method according to claim 2, characterized in that: Each abnormal point is marked on the electronic map, and the pre-trained path planning algorithm plans the corresponding inspection path for the inspection drone. After receiving the image acquisition instruction, the drone collects road image data in the application area according to the inspection path and obtains the corresponding road image data; Taking the collected road image data as input, the pre-trained YOLOv4 algorithm is used to identify and classify road defects, and the defect data is marked with time and location information.

4. The asphalt performance detection method according to claim 3, characterized in that: Taking asphalt defect data as input, the trained defect evaluation model is used to evaluate the defect degree at each pavement defect location in the application area to obtain the corresponding defect score; If the defect type at the pavement defect location is consistent with the expected defect and the defect score exceeds expectations, the corresponding location will be regarded as a defect point and an alarm command will be issued to the outside.

5. The asphalt performance detection method according to claim 4, characterized in that: Mark the defect points on the electronic map, verify the detection and identification data, check the identification results and perform secondary verification and annotation; Summarize the defect point location, defect degree and defect generation time to generate a defect status data set; Generate defect value from defect status data in defect status data set , as follows: ； Where: is the distance between the i-th and j-th defect point positions, is the mean distance, is a very small number greater than 0, is the threshold value of defect points, is the number of defect points; is the length of the detection cycle, is the degree of defect at the defect point at time t; weight coefficient: , ,and .

6. The asphalt performance detection method according to claim 5, characterized in that: Perform principal component analysis on asphalt application scenario data to determine the key factors that affect defect values, use asphalt defect identification as the target word, and pre-build a knowledge graph for defect path identification; After receiving the alarm command, the defect path of asphalt performance is output from the defect path identification knowledge graph based on the correspondence between key factors and asphalt composition data and the defect path.

7. The asphalt performance detection method according to claim 6, characterized in that: According to the performance defect path matching improvement plan, the improvement plan is simulated and tested in various test scenarios, and the acquired test data is aggregated to generate an asphalt simulation test data set; The simulated test data in the asphalt simulation test data set is used as input, and the trained performance evaluation model is used to evaluate the asphalt performance and generate a performance value, so as to evaluate the current asphalt performance with the obtained performance value.

8. The asphalt performance detection method according to claim 7, characterized in that: Execute the optimized improvement plan to optimize the asphalt components, obtain the optimized asphalt, reconstruct the performance value based on the asphalt defect data, and obtain the performance value before and after optimization at each defect point; Performance value at defect point Generate improvement , if the performance value is improved If the improvement level is lower than the preset threshold, several maintenance plans can be formulated in advance and a corresponding maintenance plan library can be generated.

9. The asphalt performance detection method according to claim 8, characterized in that: Feature extraction is performed on the asphalt defect data obtained by detection to obtain the corresponding defect features. According to the correspondence between the defect features and the maintenance plan, the maintenance plan library selects the corresponding maintenance plan for each asphalt defect; the performance value at the defect point is used to calculate the maintenance plan. Generate improvement The way is as follows: ； in, and Before and after intervention The performance value at the defect point, and is the corresponding mean, is the defect point number, , is the number of defect points, is the median improvement value of the i-th defect point, is the corresponding mean.

10. The asphalt performance detection method according to claim 9, characterized in that: After the detection interval meets the constraint conditions, the detection is performed on the corresponding detection node. The constraints are as follows: ； Where: The number of asphalt tests in the application area is: is the j-th scene outlier, is the corresponding mean, For time node and The difference between the defect values, is the expected value of the defect value difference, is the corresponding difference threshold; is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0; weight coefficient: , ,and .

11. Asphalt performance testing system, using the testing method according to any one of claims 1 to 10, characterized in that: include, The scene abnormality analysis unit collects the application scene data of asphalt in the application area. If there is an abnormality in the application scene, the scene abnormality value is generated from the abnormal state data of the abnormal scene. , if the scene is outlier If the abnormal threshold is exceeded, a collection test instruction is issued to the outside; The anomaly detection unit collects asphalt status data in the application area. If there is an anomaly in the asphalt status in the application area, the drone collects asphalt image data in the application area, identifies and classifies road surface defects, and summarizes the acquired road surface defect data into status monitoring data to generate an asphalt defect data set; The defect alarm unit uses the asphalt defect data obtained by detection to screen out defect points in the application area, and generates defect values ​​based on the defect status data of the defect points. If the defect value exceeds the preset defect threshold, an alarm instruction is issued to the outside; The test optimization unit matches the corresponding improvement plan from the defect path of asphalt performance, performs simulation test on the improvement plan and obtains test data, generates performance value from the simulation test data, and iterates and optimizes the obtained improvement plan if the performance value does not exceed the expectation; Maintenance plan: Apply the optimized asphalt in the application area and re-test to obtain asphalt defect data. If the improvement of the optimized asphalt is lower than expected, the asphalt will be tested after a test interval that meets the constraints. Based on the obtained test data, the maintenance plan library will match the corresponding maintenance plan for the asphalt defects.

Images