Roadbed compaction quality monitoring, management and maintenance method and system based on virtual and real index collaboration

Abstract

The invention relates to a roadbed compaction quality monitoring, management and maintenance method and system based on virtual and real index collaboration. The method comprises the following steps: step 1, establishing an indoor compaction test for reproducing on-site roadbed compaction conditions, and establishing a data set of the indoor test; 2, implementing a field test on roadbed pressure, and establishing a data set of the field test; 3, building a roadbed compaction quality prediction model based on virtual and real data collaborative training; and 4, carrying out construction management and control application on the roadbed compaction quality prediction model based on virtual and real index collaborative prediction. According to the method, the influence of roadbed filler parameters, road roller working parameters and external environment parameters of roadbed compaction on the compaction quality is fully considered, multiple virtual-real indexes cooperate with one another to establish an artificial neural network compaction quality prediction model of the roadbed compaction quality, and the model is optimized and improved while being migrated to engineering practical application. And real-time accurate monitoring, management and maintenance of the compaction quality are realized in the on-site roadbed compaction process.

Description

Technical Field

[0001] The present invention relates to the technical field of road construction, and in particular to a roadbed compaction quality monitoring and maintenance method and system based on the coordination of virtual and real indicators. Background Art

[0002] The detection of roadbed compaction quality is a key link in the road construction process. Traditional detection methods such as sand filling method and ring knife method use point sampling combined with a large number of indoor tests. Manual operation is prone to errors, and the detection process is labor-intensive and time-consuming. In addition, traditional compaction quality detection methods are mostly post-tests, which are lagging and detect points instead of surfaces, which is prone to insufficient or overcompacted points and rolling sections.

[0003] There are many factors and feedback indicators that affect the quality of roadbed compaction. If the influencing factors are not considered comprehensively, or if the roadbed compaction quality is judged based on a single characterization phenomenon, it is easy to produce errors or make wrong judgments. Common intelligent compaction monitoring indicators such as harmonic ratio CMV, vibration steel wheel amplitude, spectrum, etc. are all feedback indicators that cannot directly reflect the compaction quality of roadbed fillers and can be defined as "virtual" indicators. During the vibration compaction of roadbed fillers, under the action of the static pressure and exciting force of the vibrating steel wheel, the pores in the filler are filled and compacted, and the volume continues to decrease until it is dense, resulting in settlement. The settlement amount is an intuitive physical and mechanical indicator in the process of vibration compaction of the filler. The ratio of the filler rolling settlement amount to the filler thickness before rolling is defined as the settlement rate. The settlement rate can reflect the compaction of the roadbed filler at the macro level and is defined as a "real" indicator. However, a set of intelligent monitoring technology methods for roadbed compaction quality based on multiple macro- and micro-level virtual and real indicators has not yet been formed. The difficulty lies in the fact that there are many influencing factors between the compaction quality of roadbed fillers and the monitoring indicators, and there is a lack of theoretical prediction models for roadbed compaction quality that fully consider various factors. Simple physical models or empirical formulas do not consider sufficient and specific factors, and the accuracy of the test results cannot meet actual needs.

[0004] Therefore, it is necessary to propose a real-time monitoring and maintenance method for roadbed compaction quality that coordinates multiple parameters and cooperates with multiple virtual and real indicators, so as to realize accurate real-time monitoring and maintenance of the entire process of roadbed compaction quality in engineering practice. Summary of the invention

[0005] The problem that the present invention is intended to solve is that the traditional roadbed compaction quality detection method has a large workload, a long time consumption, and a lag in post-detection. The existing continuous compaction intelligent monitoring system and method do not take into account comprehensive factors, and the monitoring indicators are single, and most of them are "virtual" indicators for feedback. In view of the shortcomings of the prior art, the present invention proposes a roadbed compaction quality monitoring management and maintenance method and system based on the coordination of virtual and real indicators. The method fully considers the influence of roadbed filling parameters, roller working parameters and external environmental parameters on the compaction quality of roadbed compaction, and establishes a data set based on a large number of laboratory tests and field tests that reproduce on-site roadbed compaction conditions. The multivariate "virtual and real" indicators work together to establish an artificial neural network compaction quality prediction model for roadbed compaction quality. The model is optimized and improved while being migrated to engineering practice applications, so as to achieve real-time and accurate monitoring and management and maintenance of compaction quality during on-site roadbed compaction.

[0006] The technical solution adopted by the present invention to solve the technical problem is:

[0007] In a first aspect, the present invention provides a roadbed compaction quality monitoring and maintenance method based on virtual and real indicator collaboration, the method comprising the following steps:

[0008] Step 1: Establish an indoor compaction test that reproduces the on-site roadbed compaction conditions and establish a data set for the indoor test;

[0009] Step 2: Conduct field test on roadbed compaction and establish a data set for the field test;

[0010] Step 3: Establish a roadbed compaction quality prediction model based on virtual and real data collaborative training:

[0011] A three-layer artificial neural network compaction quality prediction model including input layer, hidden layer and output layer is constructed; the input layer includes 12 characteristic values ​​including filler soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, water content, exciting force, exciting frequency, amplitude, harmonic ratio CMV, soil pressure at different positions and settlement rate; the output layer is compaction degree;

[0012] The data set of indoor tests is used as samples to train the compaction quality prediction model of artificial neural network;

[0013] The data set of the field test is input into the trained artificial neural network compaction quality prediction model, and the trained artificial neural network compaction quality prediction model is optimized to obtain a roadbed compaction quality prediction model based on the collaborative training of virtual and real data;

[0014] Step 4: Construction control application of roadbed compaction quality prediction model based on collaborative prediction of virtual and real indicators:

[0015] Before the on-site roadbed compaction application, set up 1 to 2 test strips for on-site roadbed compaction testing, and use the real-time data of the test strips to calibrate the roadbed compaction quality prediction model based on the collaborative training of virtual and real data to obtain the optimized and calibrated roadbed compaction quality prediction model based on the collaborative training of virtual and real data. The optimized and calibrated roadbed compaction quality prediction model based on the collaborative training of virtual and real data can be used for on-site roadbed compaction implementation.

[0016] The specific implementation process of step one is: the indoor test uses a surface vibration compactor to complete the laboratory surface vibration compaction test, installs two laser displacement sensors above the tamping plate of the surface vibration compactor, installs two vertical acceleration sensors at the center and edge of the tamping plate of the surface vibration compactor, installs a harmonic ratio CMV sensor on the tamping plate of the surface vibration compactor, and the temperature sensor and humidity sensor are used to monitor external environmental parameters; set the paving thickness, install six soil pressure sensors at the top, middle and bottom positions in the paving area of ​​the test tube of the surface vibration compactor, and the six soil pressure sensors are arranged in three layers, two on each layer, one soil pressure sensor is located in the center of the same layer, and the other soil pressure sensor is located at the edge, and all soil pressure sensors located at the edge are at the same horizontal position;

[0017] The laboratory surface vibration compaction test conditions are outputted by using the indoor test condition adjustment subroutine, including roadbed filling parameters, roller working parameters and external environmental parameters, wherein the roadbed filling parameters include soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio and water content; the roller working parameters include exciting force and exciting frequency; and the external environmental parameters include temperature and humidity;

[0018] According to the energy conversion relationship between the roller and the surface vibration compactor, the working parameters of the surface vibration compactor are adjusted. The tamping plate of the surface vibration compactor is adjusted by adding or subtracting the counterweight to adjust the frequency converter of the exciting motor to change the exciting force and exciting frequency of the surface vibration compactor. Indoor tests are carried out according to the laboratory surface vibration compaction test conditions, and the exciting force L of the surface vibration compactor is set. s ;

[0019] The acceleration signal during loading is calculated based on the acceleration of the two vertical acceleration sensors, and the acceleration signal during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f S , the acceleration signal is integrated in the time domain during the loading process to obtain the first tamping plate amplitude i S1 ;

[0020] Calculate the real-time sedimentation rate s of the filler based on the real-time values ​​of the laser displacement sensors on both sides and the filling height of the filler S, where the filling height of the filler is a set known value; the electrical signal collected by the laser displacement sensor is integrated in the time domain to obtain the second tamping plate amplitude i S2 , according to formula i S =(i S1 +i S2 ) / 2 to obtain the tamping plate amplitude i S ;

[0021] According to the surface vibration compactor test method in the "Highway Geotechnical Test Code", the maximum dry density of the filler ρ is obtained. dmaxs and the dry density of the test filler ρ ds , the compaction degree K of the roadbed filler in the indoor test is obtained according to the following formula: S ;

[0022]

[0023] Record the exciting force L of the surface vibration compactor in the indoor test s , excitation frequency f S , tamping plate amplitude i S , sedimentation rate s S , roadbed filler compaction degree K S ; At the same time, directly record the soil pressure c of the bottom center soil pressure sensor S11 , Earth pressure of bottom edge earth pressure sensor c S12 、Middle center soil pressure sensor soil pressure c S21 , Earth pressure sensor at the middle edge Earth pressure c S22 、Top center earth pressure sensor earth pressure c S31 , top edge earth pressure sensor earth pressure c S32 , temperature sensor temperature g S 、Humidity sensor humidity e S ; and record the roadbed filling parameters of the indoor test, the roadbed filling parameters of the indoor test include soil type u S , soil particle gradation v S , rock content w S , block stone gradation x S , stone aspect ratio y S , moisture contentω S ;

[0024] The large amount of data recorded above constitutes the data set for indoor experiments.

[0025] The specific implementation process of step 2 is:

[0026] The field test uses a roller, and n laser displacement sensors are respectively arranged at equal heights on the front and rear sides of the roller through mounting frames. A vertical acceleration sensor and a harmonic wave ratio CMV sensor are installed on the central axis of the vibrating steel wheel of the roller. The installation directions of the vertical acceleration sensor and the harmonic wave ratio CMV sensor are consistent with the vibration direction of the steel wheel of the roller. A temperature sensor 104 and a humidity sensor for measuring ground temperature and humidity are installed on the roller. During the filling paving process, six soil pressure sensors are pre-buried at the top, middle and bottom positions of the test strip cross section every 25m along the compaction direction. The six soil pressure sensors are arranged in three layers, with two arranged on each layer. In the same layer, one soil pressure sensor is located at the center, and the other soil pressure sensor is located at the edge. All soil pressure sensors located at the edge are at the same horizontal position.

[0027] The arrangement of earth pressure sensors in the field test is consistent with that in the indoor test;

[0028] Install GPS on the roof of the roller;

[0029] At the beginning of the field test, all laser displacement sensors at the front and rear of the roller were adjusted to zero. During the roller's forward movement, the real-time value of the laser displacement sensor at the front from left to right was recorded as h. 11 、h 12 、h 13 ,…,h 1n The real-time value of the laser displacement sensor from left to right on the rear side is recorded as h 21 、h 22 、h 23 ,…,h 2n ; During the roller's backward movement, the real-time values ​​of the laser displacement sensors from left to right on the front side are recorded as h1'1, h1'2, h1'3, ..., h1' n The real-time value of the laser displacement sensor from left to right on the rear side is recorded as h' 21 、h' 22 、h' 23 ,…,h' 2n ; The acceleration of the vertical acceleration sensor on the roller is recorded as a x ;

[0030] Get the acceleration signal a during the roller loading process x , the acceleration signal of the roller during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f x , the acceleration signal of the roller during loading is integrated in the time domain to obtain the vibration steel wheel amplitude i of the first roller x1 ;

[0031] The position and movement direction of the roller are obtained by GPS positioning, and the settlement amount h of the roller during the forward movement is QJfor:

[0032]

[0033] The amount of settlement of the roller during the backward movement h HT for:

[0034]

[0035] Wherein, n is the number of single-side laser displacement sensors, and n is a natural number;

[0036] The settlement rate of each compaction is s x for: H x is the paving height of the roadbed fill, which is the set value of the field test;

[0037] When the roller moves forward Calculate the electrical signal collected by the laser displacement sensor on the front side of the roller. Calculate the electrical signal collected by the laser displacement sensor on the front side of the roller;

[0038] The electrical signal h collected by the laser displacement sensor on the front side of the roller Q Perform time domain integration to obtain the vibration steel wheel amplitude i of the second roller x2 According to formula i x =(i x1 +i x2 ) / 2 to get the vibration amplitude i of the roller steel wheel x ;

[0039] In the field test, the total volume of the paved filler is calculated based on the total volume of the paved filler per 100m 3 Take 30kg of filler for screening, measurement and weighing test, determine and record the parameters of the roadbed filler for the field test, the parameters of the roadbed filler for the field test include: soil type u x , soil particle gradation v x , rock content w x , block stone gradation x x , stone aspect ratio y x , moisture contentω x ;

[0040] In the field test, after each compaction of the roadbed, the strip was tested with sand filling method every 25m to obtain the maximum dry density ρ of the roadbed filler after each compaction. dmaxx and roadbed filler dry density ρ dx , the compaction degree K of the roadbed filler in the field test is obtained according to the following formula: x

[0041]

[0042] Record the rated exciting force L of the roller in the field test x , excitation frequency f x , roller vibration steel wheel amplitude i x , the settlement rate of each compaction s x , roadbed filler compaction degree K x , while directly recording the soil pressure c of the bottom center soil pressure sensor in the field test x11 , Earth pressure of bottom edge earth pressure sensor c x12 、Middle center soil pressure sensor soil pressure c x21 , Earth pressure sensor at the middle edge Earth pressure c x22 、Top center earth pressure sensor earth pressure c x31 , top edge earth pressure sensor earth pressure c x32 , temperature sensor temperature g x 、Humidity sensor humidity e x ;

[0043] The large amount of data recorded above constitutes the data set for the field experiment.

[0044] The process of adjusting the roadbed compaction quality prediction model based on virtual and real data collaborative training is:

[0045] Set up 1 to 2 test strips to conduct field tests on the roadbed compaction, obtain the roadbed filling parameters of the test strips, and the soil pressure c of the bottom center soil pressure sensor 11 , Earth pressure of bottom edge earth pressure sensor c 12 、Middle center soil pressure sensor soil pressure c 21 , Earth pressure sensor at the middle edge Earth pressure c 22 、Top center earth pressure sensor earth pressure c 31 , top edge earth pressure sensor earth pressure c 32 , the temperature sensor temperature g, the humidity sensor humidity e, the excitation frequency f, the vibration amplitude of the roller steel wheel i, and the settlement rate of each compaction s are input into the roadbed compaction quality prediction model based on virtual and real data collaborative training obtained in step 3. The roadbed compaction quality prediction model based on virtual and real data collaborative training is recorded as P i , after P i After processing, the real-time predicted compaction degree K of the compacted strip is obtained. YC After each rolling of the test strip, the compaction degree K is obtained by the sand filling method. SY , according to the formula Calculate the error R. When the error R is greater than 5%, continue the sand filling test during the next rolling, and use the compaction degree K obtained by this sand filling test as the value of the test result. SY and the corresponding model input data are used to iThe updated training optimization is performed again to obtain the updated roadbed compaction quality prediction model P based on virtual and real data collaborative training. i+1 , until the error R is less than or equal to 5%, the optimized and calibrated roadbed compaction quality prediction model P based on virtual and real data collaborative training is obtained. L , the adjustment is completed.

[0046] The specific value of n in the present invention can be determined according to the wheel width of the roller, and one is installed every 15-25 cm, generally not more than 12, and preferably n=6-12.

[0047] In a second aspect, the present invention provides a roadbed compaction quality monitoring and maintenance system based on virtual and real indicator collaboration, including a data real-time monitoring and acquisition module, a data analysis and processing module, and a roadbed compaction quality evaluation and maintenance module;

[0048] The data real-time monitoring and acquisition module is used to perform real-time monitoring and data acquisition of external environmental parameters and monitoring indicators during laboratory and roadbed compaction field tests and applications, and transmit the data to the data analysis and processing module; the external environmental parameters include temperature and humidity, and the monitoring indicators include vertical acceleration, harmonic ratio CMV, settlement, and earth pressure;

[0049] The data analysis and processing module is used to analyze and process the data collected by the real-time data monitoring and collection module according to conditions to obtain the virtual and real evaluation index of the roadbed compaction quality, and the virtual and real evaluation index of the roadbed compaction quality includes settlement rate, excitation frequency and amplitude;

[0050] The roadbed compaction quality evaluation and maintenance module is used to input the data obtained by the real-time data monitoring and acquisition module and the data analysis and processing module into the roadbed compaction quality prediction model, obtain the roadbed compaction degree prediction value in real time, and display it visually. The compaction quality is evaluated according to the roadbed compaction degree prediction value, and the under-compacted points are supplemented with compaction. The next step of roadbed compaction work is guided according to the visualized roadbed compaction degree prediction value.

[0051] The data real-time monitoring and acquisition module includes indoor test real-time monitoring and acquisition equipment and field test real-time monitoring and acquisition equipment. The indoor test real-time monitoring and acquisition equipment includes a surface vibration compactor, an indoor test monitoring data acquisition terminal, two laser displacement sensors capable of irradiating laser on the tamping plate of the surface vibration compactor, two vertical acceleration sensors installed at the center and edge of the tamping plate of the surface vibration compactor, a harmonic ratio CMV sensor installed on the tamping plate of the surface vibration compactor, and a temperature sensor and a humidity sensor for monitoring external environmental parameters.

[0052] The paving thickness is set, and six soil pressure sensors are installed in the paving area of ​​the test tube of the surface vibration compactor according to the top, middle and bottom positions. The six soil pressure sensors are arranged in three layers, with two arranged in each layer. In the same layer, one soil pressure sensor is located in the center, and the other soil pressure sensor is located at the edge. All soil pressure sensors located at the edge are at the same horizontal position; the indoor test monitoring data acquisition terminal is electrically connected to all soil pressure sensors, temperature sensors and humidity sensors, vertical acceleration sensors, laser displacement sensors, and harmonic ratio CMV sensors in the indoor test real-time monitoring and acquisition equipment;

[0053] The real-time monitoring and acquisition equipment for the on-site test includes a roller and an on-site monitoring data acquisition terminal. n laser displacement sensors are respectively set at equal heights on the front and rear sides of the roller through mounting frames. A vertical acceleration sensor and a harmonic wave ratio CMV sensor are installed on the central axis of the roller's vibrating steel wheel. A temperature sensor and a humidity sensor for measuring ground temperature and humidity are installed on the roller. During the filling paving process, six soil pressure sensors are pre-buried at the top, middle, and bottom positions of the test strip cross section every 25m along the compaction direction. The arrangement of the six soil pressure sensors pre-buried in each cross section of the on-site test is consistent with the arrangement of the soil pressure sensors in the indoor test.

[0054] Install GPS on the roof of the roller, install a field monitoring data collection terminal inside the roller, the field monitoring data collection terminal is electrically connected to all soil pressure sensors, temperature sensors and humidity sensors, vertical acceleration sensors, laser displacement sensors, harmonic ratio CMV sensors in the field test real-time monitoring and collection equipment, and the field monitoring data collection terminal is electrically connected to the portable power supply;

[0055] The indoor test monitoring data collection terminal and the field monitoring data collection terminal are connected to the base station via wireless, and the base station performs different data processing according to the data source of the indoor test monitoring data collection terminal or the field monitoring data collection terminal;

[0056] The roadbed compaction quality monitoring and maintenance system based on virtual and real indicator collaboration also includes a mobile terminal, which is connected to the website through the network to view data collection status and roadbed compaction quality status in real time;

[0057] The base station is used to complete data analysis and processing and roadbed compaction quality evaluation, and upload the data to the website.

[0058] The functions of the data analysis and processing module during the indoor test are:

[0059] Acceleration signal analysis and processing: The acceleration signal during loading is calculated based on the acceleration of the two vertical acceleration sensors, and the acceleration signal during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency fS , the acceleration signal is integrated in the time domain during the loading process to obtain the first tamping plate amplitude i S1 ;

[0060] Laser displacement sensor analysis and processing: Calculate the real-time sedimentation rate s of the filler based on the real-time values ​​of the laser displacement sensors on both sides and the filling height of the filler S , where the filling height of the filler is a set known value; the electrical signal collected by the laser displacement sensor is integrated in the time domain to obtain the tamping plate amplitude i S2 , according to formula i S =(i S1 +i S2 ) / 2 to obtain the tamping plate amplitude i S ;

[0061] Calculation of compaction degree of roadbed filler: According to the surface vibration compactor test method in the "Highway Geotechnical Test Code", the maximum dry density ρ of the filler is obtained. dmaxs and the dry density of the test filler is ρ ds , the compaction degree K of the roadbed filler in the indoor test is obtained according to the following formula: S ;

[0062]

[0063] Record the exciting force L of the surface vibration compactor in the indoor test s , excitation frequency f S , tamping plate amplitude i S , sedimentation rate s S , roadbed filler compaction degree K S ; At the same time, directly record the soil pressure c of the bottom center soil pressure sensor S11 、Earth pressure of the earth pressure sensor at the right edge of the bottom c S12 、Middle center soil pressure sensor soil pressure c S21 , Earth pressure sensor at the middle edge Earth pressure c S22 、Top center earth pressure sensor earth pressure c S31 , top edge earth pressure sensor earth pressure c S32 , temperature sensor temperature g S 、Humidity sensor humidity e S ; and record the roadbed filling parameters of the indoor test, the roadbed filling parameters of the indoor test include soil type u S , soil particle gradation v S , rock content w S , block stone gradation x S , stone aspect ratio y S , moisture contentω S ;

[0064] The functions of the data analysis and processing module during the field test are:

[0065] Acceleration signal analysis and processing: Obtain the acceleration signal a during the roller loading process x , the acceleration signal of the roller during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f x , the acceleration signal of the roller during loading is integrated in the time domain to obtain the vibration steel wheel amplitude i of the first roller x1 ;

[0066] Laser displacement sensor analysis and processing: GPS positioning is used to obtain the roller's position and direction of movement, and the roller's settlement amount h during its forward movement. QJ :

[0067]

[0068] The amount of settlement of the roller during the backward movement h HT :

[0069]

[0070] Where n is the number of single-side laser displacement sensors;

[0071] Sedimentation rate per compaction pass: H x is the paving height of the roadbed fill;

[0072] The electrical signal h collected by the laser displacement sensor on the front side of the roller is obtained according to the following formula: Q

[0073] or

[0074] The electrical signal h collected by the laser displacement sensor on the front side of the roller Q Perform time domain integration to obtain the vibration steel wheel amplitude i of the second roller x2 According to formula i x =(i x1 +i x2 ) / 2 to get the vibration amplitude i of the roller steel wheel x ;

[0075] Calculation of compaction degree of roadbed filler: In the field test, the total volume of the paved filler is calculated based on the compaction degree of each 100m 3 Take 30kg of filler for screening, measurement and weighing test, determine and record the parameters of the roadbed filler for the field test, the parameters of the roadbed filler for the field test include soil type u x , soil particle gradation v x , rock content w x , block stone gradation xx , stone aspect ratio y x , moisture contentω x ;

[0076] In the field test, after each compaction of the roadbed, the strip was tested with sand filling method every 25m to obtain the maximum dry density ρ of the roadbed filler after each compaction. dmaxx and roadbed filler dry density ρ dx , the compaction degree K of the roadbed filler in the field test is obtained according to the following formula: x ;

[0077]

[0078] Record the rated exciting force L of the roller in the field test x , excitation frequency f x , roller vibration steel wheel amplitude i x , the settlement rate of each compaction s x , roadbed filler compaction degree K x , while directly recording the soil pressure c of the bottom center soil pressure sensor in the field test x11 , Earth pressure of bottom edge earth pressure sensor c x12 、Middle center soil pressure sensor soil pressure c x21 , Earth pressure sensor at the middle edge Earth pressure c x22 、Top center earth pressure sensor earth pressure c x31 , top edge earth pressure sensor earth pressure c x32 , temperature sensor temperature g x 、Humidity sensor humidity e x .

[0079] The roadbed compaction quality prediction model takes 12 characteristic values, including filler soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, water content, exciting force, exciting frequency, amplitude, harmonic ratio CMV, soil pressure at different positions, and settlement rate, as input, and is composed of a neural network with compaction degree as output.

[0080] The base station is loaded with a field-laboratory roadbed compaction intelligent collaborative perception program, and the field-laboratory roadbed compaction intelligent collaborative perception program includes an indoor test condition deployment subprogram.

[0081] The indoor test condition allocation subroutine is used to randomly allocate the roadbed filling parameters, roller working parameters, and external environmental parameters in the indoor test to generate a large number of test condition combinations; the roadbed filling parameters include soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, and moisture content; the external environmental parameters include temperature and humidity; the roller working parameters include exciting force and exciting frequency; the indoor test condition allocation subroutine has a collection of various roadbed filling parameters, various roller working parameters, and various external environmental parameters;

[0082] Obtaining the energy conversion relationship between the roller and the surface vibration compactor, and obtaining the working parameters of the surface vibration compactor according to the energy conversion relationship between the roller and the surface vibration compactor, so that the working parameters of the surface vibration compactor in the indoor test are consistent with the working parameters of the roller on site;

[0083] The external environmental parameters of the indoor test are set according to the external environmental parameters of the on-site roadbed compaction conditions.

[0084] Reproduce the roadbed filling parameters indoors according to the actual situation of the filling on site.

[0085] The external environmental parameters of the on-site roadbed compaction conditions, roadbed filling parameters, and roller working parameters are obtained by inputting the indoor test condition adjustment subroutine. The indoor test condition adjustment subroutine outputs multiple groups of all parameters required for indoor tests under the current on-site roadbed compaction test conditions, thereby realizing intelligent collaborative perception of on-site-laboratory roadbed compaction.

[0086] Compared with the prior art, the present invention has the following beneficial effects:

[0087] 1. The present invention establishes a laboratory test to reproduce the on-site roadbed compaction conditions, realizes the coordinated reproduction of indoor and field tests, and reproduces the roadbed filling parameters, roller working parameters, external environmental parameters and other related parameters of the field test in the laboratory test, accurately controls the test conditions, and monitors and collects test data in real time and comprehensively. Based on the on-site-laboratory roadbed compaction intelligent collaborative perception program, the on-site roadbed compaction conditions are reproduced in the laboratory, and the on-site roadbed compaction virtual and real indicator data are monitored and collected in real time, and an accurate, comprehensive and reliable corresponding data set containing indoor and outdoor virtual and real indicators is established.

[0088] 2. The intelligent collaborative perception program of on-site-laboratory roadbed compaction of the present invention can realize the collaborative interaction of laboratory and on-site roadbed compaction conditions; the system can monitor the virtual and real, macro and micro dynamic indicators in indoor and outdoor tests and roadbed compaction applications in real time and collect and process data. The roadbed filling parameters, roller working parameters, external environmental parameters, monitoring indicators in the laboratory and on-site roadbed compaction process, such as "virtual" feedback indicators amplitude, harmonic ratio CMV, excitation frequency and intuitive physical and mechanical indicators settlement, micro-indicator soil pressure and other data can be collected, uploaded, analyzed and processed in real time to realize real-time monitoring and management and maintenance of roadbed compaction quality.

[0089] 3. The method of the present invention effectively applies the roadbed compaction quality prediction model in engineering practice, making the roadbed compaction quality prediction model closer to the actual application scenario. Before the on-site roadbed compaction, a test strip is set up to collect test data and monitor dynamic virtual and real macro and micro indicators. Then, the compaction prediction value obtained by the roadbed compaction quality prediction model is compared with the compaction test value obtained by the traditional detection method. The model is optimized and adjusted again based on the error between the two. After reaching the accuracy requirements, it is put into use, and the roadbed compaction management and maintenance is realized based on the real-time monitored compaction.

[0090] 4. The present invention establishes data sets for field tests and indoor tests based on indoor and outdoor collaborative reproduction of on-site compaction of roadbed. Laboratory tests reproduce on-site roadbed compaction conditions, reduce on-site testing, reduce difficulty, and reduce workload. In indoor and outdoor collaborative tests, real-time monitoring and collection of virtual and real indicator data are carried out, and feedback "virtual" indicators and intuitive physical and mechanical indicators in the filler compaction process are monitored and collected simultaneously. Real-time monitoring and data collection of virtual and real indicators, multiple sensors (vertical acceleration sensor, laser displacement sensor, harmonic ratio CMV sensor, soil pressure sensor, temperature sensor, humidity sensor) work simultaneously to achieve real-time monitoring and data collection of multiple virtual and real indicators, and establish a roadbed compaction quality prediction model that coordinates virtual and real indicators. The factors considered are more comprehensive, the accuracy is improved, and it is more suitable for the working environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 It is a flow chart of the roadbed compaction quality monitoring and maintenance method based on the coordination of virtual and real indicators of the present invention.

[0092] Figure 2 This is a flow chart of the working principle of on-site roadbed compaction based on the roadbed compaction quality monitoring and maintenance method coordinated with virtual and real indicators.

[0093] Figure 3 This is a schematic diagram of the installation of the indoor surface vibration compaction tester.

[0094] (In the figure, 101 - laser displacement sensor, 102 - vertical acceleration sensor, 103 - harmonic ratio CMV sensor, 104 - temperature sensor, 105 - humidity sensor, 106 - soil pressure sensor, 107 - indoor test monitoring data acquisition terminal, 001 tamping plate, 002 test tube, 003 bracket, 004 laptop computer)

[0095] Figure 4 This is a schematic diagram of the installation of the on-site roadbed compaction roller equipment.

[0096] (In the figure, 101 - laser displacement sensor, 201 - laser displacement sensor mounting bracket, 102 - vertical acceleration sensor, 103 - harmonic ratio CMV sensor, 104 - temperature sensor, 105 - humidity sensor, 106 - soil pressure sensor, 202 - soil pressure wireless receiver, 203 - GPS, 204 - field monitoring data acquisition terminal)

[0097] Figure 5 It is a structural diagram of the artificial neural network roadbed compaction quality prediction model. DETAILED DESCRIPTION

[0098] In order to further understand the content, features and functions of the present invention, the following are specific embodiments of the present invention. The specific embodiments are only used to further illustrate the present invention and do not limit the protection scope of the claims of this application.

[0099] The roadbed compaction quality monitoring and maintenance system based on the coordination of virtual and real indicators of the present invention is an intelligent compaction integrated device. The roadbed compaction quality monitoring and maintenance based on the coordination of virtual and real indicators is implemented in the intelligent compaction integrated device in the application, and the device includes a real-time data monitoring and acquisition module, a data analysis and processing module, and a roadbed compaction quality evaluation and maintenance module. The real-time data monitoring and acquisition module refers to the real-time monitoring and data acquisition of external environmental parameters (such as temperature, humidity) and monitoring indicators (such as vertical acceleration, harmonic ratio CMV, settlement, soil pressure) during the laboratory and on-site roadbed compaction process and transmits the above data to the data analysis and processing module. The data analysis and processing module refers to the analysis and processing of the data collected by the real-time data monitoring according to the conditions to obtain the virtual and real evaluation indicators of the roadbed compaction quality (such as settlement rate, soil pressure at different positions, excitation frequency, amplitude). The roadbed compaction quality evaluation and maintenance module refers to inputting the data obtained by the real-time data monitoring and acquisition module and the data analysis and processing module into the roadbed compaction quality prediction model, obtaining the roadbed compaction degree prediction value in real time, and visually displaying it. The compaction quality is evaluated according to the roadbed compaction degree prediction value, and the under-compacted points are supplemented with compaction while avoiding the occurrence of over-compacting. The next step of roadbed compaction is guided according to the visual prediction value. That is, if the compaction degree required by the project is not met, the current compaction work does not meet the compaction requirements and needs to be rolled again. If it is met, the rolling of the current position is stopped to achieve construction quality control of the entire roadbed compaction process.

[0100] The data real-time monitoring and acquisition module includes indoor test real-time monitoring and acquisition equipment and on-site roadbed compaction real-time monitoring and acquisition equipment. The indoor test real-time monitoring and acquisition equipment includes a surface vibration compactor, an indoor test monitoring data acquisition terminal 107, two laser displacement sensors 101 capable of irradiating laser on the tamping plate 001 of the surface vibration compactor, two vertical acceleration sensors 102 installed at the center and edge of the tamping plate of the surface vibration compactor, a harmonic ratio CMV sensor 103 installed on the tamping plate of the surface vibration compactor, a temperature sensor 104 and a humidity sensor 105 for monitoring external environmental parameters, setting the paving thickness, and vertically placing six soil pressure sensors 106 at the bottom center and one side edge, the middle center and one side edge, and the top center and one side edge of the test tube 002 of the surface vibration compactor. The indoor test monitoring data acquisition terminal 107 is electrically connected to all the above sensors;

[0101] The on-site roadbed compaction real-time monitoring and collection equipment includes a roller and an on-site monitoring data collection terminal 204. n laser displacement sensors 101 are respectively arranged at equal heights on the front and rear sides of the roller through mounting frames. A vertical acceleration sensor 102 and a harmonic wave ratio CMV sensor 103 are installed on the central axis of the roller's vibrating steel wheel. The installation directions of the two sensors are consistent with the vibration direction of the roller's steel wheel. A temperature sensor 104 and a humidity sensor 105 for measuring ground temperature and humidity are installed on the roller. During the filling paving process, six soil pressure sensors 106 are pre-buried at the bottom center and one side edge, the middle center and one side edge, and the top center and one side edge of the test strip cross section every 25m along the compaction direction. The soil pressure sensors at the bottom, middle, and top are located on the same side, which is consistent with the arrangement of soil pressure sensors in indoor tests.

[0102] The top soil pressure sensor of indoor test and on-site roadbed compaction should be 1-2cm lower than the surface of filler and roadbed to prevent damage caused by direct rolling on the soil pressure sensor;

[0103] A GPS 203 is installed on the roof of the roller, and a field monitoring data collection terminal 204 is installed inside the roller. The field monitoring data collection terminal 204 is electrically connected to all sensors in the field roadbed compaction real-time monitoring and collection equipment, and the field monitoring data collection terminal 204 is electrically connected to a portable power source;

[0104] The indoor test monitoring data collection terminal 107 and the field monitoring data collection terminal 204 transmit the collected data to the base station through wireless connection, and perform different data processing according to the data source laboratory or field. The mobile terminal (mobile phone or computer) can connect to the corresponding website through the network to view the data collection status and roadbed compaction quality in real time.

[0105] The function of the base station is to complete data analysis and processing and roadbed compaction quality prediction and evaluation.

[0106] The indoor test monitoring data collection terminal 107 and the field monitoring data collection terminal 204 are used to collect data from each sensor and transmit the data to the base station.

[0107] The data analysis and processing module refers to the analysis, processing and calculation of the test data collected by the real-time data monitoring and acquisition module in the data analysis and processing module and the roadbed compaction quality evaluation and maintenance center base station according to the data source laboratory or on-site.

[0108] The roadbed compaction quality evaluation and maintenance module inputs the collected and analyzed data into the roadbed compaction quality prediction model to obtain the real-time compaction degree of the roadbed filling material, and performs additional compaction at points that have not reached the compaction strength, while avoiding the occurrence of over-pressure, and accurately constructs and manages maintenance. The roadbed compaction quality evaluation and maintenance module is completed in the base station.

[0109] The intelligent compaction integrated device is loaded with a field-laboratory roadbed compaction intelligent collaborative perception program.

[0110] The indoor test condition allocation subroutine is used to randomly allocate roadbed filling parameters, roller working parameters, and external environmental parameters in the indoor test to generate a large number of test condition combinations; the indoor test condition allocation subroutine has a collection of various soil types, various soil particle gradations, various block stone contents, various block stone gradations, various block stone aspect ratios, various moisture contents, various exciting forces, various exciting frequencies, various temperatures, and various humidity;

[0111] First, according to the existing conditions, a large number of on-site roadbed compaction conditions are collected, including: roadbed filling parameters are at least one of soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, and water content; roller working parameters are at least one of exciting force and exciting frequency; external environment parameters are at least one of temperature and humidity;

[0112] Obtaining the energy conversion relationship between the roller and the surface vibration compactor, and adjusting the working parameters of the surface vibration compactor according to the energy conversion relationship between the roller and the surface vibration compactor, so that the working parameters of the surface vibration compactor in the indoor test are consistent with the working parameters of the roller on site;

[0113] The external environmental parameters of the indoor test are set according to the external environmental parameters of the on-site roadbed compaction conditions.

[0114] Reproduce the roadbed filling parameters indoors according to the actual situation of the roadbed compaction filling on site.

[0115] The indoor test condition adjustment subroutine is used to input the above parameters known according to the on-site roadbed compaction conditions, and the indoor test condition adjustment subroutine outputs multiple groups of all parameters of the indoor test corresponding to the current on-site roadbed compaction test conditions, so as to realize the intelligent collaborative perception of on-site-laboratory roadbed compaction;

[0116] A large number of roadbed filling parameters, roller working parameters, and external environmental parameters are obtained by using the indoor test random allocation subroutine to generate a large number of test condition combinations;

[0117] In the process of collecting data, there are cases where the above data are not fully recorded for the relevant on-site roadbed compaction. Based on the above relevant data that can be collected, the indoor test condition allocation subroutine is used to determine the relevant parameters of the indoor test corresponding to the current on-site roadbed compaction test conditions. The indoor test condition allocation subroutine can allocate fillers according to the generated test condition combination; adjust the working parameters of the surface vibration compactor according to the energy conversion relationship between the roller and the surface vibration compactor; and adjust the environmental conditions. Carry out indoor surface vibration compaction tests according to the reproducibility requirements.

[0118] The present invention provides a real-time monitoring and maintenance method for roadbed compaction quality based on virtual and real index coordination. The implementation process of the method is as follows: Figure 1 As shown in the figure, 1) establish an indoor compaction test that reproduces the on-site roadbed compaction conditions and establish a data set for the indoor test; 2) conduct a field test on roadbed compaction and establish a data set for the field test; 3) establish a roadbed compaction quality prediction model based on collaborative training of virtual and real data, and 4) apply the roadbed compaction quality prediction model to construction control based on collaborative prediction of virtual and real indicators.

[0119] The specific implementation process of the real-time monitoring and maintenance method of roadbed compaction quality based on the coordination of virtual and real indicators is as follows:

[0120] Step 1: Establish an indoor compaction test that reproduces the on-site roadbed compaction conditions and establish a data set for the indoor test;

[0121] Step 1: Specific implementation process: Use the indoor test condition adjustment subroutine to output the laboratory surface vibration compaction test conditions, including roadbed filling parameters, roller working parameters and external environmental parameters. The roadbed filling parameters include filling soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, and moisture content; the roller working parameters include exciting force and exciting frequency; the external environmental parameters include temperature and humidity;

[0122] According to the energy conversion relationship between the roller and the surface vibration compactor, the working parameters of the surface vibration compactor are adjusted, the counterweight of the surface vibration compactor tamping plate is added or subtracted, the vibration motor inverter is adjusted, and the vibration force and vibration frequency of the surface vibration compactor are changed. The relative positions of the sensors installed on the surface vibration compactor in the indoor test real-time monitoring collection equipment are as follows Figure 3As shown: (1) The laser displacement sensor 101 is installed on the brackets 003 on both sides of the surface vibration compactor. The laser can be irradiated vertically on the tamping plate of the surface vibration compactor. The accuracy of the laser displacement sensor should be greater than 0.001mm. (2) Vertical acceleration sensor 102 is vertically mounted at the center and edge of the tamping plate of the surface vibration compactor; (3) Harmonic ratio CMV sensor 103 is vertically mounted on the tamping plate of the surface vibration compactor; (4) Temperature sensor 104 is vertically mounted on a bracket on one side of the surface vibration compactor; (5) Humidity sensor 105 is mounted on a bracket on one side of the surface vibration compactor; (6) Soil pressure sensor 106 is vertically mounted during the process of filling the test tube with filler, and a total of 6 sensors are mounted at the center and edge of the bottom, center and edge of the middle, and center and edge of the top of the test tube; (7) Indoor test monitoring data acquisition terminal 107, the indoor test monitoring data acquisition terminal is loaded with data acquisition software for each sensor and is equipped with a removable power supply for powering itself and each sensor, and can obtain real-time data collected by each sensor. The indoor test monitoring data acquisition terminal is placed on one side of the surface vibration compactor and does not affect the experimental operation; all sensors are connected to the indoor test monitoring data acquisition terminal 107 via wired or wireless connections. The indoor test monitoring data acquisition terminal 107 is equipped with a mobile phone card, which can transmit real-time data to the base station by connecting to the mobile network. The laptop computer 004 logs on to the relevant website to realize the real-time visualization of the sensor collected data.

[0123] In this embodiment, the base station in the intelligent compaction integrated equipment is fixed and connected via a network, and does not need to be moved during the test. The intelligent compaction integrated equipment is loaded with a data analysis and processing module and a roadbed compaction quality evaluation and maintenance module.

[0124] Start the indoor test, the indoor test real-time monitoring and acquisition equipment is powered on, and the laser displacement sensors 101 on both sides are zeroed. During the test, the real-time value h1 of the left laser displacement sensor, the real-time value h2 of the laser displacement sensor, the acceleration a1 of the vertical acceleration sensor at the center of the tamping plate, and the acceleration a2 of the vertical acceleration sensor at the edge of one side of the tamping plate are recorded. The above data are transmitted to the base station via the indoor test monitoring data acquisition terminal 107.

[0125] The functions of the data analysis and processing module during the indoor test are:

[0126] Acceleration signal analysis and processing: The acceleration signal during loading is calculated based on the acceleration of the two vertical acceleration sensors, and the acceleration signal during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f S , the acceleration signal is integrated in the time domain during the loading process to obtain the first tamping plate amplitude i S1 ;

[0127] Laser displacement sensor analysis and processing: Calculate the real-time sedimentation rate s of the filler based on the real-time values ​​of the laser displacement sensors on both sides and the filling height of the filler S , where the filling height of the filler is a set known value; the electrical signal collected by the laser displacement sensor is integrated in the time domain to obtain the second tamping plate amplitude i S2 , obtained by laser displacement sensor S2 And the acceleration signal time domain integration i S1 Co-verification of the tamping plate amplitude: S =(i S1 +i S2 ) / 2, at this time i S is the tamping plate amplitude;

[0128] Calculation of compaction degree of roadbed filler: According to the surface vibration compactor test method in the "Highway Geotechnical Test Code", the maximum dry density ρ of the filler is obtained. dmaxs and the dry density of the test filler ρ ds , the compaction degree K of the roadbed filler in the indoor test is obtained according to the following formula: S .

[0129]

[0130] Record the surface vibration compactor of the indoor test and set the excitation force L s , excitation frequency f S , amplitude i S , sedimentation rate s S , roadbed filler compaction degree K S ; At the same time, directly record the soil pressure c of the bottom center soil pressure sensor S11 , Earth pressure of bottom edge earth pressure sensor c S12 、Middle center soil pressure sensor soil pressure c S21 , Earth pressure sensor at the middle edge Earth pressure c S22 、Top center earth pressure sensor earth pressure c S31 , top edge earth pressure sensor earth pressure c S32 , temperature sensor temperature g S 、Humidity sensor humidity e S ; and record the roadbed filling parameters of the indoor test, which include soil type u S , soil particle gradation v S , rock content w S , block stone gradation x S , stone aspect ratio y S (Take 10% of the total amount of filler blocks, measure the length ratio of the long side and the short side of each block and find the average to obtain the length-to-diameter ratio of the block y S ), moisture content ω SThe large amount of data recorded above constitutes the data set for the indoor test.

[0131] Among them, the acceleration signal a during loading S :

[0132] Real-time sedimentation rate of filler S : H S is the filling height of the filler.

[0133] Step 2: Conduct field tests on roadbed compaction and establish a data set for the field tests

[0134] Step 2: Specific implementation process: On-site roadbed compaction test real-time monitoring and collection equipment, the relative positions of the sensors installed on the roller are as follows: Figure 4 As shown: (1) Two laser displacement sensor mounting brackets 201 are installed at the same horizontal height in front and behind the roller; (2) The laser displacement sensor 101 is installed on the laser displacement sensor mounting bracket, one is installed at each end of the laser displacement sensor mounting bracket, and one is installed every 20 cm in the middle part. If it is less than 20 cm, it will be calculated as 20 cm. A total of n laser displacement sensors are installed, and generally no more than 12 on one side; (3) The vertical acceleration sensor 102 is vertically placed on the central axis of the roller's vibrating steel wheel, in the same direction as the roller's vibration direction; (4) The harmonic ratio CMV sensor 103 is vertically placed on the central axis of the roller's vibrating steel wheel, in the same direction as the roller's vibration direction; (5) The temperature sensor 104 is perpendicular to the road base surface and is adsorbed on the roller; (6) The humidity sensor 105 is adsorbed on the roller On the front crossbeam; (7) Earth pressure sensor 106. During the filling process, six earth pressure sensors are embedded in the bottom center, edge, middle center, edge and top center, edge of the cross section of the test strip (the width of the test strip is consistent with the wheel width) every 25 meters. The top earth pressure sensor should be 1-2 cm lower than the roadbed surface to prevent the vibrating steel wheel from directly rolling on the earth pressure sensor and causing damage; (8) Earth pressure wireless receiver 202 is vertically adsorbed on the rear cover of the roller; (9) GPS 203 is vertically adsorbed on the roof of the roller; (10) The data acquisition software of each sensor is loaded in the field monitoring data acquisition terminal 204, and is equipped with a movable power supply for powering each sensor, which can obtain real-time data collected by each sensor and is adsorbed and fixed on the rear side of the roller cab without affecting the operation;

[0135] The soil pressure sensor is wirelessly connected to the field monitoring data collection terminal 204, and the other sensors are wired to the field monitoring data collection terminal 204. The field monitoring data collection terminal 204 is equipped with a mobile phone card, which can transmit real-time data to the base station by connecting to the mobile network. The laptop computer logs on to the relevant website to realize the real-time visualization of the sensor collection data and the roadbed compaction degree.

[0136] The field test was started. The real-time monitoring and acquisition equipment of the field roadbed compaction test was powered on. All laser displacement sensors in front and behind the roller were adjusted to zero. During the roller's forward movement, the real-time value of the laser displacement sensor on the front side from left to right was recorded as h. 11 、h 12 、h 13 ......h 1n The real-time value of the laser displacement sensor from left to right on the rear side is recorded as h 21 、h 22 、h 23 ......h 2n ; During the roller's backward movement, the real-time values ​​of the laser displacement sensors from left to right on the front side are recorded as h1'1, h1'2, h1'3...h1' n The real-time value of the laser displacement sensor from left to right on the rear side is recorded as h' 21 、h' 22 、h' 23 ......h' 2n ; The acceleration of the vertical acceleration sensor on the roller is recorded as a x , the above data is transmitted to the data analysis and processing module for analysis and calculation to obtain:

[0137] Acceleration signal analysis and processing: Obtain the acceleration signal a during the roller loading process x , the acceleration signal of the roller during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f x , the acceleration signal of the roller during loading is integrated in the time domain to obtain the vibration steel wheel amplitude i of the first roller x1 ;

[0138] Laser displacement sensor analysis and processing: GPS positioning can be used to determine the roller's position and direction of movement, as well as the roller's settlement amount during its forward movement. QJ :

[0139]

[0140] The amount of settlement of the roller during the backward movement h HT :

[0141]

[0142] Where n is the number of single-side laser displacement sensors;

[0143] Sedimentation rate per compaction pass: H x is the paving height of the roadbed filler, which is the set value of the on-site roadbed compaction test, that is, a known value;

[0144] When the roller moves forward Calculate the electrical signal collected by the laser displacement sensor on the front side of the roller. Calculate the electrical signal collected by the laser displacement sensor on the front side of the roller;

[0145] The electrical signal h collected by the laser displacement sensor on the front side of the roller Q Perform time domain integration to obtain the vibration steel wheel amplitude i of the second roller x2 , obtained by laser displacement sensor x2 And the acceleration signal time domain integration i x1 Co-verification of the vibration amplitude of the roller drum: x =(i x1 +i x2 ) / 2, at this time i x It is the amplitude of the vibrating steel wheel of the roller;

[0146] Calculation of compaction degree of roadbed filler: In the field test, the total volume of the paved filler is calculated based on the compaction degree of each 100m 3 Take 30kg of filler for screening, measurement and weighing test, determine and record the parameters of the roadbed filler for the field test. The parameters of the roadbed filler for the field test include soil type u x , soil particle gradation v x , rock content w x , block stone gradation x x , stone aspect ratio y x (Take 10% of the total amount of filler blocks, measure the ratio of the long side and short side of each block and find the average value), moisture content ω x ;

[0147] In the field test, after each compaction of the roadbed, the strip was tested with sand filling method every 25m to obtain the maximum dry density ρ of the roadbed filler after each compaction. dmaxx and the dry density of roadbed filler is ρ dx , the compaction degree K of the roadbed filler in the field test is obtained according to the following formula: x

[0148]

[0149] Record the rated exciting force L of the roller in the field test x , excitation frequency f x , roller vibration steel wheel amplitude ix , the settlement rate of each compaction sx, the compaction degree of roadbed filler K x , while directly recording the soil pressure c of the bottom center soil pressure sensor in the field test x11 , Earth pressure of bottom edge earth pressure sensor c x12 、Middle center soil pressure sensor soil pressure c x21 , Earth pressure sensor at the middle edge Earth pressure c x22 、Top center earth pressure sensor earth pressure c x31 , top edge earth pressure sensor earth pressure c x32 , temperature sensor temperature g x 、Humidity sensor humidity e x ; All the data recorded above are the data sets of the field test.

[0150] Step 3: Establish a roadbed compaction quality prediction model based on collaborative training of virtual and real data.

[0151] The specific implementation process of step three is the construction of the roadbed compaction quality prediction model. Taking the data set of the indoor test as a sample, a three-layer artificial neural network compaction quality prediction model including input layer, hidden layer and output layer is established. The input layer includes 12 characteristic values, including filler soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, water content, exciting force, exciting frequency, amplitude, harmonic ratio CMV, soil pressure at different positions, and settlement rate; the number of neurons in the hidden layer is 9; the number of output neurons is 1, and the compaction degree is the final target value;

[0152] The training of the roadbed compaction quality prediction model divides the data set consisting of a large amount of indoor test data into 75% training set and 25% test set, and the BP-LM algorithm is used for weights and thresholds;

[0153] Optimization of the roadbed compaction quality model: input the data set of the field test into the trained roadbed compaction quality prediction model, train and optimize the model, and obtain the roadbed compaction quality prediction model based on the collaborative training of virtual and real data.

[0154] Step 4: Construction control application of roadbed compaction quality prediction model based on collaborative prediction of virtual and real indicators.

[0155] The specific implementation process of step 4 (such as Figure 2 ), before the on-site roadbed compaction application, set up 1 to 2 test strips to calibrate the roadbed compaction quality prediction model based on the collaborative training of virtual and real data and collect soil pressure data.

[0156] According to step 2, the on-site roadbed compaction test of the test strip is carried out, and the data collected by real-time monitoring, the data calculated by the data analysis and processing module and the test data are: soil type u, soil particle gradation v, block stone content w, block stone gradation x, block stone aspect ratio y (take 10% of the total amount of filler blocks, measure the length ratio of the long side and short side of each block stone and calculate the average value), moisture content ω, soil pressure c of the soil pressure sensor at the center of the filler bottom 11 , Earth pressure of bottom edge earth pressure sensor c 12 、Middle center soil pressure sensor soil pressure c 21 , Earth pressure sensor at the middle edge Earth pressure c 22 、Top center earth pressure sensor earth pressure c 31 , top edge earth pressure sensor earth pressure c 32 , the temperature sensor temperature g, the humidity sensor humidity e, the excitation frequency f, the amplitude i, and the settlement rate of each compaction s are input into the roadbed compaction quality prediction model P based on virtual and real data collaborative training obtained in step 3 i In this paper, the roadbed compaction quality prediction model P based on virtual and real data collaborative training is proposed. i The real-time predicted compaction degree K of the compacted strip can be obtained YC After each rolling of the test strip, the compaction degree K is obtained by the sand filling method. SY , compare the two and find the error R: When the error R is greater than 5%, the next rolling test is continued, and the compaction degree K obtained by this sand filling test is calculated. SY The corresponding model input data is used to predict the roadbed compaction quality based on the collaborative training of virtual and real data. i The updated training optimization is performed again to obtain the updated roadbed compaction quality prediction model P based on virtual and real data collaborative training. i+1 , until the error R is less than or equal to 5%, the final roadbed compaction quality prediction model is obtained, which is the optimized and adjusted roadbed compaction quality prediction model P based on virtual and real data collaborative training L , the adjustment is completed to optimize the roadbed compaction quality prediction model P based on virtual and real data collaborative training. L The predicted compaction degree obtained is used as the real-time compaction degree of roadbed compaction.

[0157] After the roadbed compaction quality prediction model based on virtual and real data collaborative training is optimized and adjusted, the real-time monitored on-site soil type u, soil particle gradation v, block stone content w, block stone gradation x, block stone aspect ratio y (taking 10% of the total amount of filler blocks, measuring the length ratio of the long side and short side of each block stone to find the average), moisture content ω, and soil pressure c of the soil pressure sensor at the bottom of the filler are used. 11 , Earth pressure of bottom edge earth pressure sensor c 12 、Middle center soil pressure sensor soil pressure c21 , Earth pressure sensor at the middle edge Earth pressure c 22 、Top center earth pressure sensor earth pressure c 31 , top edge earth pressure sensor earth pressure c 32 The data are input into the optimized and calibrated roadbed compaction quality prediction model based on collaborative training of virtual and real data, and the data analysis and processing module is used to obtain the excitation frequency f, amplitude i, settlement rate s of each compaction, and real-time collected temperature sensor temperature g and humidity sensor humidity e. The data are input into the optimized and calibrated roadbed compaction quality prediction model based on collaborative training of virtual and real data. The on-site roadbed compaction degree can be obtained in real time, and the on-site roadbed compaction quality can be monitored and evaluated in real time. The roller is equipped with a GPS positioning module, and the real-time speed of the roller can be obtained at the same time. The precise positioning of any point in the strip can be achieved, and the points with insufficient compaction can be accurately supplemented to avoid over-compaction.

[0158] The present invention realizes the reproduction of on-site roadbed compaction conditions in the laboratory through the on-site-laboratory roadbed compaction intelligent collaborative perception program, carries out the collection of virtual and real indicators of on-site roadbed compaction quality, and establishes an accurate, comprehensive and reliable indoor and outdoor virtual and real indicator compaction data set. Through the on-site-laboratory roadbed compaction intelligent collaborative perception program, in the laboratory, according to the roadbed filling parameters, roller working parameters, and external environmental parameters in the on-site roadbed compaction process, each parameter can be accurately adjusted into groups, and the variables can be controlled for testing. Real-time monitoring and collection of target data establishes indoor and outdoor test data sets of the same standard.

[0159] The system of the present invention can monitor, collect and process data of monitoring indicators (such as soil pressure at different positions, acceleration, harmonic ratio CMV, and settlement) in real time during the filling compaction process in laboratory tests; and monitor, collect and process data of monitoring indicators (such as position, direction of travel, speed, soil pressure at different positions, acceleration, harmonic ratio CMV, temperature, humidity, and settlement) in real time during field tests and engineering applications. A roadbed compaction quality prediction model is established based on the virtual and real indicators monitored in real time. In laboratory tests and field tests, the real-time monitoring or processed and analyzed data includes both physical and mechanical "real" indicators such as settlement and soil pressure at different positions that can intuitively reflect the compaction situation, and "virtual" indicators such as harmonic ratio CMV, amplitude, and excitation frequency that can indirectly feedback the compaction situation. Based on laboratory tests and field tests, the factors affecting the roadbed compaction quality - roadbed filling parameters, roller working parameters and external environmental parameters are coordinated, and the "virtual" and "real" indicators are coordinated to establish a roadbed compaction quality neural network prediction model.

[0160] The present invention verifies and optimizes the compaction quality prediction model in on-site roadbed compaction applications, improves the accuracy of the prediction model, and thus makes an accurate evaluation of the roadbed compaction quality monitoring results and proposes targeted management and maintenance methods.

[0161] Any matters not described in the present invention are applicable to the prior art.

Claims

1. A roadbed compaction quality monitoring and maintenance method based on virtual and real index collaboration, characterized in that: The method comprises the following steps: Step 1: Establish an indoor compaction test that reproduces the on-site roadbed compaction conditions and establish a data set for the indoor test; Step 2: Conduct field test on roadbed compaction and establish a data set for the field test; Step 3: Establish a roadbed compaction quality prediction model based on virtual and real data collaborative training: A three-layer artificial neural network compaction quality prediction model including input layer, hidden layer and output layer is constructed; the input layer includes 12 characteristic values ​​including filler soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, water content, exciting force, exciting frequency, amplitude, harmonic ratio CMV, soil pressure at different positions and settlement rate; the output layer is compaction degree; The data set of indoor tests is used as samples to train the compaction quality prediction model of artificial neural network; The data set of the field test is input into the trained artificial neural network compaction quality prediction model, and the trained artificial neural network compaction quality prediction model is optimized to obtain a roadbed compaction quality prediction model based on the collaborative training of virtual and real data; Step 4: Construction control application of roadbed compaction quality prediction model based on collaborative prediction of virtual and real indicators: Before the on-site roadbed compaction application, set up 1 to 2 test strips for on-site roadbed compaction testing, and use the real-time data of the test strips to calibrate the roadbed compaction quality prediction model based on the collaborative training of virtual and real data to obtain the optimized and calibrated roadbed compaction quality prediction model based on the collaborative training of virtual and real data. The optimized and calibrated roadbed compaction quality prediction model based on the collaborative training of virtual and real data can be used for on-site roadbed compaction implementation.

2. The roadbed compaction quality monitoring and maintenance method based on virtual and real index coordination according to claim 1 is characterized in that: The specific implementation process of establishing an indoor compaction test that reproduces the on-site roadbed compaction conditions in the step 1 is: the indoor test uses a surface vibration compactor to complete the laboratory surface vibration compaction test, two laser displacement sensors are installed above the tamping plate of the surface vibration compactor, two vertical acceleration sensors are installed at the center and edge of the tamping plate of the surface vibration compactor, a harmonic ratio CMV sensor is installed on the tamping plate of the surface vibration compactor, and a temperature sensor and a humidity sensor are used to monitor external environmental parameters; the paving thickness is set, and six soil pressure sensors are installed at the top, middle and bottom positions in the paving area of ​​the test tube of the surface vibration compactor. The six soil pressure sensors are arranged in three layers, two on each layer, one soil pressure sensor is located at the center of the same layer, and the other soil pressure sensor is located at the edge, and all soil pressure sensors located at the edge are at the same horizontal position; The laboratory surface vibration compaction test conditions are outputted by using the indoor test condition adjustment subroutine, including roadbed filling parameters, roller working parameters and external environmental parameters. The roadbed filling parameters include soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio and water content; the roller working parameters include exciting force and exciting frequency; the external environmental parameters include temperature and humidity; According to the energy conversion relationship between the roller and the surface vibration compactor, the working parameters of the surface vibration compactor are adjusted. The tamping plate of the surface vibration compactor is adjusted by adding or subtracting the counterweight to adjust the frequency converter of the exciting motor to change the exciting force and exciting frequency of the surface vibration compactor. Indoor tests are carried out according to the laboratory surface vibration compaction test conditions, and the exciting force L of the surface vibration compactor is set. s ; The acceleration signal during loading is calculated based on the acceleration of the two vertical acceleration sensors, and the acceleration signal during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f S , the acceleration signal is integrated in the time domain during the loading process to obtain the first tamping plate amplitude i S1 ; Calculate the real-time sedimentation rate s of the filler based on the real-time values ​​of the laser displacement sensors on both sides and the filling height of the filler S , where the filling height of the filler is a set known value; the electrical signal collected by the laser displacement sensor is integrated in the time domain to obtain the second tamping plate amplitude i S2 , according to formula i S =(i S1 +i S2 ) / 2 to obtain the tamping plate amplitude i S ; According to the surface vibration compactor test method in the "Highway Geotechnical Test Code", the maximum dry density of the filler ρ is obtained. dmaxs and the dry density of the test filler ρ ds , the compaction degree K of the roadbed filler in the indoor test is obtained according to the following formula: S ; Record the exciting force L of the surface vibration compactor in the indoor test s , excitation frequency f S , tamping plate amplitude i S , sedimentation rate s S , roadbed filler compaction degree K S ; At the same time, directly record the soil pressure c of the bottom center soil pressure sensor S11 , Earth pressure of bottom edge earth pressure sensor c S12 、Middle center soil pressure sensor soil pressure c S21 , Earth pressure sensor at the middle edge Earth pressure c S22 、Top center soil pressure sensor soil pressure c S31 , top edge earth pressure sensor earth pressure c S32 , temperature sensor temperature g S 、Humidity sensor humidity S ; and record the roadbed filling parameters of the indoor test, the roadbed filling parameters of the indoor test include soil type u S , soil particle gradation v S , rock content w S , block stone grading x S , stone aspect ratio y S , moisture contentω S ; The large amount of data recorded above constitutes the data set for indoor experiments.

3. The roadbed compaction quality monitoring and maintenance method based on virtual and real index coordination according to claim 1 is characterized in that: The specific implementation process of the field test of roadbed pressure in step 2 is: The field test uses a roller. n laser displacement sensors are respectively set at equal heights on the front and rear sides of the roller through mounting frames, where n is a natural number. A vertical acceleration sensor and a harmonic wave ratio CMV sensor are installed on the central axis of the roller's vibrating steel wheel. The installation directions of the vertical acceleration sensor and the harmonic wave ratio CMV sensor are consistent with the vibration direction of the roller's steel wheel. A temperature sensor and a humidity sensor for measuring ground temperature and humidity are installed on the roller. During the filling paving process, six soil pressure sensors are pre-buried at the top, middle, and bottom positions of the test strip cross section every 25m along the compaction direction. The six soil pressure sensors are arranged in three layers, with two on each layer. In the same layer, one soil pressure sensor is located in the center, and the other soil pressure sensor is located at the edge. All soil pressure sensors located at the edge are at the same horizontal position. The arrangement of earth pressure sensors in the field test is consistent with that in the indoor test; Install GPS on the roof of the roller.

4. The roadbed compaction quality monitoring and maintenance method based on virtual and real index coordination according to claim 1 is characterized in that: At the beginning of the field test, all laser displacement sensors at the front and rear of the roller were adjusted to zero. During the roller's forward movement, the real-time value of the laser displacement sensor at the front from left to right was recorded as h. 11 、h 12 、h 13 ,…,h 1n The real-time value of the laser displacement sensor from left to right on the rear side is recorded as h 21 、h 22 、h 23 ,…,h 2n ; During the roller's backward movement, the real-time values ​​of the laser displacement sensors from left to right on the front side are recorded as h1'1, h1'2, h1'3, ..., h1' n , the real-time value of the laser displacement sensor from left to right on the rear side is recorded as h' 21 、h' 22 、h' 23 ,…,h' 2n ; The acceleration of the vertical acceleration sensor on the roller is recorded as a x ; Get the acceleration signal a during the roller loading process x , the acceleration signal of the roller during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f x , the acceleration signal of the roller during loading is integrated in the time domain to obtain the vibration steel wheel amplitude i of the first roller x1 ; The position and movement direction of the roller are obtained by GPS positioning, and the settlement amount h of the roller during the forward movement is QJ for: The amount of settlement of the roller during the backward movement h HT for: Wherein, n is the number of single-side laser displacement sensors, and n is a natural number; The settlement rate of each compaction is s x for: H x is the paving height of the roadbed fill, which is the set value of the field test; When the roller moves forward Calculate the electrical signal collected by the laser displacement sensor on the front side of the roller. Calculate the electrical signal collected by the laser displacement sensor on the front side of the roller; The electrical signal h collected by the laser displacement sensor on the front side of the roller Q Perform time domain integration to obtain the vibration steel wheel amplitude i of the second roller x2 According to formula i x =(i x1 +i x2 ) / 2 to get the vibration amplitude i of the roller steel wheel x ; In the field test, the total volume of the paved filler is calculated based on the total volume of the paved filler per 100m 3 Take 30kg of filler for screening, measurement and weighing test, determine and record the parameters of the roadbed filler for the field test, the parameters of the roadbed filler for the field test include: soil type u x , soil particle gradation v x , rock content w x , block stone grading x x , stone aspect ratio y x , moisture contentω x ; In the field test, after each compaction of the roadbed, the strip was tested with sand filling method every 25m to obtain the maximum dry density ρ of the roadbed filler after each compaction. dmaxx and roadbed filler dry density ρ dx , the compaction degree K of the roadbed filler in the field test is obtained according to the following formula: x Record the rated exciting force L of the roller in the field test x , excitation frequency f x , roller vibration steel wheel amplitude i x , the settlement rate of each compaction s x , roadbed filler compaction degree K x , while directly recording the soil pressure c of the bottom center soil pressure sensor in the field test x11 , Earth pressure of bottom edge earth pressure sensor c x12 、Middle center soil pressure sensor soil pressure c x21 , Earth pressure sensor at the middle edge Earth pressure c x22 、Top center soil pressure sensor soil pressure c x31 , top edge earth pressure sensor earth pressure c x32 , temperature sensor temperature g x 、Humidity sensor humidity x ; The large amount of data recorded above constitutes the data set for the field experiment.

5. The roadbed compaction quality monitoring and maintenance method based on virtual and real index coordination according to claim 3 is characterized in that: The process of adjusting the roadbed compaction quality prediction model based on virtual and real data collaborative training in step 4 is: Set up 1 to 2 test strips to conduct field tests on the roadbed compaction, obtain the roadbed filling parameters of the test strips, and the soil pressure c of the bottom center soil pressure sensor 11 , Earth pressure of bottom edge earth pressure sensor c 12 、Middle center soil pressure sensor soil pressure c 21 , Earth pressure sensor at the middle edge Earth pressure c 22 、Top center soil pressure sensor soil pressure c 31 , top edge earth pressure sensor earth pressure c 32 , the temperature sensor temperature g, the humidity sensor humidity e, the excitation frequency f, the vibration amplitude of the roller steel wheel i, and the settlement rate of each compaction s are input into the roadbed compaction quality prediction model based on virtual and real data collaborative training obtained in step 3. The roadbed compaction quality prediction model based on virtual and real data collaborative training is recorded as P i , after P i After processing, the real-time predicted compaction degree K of the compacted strip is obtained. YC After each rolling of the test strip, the compaction degree K is obtained by the sand filling method test. SY , according to the formula Calculate the error R. When the error R is greater than 5%, continue the sand filling test after the next rolling, and use the compaction degree K obtained by this sand filling test as the value of the sand filling test result. SY and the corresponding model input data are used to i The updated training optimization is performed again to obtain the updated roadbed compaction quality prediction model P based on virtual and real data collaborative training. i+1 , until the error R is less than or equal to 5%, the optimized and calibrated roadbed compaction quality prediction model P based on virtual and real data collaborative training is obtained. L , the adjustment is completed.

6. A roadbed compaction quality monitoring and maintenance system based on virtual and real index collaboration, characterized in that: It includes real-time data monitoring and acquisition module, data analysis and processing module and roadbed compaction quality evaluation and maintenance module; The data real-time monitoring and acquisition module is used to monitor and collect data of external environmental parameters and monitoring indicators in real time during laboratory and roadbed pressure field tests and applications, and transmit the above data to the data analysis and processing module; the external environmental parameters include temperature and humidity, and the monitoring indicators include vertical acceleration, harmonic ratio CMV, settlement, and earth pressure; The data analysis and processing module is used to analyze and process the data collected by the real-time data monitoring and collection module according to conditions to obtain the virtual and real evaluation index of the roadbed compaction quality, and the virtual and real evaluation index of the roadbed compaction quality includes settlement rate, excitation frequency and amplitude; The roadbed compaction quality evaluation and maintenance module is used to input the data obtained by the real-time data monitoring and acquisition module and the data analysis and processing module into the roadbed compaction quality prediction model, obtain the roadbed compaction degree prediction value in real time, and display it visually. The compaction quality is evaluated according to the roadbed compaction degree prediction value, and the under-compacted points are supplemented with compaction. The next step of roadbed compaction work is guided according to the visualized roadbed compaction degree prediction value.

7. The roadbed compaction quality monitoring and maintenance system based on virtual and real index coordination according to claim 6 is characterized in that: The data real-time monitoring and acquisition module includes indoor test real-time monitoring and acquisition equipment and field test real-time monitoring and acquisition equipment. The indoor test real-time monitoring and acquisition equipment includes a surface vibration compactor, an indoor test monitoring data acquisition terminal, two laser displacement sensors capable of irradiating laser on the tamping plate of the surface vibration compactor, two vertical acceleration sensors installed at the center and edge of the tamping plate of the surface vibration compactor, a harmonic ratio CMV sensor installed on the tamping plate of the surface vibration compactor, and a temperature sensor and a humidity sensor for monitoring external environmental parameters. The real-time monitoring and acquisition equipment for the on-site test includes a roller and an on-site monitoring data acquisition terminal. n laser displacement sensors are respectively installed at equal heights on the front and rear sides of the roller through mounting frames. A vertical acceleration sensor and a harmonic wave ratio CMV sensor are installed on the central axis of the roller's vibrating steel wheel. A temperature sensor and a humidity sensor for measuring ground temperature and humidity are installed on the roller. The indoor test monitoring data collection terminal and the field monitoring data collection terminal are connected to the base station via wireless, and the base station performs different data processing according to the data source, the indoor test monitoring data collection terminal or the field monitoring data collection terminal.

8. The roadbed compaction quality monitoring and maintenance system based on virtual and real index coordination according to claim 6 is characterized in that: Set the paving thickness, and install six soil pressure sensors at the top, middle, and bottom positions in the paving area of ​​the test tube of the surface vibration compactor. The six soil pressure sensors are arranged in three layers, two in each layer. In the same layer, one soil pressure sensor is located in the center, and the other soil pressure sensor is located at the edge. All soil pressure sensors located at the edge are at the same horizontal position; The indoor test monitoring data acquisition terminal is electrically connected to all soil pressure sensors, temperature sensors and humidity sensors, vertical acceleration sensors, laser displacement sensors, and harmonic ratio CMV sensors in the indoor test real-time monitoring and acquisition equipment; During the filling paving process, six soil pressure sensors are embedded in the top, middle and bottom positions of the test strip cross section along the compaction direction. The arrangement of the six soil pressure sensors embedded in each cross section of the field test is consistent with the arrangement of the soil pressure sensors in the indoor test. A field monitoring data acquisition terminal is installed in the roller, and the field monitoring data acquisition terminal is electrically connected to all soil pressure sensors, temperature sensors and humidity sensors, vertical acceleration sensors, laser displacement sensors, and harmonic ratio CMV sensors in the field test real-time monitoring and acquisition equipment; The roadbed compaction quality monitoring and maintenance system based on virtual and real indicator collaboration also includes a mobile terminal, which is connected to the website through the network to view data collection status and roadbed compaction quality status in real time; The base station is used to complete data analysis and processing and roadbed compaction quality evaluation, and upload the data to the website.

9. The roadbed compaction quality monitoring and maintenance system based on virtual and real index coordination according to claim 6 is characterized in that: The functions of the data analysis and processing module during the indoor test are: Acceleration signal analysis and processing: The acceleration signal during loading is calculated based on the acceleration of the two vertical acceleration sensors, and the acceleration signal during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f S , the acceleration signal is integrated in the time domain during the loading process to obtain the first tamping plate amplitude i S1 ; Laser displacement sensor analysis and processing: Calculate the real-time sedimentation rate s of the filler based on the real-time values ​​of the laser displacement sensors on both sides and the filling height of the filler S , where the filling height of the filler is a set known value; the electrical signal collected by the laser displacement sensor is integrated in the time domain to obtain the tamping plate amplitude i S2 , according to formula i S =(i S1 +i S2 ) / 2 to obtain the tamping plate amplitude i S ; Calculation of compaction degree of roadbed filler: According to the surface vibration compactor test method in the "Highway Geotechnical Test Code", the maximum dry density of the filler ρ is obtained. dmaxs and the dry density of the test filler is ρ ds , the compaction degree K of the roadbed filler in the indoor test is obtained according to the following formula: S ; Record the exciting force L of the surface vibration compactor in the indoor test s , excitation frequency f S , tamping plate amplitude i S , sedimentation rate s S , roadbed filler compaction degree K S ; At the same time, directly record the soil pressure c of the bottom center soil pressure sensor S11 、Earth pressure of the earth pressure sensor at the right edge of the bottom c S12 、Middle center soil pressure sensor soil pressure c S21 , Earth pressure sensor at the middle edge Earth pressure c S22 、Top center soil pressure sensor soil pressure c S31 , top edge earth pressure sensor earth pressure c S32 , temperature sensor temperature g S 、Humidity sensor humidity S ; and record the roadbed filling parameters of the indoor test, the roadbed filling parameters include soil type u S , soil particle gradation v S , rock content w S , block stone grading x S , stone aspect ratio y S , moisture contentω S ; The functions of the data analysis and processing module during field testing are: Acceleration signal analysis and processing: Obtain the acceleration signal a during the roller loading process x , the acceleration signal of the roller during loading is filtered, Fourier transformed and spectrum analyzed to obtain the excitation frequency f x , the acceleration signal of the roller during loading is integrated in the time domain to obtain the vibration steel wheel amplitude i of the first roller x1 ; Laser displacement sensor analysis and processing: GPS positioning is used to obtain the roller's position and direction of movement, and the roller's settlement amount h during its forward movement. QJ : The amount of settlement of the roller during the backward movement h HT : Where n is the number of single-side laser displacement sensors; Sedimentation rate per compaction pass: H x is the paving height of the roadbed fill; The electrical signal h collected by the laser displacement sensor on the front side of the roller is obtained according to the following formula: Q or The electrical signal h collected by the laser displacement sensor on the front side of the roller Q Perform time domain integration to obtain the vibration steel wheel amplitude i of the second roller x2 According to formula i x =(i x1 +i x2 ) / 2 to get the vibration amplitude i of the roller steel wheel x ; Calculation of compaction degree of roadbed filler: In the field test, the total volume of the paved filler is calculated based on the compaction degree of each 100m 3 Take 30kg of filler for screening, measurement and weighing test to determine and record the parameters of the roadbed filler, including soil type u x , soil particle gradation v x , rock content w x , block stone grading x x , stone aspect ratio y x , moisture contentω x ; In the field test, after each compaction of the roadbed, the strip was tested with sand filling method every 25m to obtain the maximum dry density ρ of the roadbed filler after each compaction. dmaxx and roadbed filler dry density ρ dx , the compaction degree K of the roadbed filler in the field test is obtained according to the following formula: x ; Record the rated exciting force L of the roller tested on site x , excitation frequency f x , roller vibration steel wheel amplitude i x , the settlement rate of each compaction s x , roadbed filler compaction degree K x , while directly recording the soil pressure c of the bottom center soil pressure sensor in the field test x11 , Earth pressure of bottom edge earth pressure sensor c x12 、Middle center soil pressure sensor soil pressure c x21 , Earth pressure sensor at the middle edge Earth pressure c x22 、Top center soil pressure sensor soil pressure c x31 , top edge earth pressure sensor earth pressure c x32 , temperature sensor temperature g x 、Humidity sensor humidity x .

10. The roadbed compaction quality monitoring and maintenance system based on virtual and real index coordination according to claim 6 is characterized in that: The roadbed compaction quality prediction model is composed of a neural network with 12 characteristic values ​​as input, including filler soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, water content, exciting force, exciting frequency, amplitude, harmonic ratio CMV, soil pressure at different positions, and settlement rate, and compaction degree as output; The base station is loaded with a field-laboratory roadbed compaction intelligent collaborative perception program, and the field-laboratory roadbed compaction intelligent collaborative perception program includes an indoor test condition deployment subprogram; The indoor test condition allocation subroutine is used to randomly allocate the roadbed filling parameters, roller working parameters, and external environmental parameters in the indoor test to generate a large number of test condition combinations; the roadbed filling parameters include soil type, soil particle gradation, block stone content, block stone gradation, block stone aspect ratio, and moisture content; the external environmental parameters include temperature and humidity; the roller working parameters include exciting force and exciting frequency; the indoor test condition allocation subroutine has a collection of various roadbed filling parameters, various roller working parameters, and various external environmental parameters; Obtaining the energy conversion relationship between the roller and the surface vibration compactor, and obtaining the working parameters of the surface vibration compactor according to the energy conversion relationship between the roller and the surface vibration compactor, so that the working parameters of the surface vibration compactor in the indoor test are consistent with the working parameters of the roller on site; The external environmental parameters of the indoor test are set according to the external environmental parameters of the on-site roadbed compaction conditions. Reproduce the roadbed filling parameters indoors according to the actual situation of the filling on site. The external environmental parameters of the on-site roadbed compaction conditions, roadbed filling parameters, and roller working parameters are obtained by inputting the indoor test condition adjustment subroutine. The indoor test condition adjustment subroutine outputs multiple groups of all parameters required for indoor tests under the current on-site roadbed compaction test conditions, thereby realizing intelligent collaborative perception of on-site-laboratory roadbed compaction.

Images