A method and system for acquiring dynamic mechanical response data of road structure

Abstract

The present invention relates to a method for acquiring mechanical response data of road structures under the action of vehicle dynamic loads, and belongs to the field of long-term performance observation of highways in the field. The present invention buries a mechanical sensor matrix inside the road to continuously and in real time acquire the mechanical data of the internal structure of the road at a high sampling frequency of more than 2000 Hz. The collected data is processed synchronously by a high-performance front-end data processing unit to realize a series of synchronous processing such as dynamic response data detection, low information density data compression, and high and low frequency data separation. On this basis, after the metadata information of each independent data unit is extracted and processed, it is encapsulated into a standardized data packet with a unique number, and finally a highway scientific data resource with high availability and credibility is formed.

Description

Technical Field

[0001] The present invention relates to the technical field of highway management, and in particular to a method for acquiring road structure mechanical response data of a field highway under a vehicle dynamic load, and the formation, management and application of subsequent scientific data resources. Background Art

[0002] Today, the world's highway technology is in a period of transformation and development. To this end, all countries are further strengthening the accumulation of long-term highway performance observation data, aiming to establish a complete and detailed highway science data system. On this basis, they systematically summarize and study the evolution law and decay mechanism of long-term highway performance, and develop a new generation of highway construction and maintenance technology system.

[0003] Among them, the road structure mechanical response characteristics, as important characteristic data for evaluating road quality and estimating road life, have been widely used in research fields such as road mechanical models, back calculation methods, and principle derivation; however, due to its own problems such as large data volume, low processing efficiency, data transmission link limitations, and non-standard data, it is difficult to systematically and scientifically obtain stable observation data in actual long-term highway performance observation. As a result, the research results such as road structure models, new processes, and new methods obtained are difficult to be widely used in actual work, and lack systematic scientific data for effectiveness verification. Summary of the invention

[0004] The purpose of the present invention is to provide a method for effectively acquiring dynamic mechanical response data of road structures in real time, so as to solve the problem of scientific aggregation of massive mechanical response data in long-term field performance observation of highways. The present invention buries a mechanical sensor matrix inside the road and uses a high-frequency data acquisition unit to perform real-time and uninterrupted data acquisition to obtain complete mechanical response data of road structures when vehicles pass by; the collected massive response data will be saved in a high-speed cache, and then processed in real time by a high-performance front-end data processing unit, invalid data will be filtered to increase the density of effective data and reduce the total amount of data transmission; the processed and refined data will be packaged and encapsulated with a single response as the minimum unit, and standardized processing such as data numbering, timestamp calibration, and metadata information addition will be performed on it, finally forming a highway scientific data resource with high availability and credibility.

[0005] A method for acquiring dynamic mechanical response data of a road structure comprises the following steps:

[0006] 1. High-frequency real-time collection of road structure mechanics data

[0007] The mechanical sensors are buried in each structural layer of the road, and their signal data lines are connected to the high-frequency data acquisition unit after passing through the acquisition channel protection module. The working parameters of the data acquisition unit are set, and the acquisition frequency of the data acquisition unit is set to 2000Hz or higher. The working mode of the acquisition unit is set to 24-hour uninterrupted acquisition;

[0008] 2. Filter low value density data

[0009] The data stream volatility is calculated through the dynamic load response detection algorithm. When the volatility is greater than the set threshold, the current moment is recorded as the start time of the mechanical response of the road structure, and the data stream status is updated to "dynamic response". When the data stream volatility is lower than the set threshold, the current moment is recorded as the end time of the response, and the data stream status is updated to "no dynamic response". The non-response data is compressed according to the low-frequency sampling frequency based on the recorded time.

[0010] 3. Processing of Mechanical Response Data

[0011] Processing the processed data stream, including: data timestamp calibration; separation of high-frequency and low-frequency data, forming independent data units for each response; extraction and processing of response data metadata information; and extraction and processing of non-response data metadata information;

[0012] 4. Data Resource Packaging

[0013] Each independent data unit and its metadata information are uniformly packaged and numbered according to standards to form scientific data resources.

[0014] The data flow volatility P is calculated as follows:

[0015]

[0016] Where P is the volatility of the current data stream, x i is the real-time data in the dynamic window, and B is the average baseline value of the data stream.

[0017] The high frequency is 2000 Hz, the low frequency is 1 Hz, and the method for compressing the non-response data according to the low frequency sampling frequency is: the data stream from 0.5s after the end of each response to 0.5s before the start of the next response is averaged at a sampling rate of 1 Hz. The formula for the 1 Hz sampling rate average processing is: where x j is the sampling value after averaging for 1 second, F is the sampling rate of the high-frequency acquisition unit, x i The sensor response data.

[0018] The response data metadata information extraction and processing includes the following data information: data start time, data end time, number of data sampling points, data dimension, data accuracy, data baseline, overall mean of response data, upper 5% data mean and lower 5% data mean; the non-response data metadata information extraction and processing includes the following data information: data file size, number of data sampling points, data dimension and data accuracy.

[0019] The data dimension is the number of sensors plus the data timestamp, and the calculation formula is: M=m+1, where m is the number of sensors;

[0020] The data baseline calculation formula is: Where B is the data baseline, N is 5% of the total single response data, and x i Sensor response data;

[0021] The overall mean calculation formula of the response data is: in is the overall mean of the response data, M is the total amount of single response data, x i Sensor response data;

[0022] The calculation formula for the mean of the upper 5% data is: in is the mean of the top 5% data, N is 5% of the total single response data, x′ i Arrange the top 5% of the sensor response data in descending order;

[0023] The formula for calculating the mean of the lower 5% data is: in is the mean of the next 5% of data, N is 5% of the total single response data, x″ i The sensor response data is sorted in descending order and contains the last 5% of the data.

[0024] The number is for the response data resource and the non-response data resource number, and includes the following information: collection observation station, collection date, hour time, response sequence number within this hour, and start data timestamp, etc., and the code is unique.

[0025] The method also includes establishing an archive database of mechanical sensors, and the database storage fields include: number, sensor type, model, manufacturer, burial location, burial depth, measurement direction, correction model, model parameters, zero position reading, and zero load reading.

[0026] The acquisition system of the road structure dynamic mechanical response data designed by the above acquisition method includes a data acquisition subsystem, a data processing and temporary storage subsystem, and a data processing and packaging subsystem;

[0027] The data acquisition subsystem includes mechanical sensors located at each structural layer of the road, an acquisition channel protection module and a high-frequency data acquisition unit. The signal data line of the mechanical sensor is connected to the high-frequency data acquisition unit through the acquisition channel protection module; the high-frequency data acquisition unit acquires dynamic mechanical response data in the mechanical sensor;

[0028] The data processing and temporary storage subsystem includes a high-performance front-end data processing unit A, a first-level cache unit and a second-level cache unit. The response data collected by the high-frequency data collection unit is stored in the first-level cache unit in the form of a data stream; the high-performance front-end data processing unit A reads the response data from the first-level cache unit, calculates the data stream volatility in real time, and updates the data stream state at the same time; when a vehicle passes by, when the calculated volatility is greater than a set threshold, the current moment is recorded as the start time of the mechanical response of the road structure, and the data stream state is updated to "dynamic response"; when the vehicle passes, the data stream volatility is less than the set threshold, the current moment is recorded as the response end time, and the data stream state is updated to "no dynamic response"; the data from the response end time to the next response start time is compressed as a whole according to the low-frequency sampling rate, and the dynamic response and no dynamic response data streams are temporarily stored in the second-level cache unit;

[0029] The data processing and archiving subsystem includes a high-performance front-end data processing unit B, which reads data from the secondary cache unit and calibrates the timestamp information of the data according to the local standard time; separates the response data and the non-response data according to the recorded response time, and forms an independent file with a single response as the unit of high-frequency response data and an independent file with an hour as the unit of low-frequency non-response data; extracts the basic information in the response data and the non-response data: the start time, the end time, the number of data sampling points, the data dimension and the data accuracy; calculates the data information in the response data: the data baseline, the overall mean of the response data, the upper 5% data mean and the lower 5% data mean; uniformly numbers the response data and the non-response data, and encapsulates the entity data, basic information and feature information into an independent data resource package.

[0030] The mechanical sensors are arranged with the center line of the wheel track on the left side of the road as the reference line. The mechanical sensors are buried layer by layer to the bottom of each structural layer of the road. The strain sensors are buried in the asphalt layer and the water-stable layer in a 2x2 matrix to measure the lateral and longitudinal strains of the bottom of the structural layer when the vehicle passes. The stress sensors are buried in all structural layers at the center of the strain sensor matrix, including the asphalt layer, the water-stable layer and each layer of the roadbed, to measure the vertical stress of the bottom of the structural layer when the vehicle passes. The strain sensors include a pair of lateral strain sensors and vertical strain sensors, which are arranged in a T-shape in the matrix.

[0031] The acquisition channel protection module includes a surge pulse protector, an electromagnetic interference filter, a low-frequency filter and a signal isolator.

[0032] The specific steps of the method of the present invention are:

[0033] 1. Create a sensor profile

[0034] (1) Sensor numbering: uniformly number the mechanical sensors to be buried, including strain sensors and stress sensors, to ensure the uniqueness of the numbers and the traceability of subsequent data.

[0035] (2) Sensor calibration: the strain sensor and stress sensor are calibrated before burial.

[0036] (3) Sensor parameter recording, recording the calibration model, parameters and sensor readings in the initial zero state.

[0037] 2. High-frequency real-time collection of road structural mechanics data

[0038] (1) Sensor placement: Taking the center line of the wheel track on the left side of the road as the reference line, mechanical sensors are buried layer by layer at the bottom of each structural layer of the road. Strain sensors are buried in the asphalt layer and the water-stable layer in a 2x2 matrix to measure the lateral and longitudinal strains at the bottom of the structural layer when a vehicle passes by. Stress sensors are buried in all structural layers at the center of the strain sensor matrix, including the asphalt layer, the water-stable layer, and each layer of the roadbed, to measure the vertical stress at the bottom of the structural layer when a vehicle passes by.

[0039] (2) Install the acquisition channel protection module to isolate the interference between the acquisition channels, eliminate the disturbance of the power supply to the acquisition signal, prevent the damage of the acquisition equipment to the surge pulse, and suppress external electromagnetic interference.

[0040] (3) Set the working parameters of the data acquisition unit, set the data acquisition unit acquisition frequency to 2000 Hz or higher, and set the acquisition unit working mode to 24-hour uninterrupted acquisition.

[0041] (4) Record the zero-load sensor response and the initial response data when there is no vehicle load after the sensor is buried.

[0042] 3. Filter low value density data

[0043] (1) Temporary storage of collected data: the massive high-frequency data collected in real time is temporarily stored in the first-level high-speed memory in the form of data stream.

[0044] (2) Dynamic load response detection: by detecting the fluctuation of the data stream and recording the time when the fluctuation exceeds the threshold, it is the starting time point when the road structure produces a mechanical response when the vehicle passes through the sensor matrix; the time when the data stream fluctuation recovers is recorded as the end time point of the mechanical response.

[0045] (3) Compress the non-response data, process the data stream from 0.5s after the end of each response to 0.5s before the next response as the average value, reduce the data sampling frequency to 1Hz, and significantly reduce the amount of low information density data.

[0046] (4) Data secondary cache: pushes the compressed and filtered data stream to the secondary high-speed memory, waiting for subsequent processing.

[0047] 4. Processing of Mechanical Response Data

[0048] (1) Time calibration: calibrate the data stream in the secondary cache with the local standard time and correct the corresponding timestamp information to ensure the time accuracy of the acquired data.

[0049] (2) Data separation: extract high-frequency data from the second-level cache from 0.5 seconds before the initial time of each response to 0.5 seconds after the end time, and retain low-frequency data streams and store them independently into files every hour to form non-responsive data.

[0050] (3) Extraction of basic information of response data. The basic information that needs to be extracted includes: data start time, end time, number of data sampling points, data dimension, and data accuracy.

[0051] (4) Processing of response data characteristic information: the characteristic information that needs to be processed and calculated includes: data baseline, overall mean of response data, upper 5% data mean, and lower 5% data mean.

[0052] (5) Extraction of basic information of unresponsive data. The basic information that needs to be extracted includes: data file size, number of data sampling points, data dimension, and data accuracy.

[0053] 5. Data resource encapsulation

[0054] (1) Response data number. Each response data is uniquely numbered in the format of the unified code of the collection observation station or the unique English abbreviation + collection date (2021-07-08) + hour (24-hour format) + response sequence number within this hour + timestamp of the starting data.

[0055] (2) Response data resource encapsulation: encapsulate the basic information, characteristic information and entity data of the response data into independent data packets and name them according to their data numbers.

[0056] (3) No response data number: uniquely number the no response data in the format of the unified code of the collection observation station or the unique English abbreviation + collection date (2021-07-08) + hour (24-hour format).

[0057] (4) Encapsulation of unresponsive data resources: encapsulating the basic information and entity data of unresponsive data into independent data packets and naming them according to their data numbers.

[0058] Effects of the present invention:

[0059] (1) The total amount of data is greatly reduced by compressing non-responsive data with low information density.

[0060] Since the dynamic acquisition unit collects data at a frequency of 2000Hz, a single field observation station will generate 30 to 50GB of data every day. In addition to the effective response data when the vehicle passes, there is also a large amount of no-load response data. The present invention separates the data stream and reduces the sampling frequency of the no-response data, thereby greatly reducing the total amount of data without losing the effective information of the overall data. Based on 5,000 vehicles / day and the average effective response time of 4 seconds, the total amount of data will be reduced by about 70% if the effective data remains unchanged.

[0061] (2) Time calibration ensures that dynamic data is synchronized with local standard time.

[0062] Due to the accumulated error of the acquisition device's own clock, it will produce a large time deviation after long-term operation, which will lead to confusion in data recording and make it difficult to achieve time synchronization with other data, making it impossible to effectively associate in subsequent calculation and analysis. The present invention effectively solves the time synchronization problem of long-term data acquisition by setting time calibration and performing real-time time correction for a single response, avoiding the inability to effectively associate data in the subsequent data analysis process.

[0063] (3) High availability of collected data is ensured by independent and standardized packaging of single response data and its metadata.

[0064] Since dynamic mechanical response data has the characteristics of large data volume, wide spatial distribution, long time span, etc. when conducting actual field observations, and there is no systematic and scientific management mechanism, it is difficult to trace the data in later applications, and its spatial and temporal coordinates cannot be accurately located, making a large amount of data isolated data, greatly reducing its application value. The present invention forms a complete set of basic metadata by processing various basic information and feature information of data in real time, and independently encapsulates it in a unified standardized manner with entity data to form scientific data resources, effectively solving the problems of subsequent data submission and management difficulties, difficulty in correlation research and application, and poor reliability, and greatly improving the use value of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 Road structure mechanics sensor layout (top view),

[0066] Figure 2 Road structure mechanics sensor layout diagram (main section view), Figure 3 Sensor data acquisition system connection diagram,

[0067] Figure 4 Flow chart of the method of the present invention,

[0068] Figure 5 Initial sensor acquisition data flow diagram,

[0069] Figure 6 After separation, single response data,

[0070] Figure 7 No response data after separation.

[0071] The reference numerals are listed as follows: 1 is a stress sensor, 2 is a lateral strain sensor, and 3 is a vertical strain sensor. 4 is an acquisition channel protection module, 5 is a sensor input terminal, 6 is a surge pulse protector, 7 is an electromagnetic interference suppression filter, 8 is a low-frequency filter, 9 is a signal isolator, 10 is a high-speed acquisition unit input terminal, 11 is a high-frequency data acquisition unit, 12 is a first-level cache, 13 is a high-performance front-end data processing unit A, 14 is a second-level cache, and 15 is a high-performance front-end data processing unit B. DETAILED DESCRIPTION

[0072] The present invention is further described in detail below in conjunction with the embodiments.

[0073] The specific process of the method of the present invention is as follows:

[0074] (1) Establish a sensor archive database. The data table storage fields include: number, sensor type, model, manufacturer, burial location, burial depth, measurement direction, correction model, model parameters, zero position reading, and zero load reading.

[0075] (2) Number the sensor and create a corresponding sensor file in the database, entering the number, sensor type, model, and manufacturer information;

[0076] (3) Use customized tooling to independently calibrate each sensor and enter the calibrated data parameters into the sensor archive database, including: correction model, model parameters, and zero position reading;

[0077] (4) According to the attached Figure 1 , 2 The sensors are placed at the corresponding positions at the bottom of each structural layer of the road, and the buried information is entered into the database, including: buried location, buried depth, and measurement direction;

[0078] The arrangement position of the mechanical sensor is based on the center line of the wheel track on the left side of the road. The mechanical sensors are buried layer by layer to the bottom of each structural layer of the road. The lateral strain sensor 2 and the vertical strain sensor 3 are arranged in a T shape and buried in the asphalt layer and the water-stabilizing layer in a 2x2 matrix form, which are used to measure the lateral and longitudinal strains of the bottom of the structural layer when a vehicle passes by. The stress sensor 1 is buried in all structural layers at the center of the strain sensor matrix, including the asphalt layer, the water-stabilizing layer and each layer of the roadbed, and is used to measure the vertical stress at the bottom of the structural layer when a vehicle passes by.

[0079] (5) Connect the sensor input terminals 5 to the acquisition channel protection module 4 one by one, and install the protection module to the high-speed acquisition unit input terminal 10; the acquisition channel protection module includes a surge pulse protector 6, an electromagnetic interference suppression filter 7, a low-frequency filter 8 and a signal isolator 9 connected in sequence.

[0080] (6) Setting the acquisition frequency of the high-frequency data acquisition unit 11 to 2000 Hz or higher, and setting the acquisition unit working mode to 24-hour uninterrupted acquisition;

[0081] (7) reading sensor data when no vehicle passes through, and storing it in the zero load reading field of the sensor archive database;

[0082] (8) The high-frequency data acquisition unit 11 stores the sensor data in the form of a data stream to the first-level cache 12;

[0083] (9) The high-performance front-end data processing unit A13 reads the sensor data from the first-level cache 12 and uses a fluctuation detection algorithm of a dynamic window difference to calculate the data flow fluctuation rate and the current data flow status in real time;

[0084]

[0085] Where P is the volatility of the current data stream, x i is the real-time data in the dynamic window, and B is the average baseline value of the data stream.

[0086] (10) When the fluctuation rate is greater than the set threshold, the current time is recorded as the start time of the mechanical response of the road structure, and the data stream state is updated to "dynamic response";

[0087] (11) When the data flow fluctuation rate stabilizes and returns to below the threshold, the current time is recorded as the time when the response ends, and the data flow status is updated to "no dynamic response";

[0088] (12) Based on the recorded time points, the data stream from 0.5 s after the end of each response to 0.5 s before the start of the next response is averaged at a sampling rate of 1 Hz;

[0089] 1Hz sampling rate average processing formula: where x j is the sampling value after averaging for 1 second, F is the sampling rate of the high-frequency acquisition unit, x i Sensor response data;

[0090] (13) The high-performance front-end data processing unit A13 pushes the compressed and filtered data stream to the secondary cache 14;

[0091] (14) The high-performance front-end data processing unit B15 calibrates the timestamp information of the data according to the local standard time;

[0092] (15) The high-performance front-end data processing unit B15 separates the response data and the non-response data according to the recorded response time, and forms an independent file with the high-frequency response data in units of single response, and forms an independent file with the low-frequency non-response data in units of hours;

[0093] (16) Extract basic information from the response data: start time, end time, number of data sampling points, data dimension (number of sensors + 1), and data accuracy;

[0094] (17) Calculate the characteristic information in the response data: data baseline, overall mean of the response data, mean of the upper 5% data, and mean of the lower 5% data;

[0095] Data baseline calculation formula: Where B is the data baseline, N is 5% of the total single response data, and x i Sensor response data;

[0096] The overall mean calculation formula of the response data is: in is the overall mean of the response data, M is the total amount of single response data, x i Sensor response data;

[0097] The formula for calculating the mean of the top 5% data is: in is the mean of the top 5% data, N is 5% of the total single response data, x′ i Arrange the top 5% of the sensor response data in descending order;

[0098] The formula for calculating the mean of the lower 5% data is: in is the mean of the next 5% of data, N is 5% of the total single response data, x″ i The sensor response data is arranged in descending order and the last 5% of the data;

[0099] (18) Uniformly number the response data, and encapsulate the entity data, basic information, and feature information into independent data resource packages;

[0100] (19) Extract basic information from non-response data: data file size, number of data sampling points, data dimension, and data accuracy;

[0101] The data dimension is calculated as follows: M = m + 1, where m is the number of sensors.

[0102] (20) The non-responsive data are uniformly numbered, and the entity data and basic information are encapsulated into independent data resource packages.

[0103] Taking the dynamic response data of a certain field observation station as an example, the observation station has 5,831 vehicles passing through it every day. The road structure consists of 2 asphalt structural layers, 3 water-stabilizing structural layers, and 2 roadbed structural layers. A total of 47 mechanical sensors are buried with a collection frequency of 2000 Hz. The total amount of data generated is about 34GB per day.

[0104] According to the content of the present invention, the specific implementation steps are as follows:

[0105] (1) Establish a sensor archive database, calibrate, bury and initially read the sensors, and store the relevant information in the database as follows:

[0106]

[0107] (2) After connecting the 47 sensor terminals one by one to the sensor input terminals of the acquisition channel protection module, the output terminals of the acquisition channel protection module are connected to the input terminals of the high-frequency data acquisition unit;

[0108] (3) After the acquisition device is connected, set the acquisition frequency of the high-frequency data acquisition unit to 2000 Hz, and set the working mode to 24-hour uninterrupted acquisition;

[0109] (4) After the high-frequency data acquisition unit stores the sensor data in the first-level cache in the form of data stream, the high-performance data processing unit reads the sensor data from the first-level cache, calculates the data stream volatility in real time, and updates the data stream status at the current moment;

[0110] (5) When the vehicle passes by, the calculated volatility is 0.08, which is greater than the set threshold of 0.05. The current time 7:14:36.254 is recorded as the start time of the mechanical response of the road structure, and the data stream state is updated to "dynamic response";

[0111] (6) After the vehicle passes, the data flow volatility stabilizes and returns to 0.001, which is less than the set threshold. The current time 7:14:37.978 is recorded as the response end time, and the data flow status is updated to "no dynamic response";

[0112] (7) According to the recorded time point, the next response start time is 7:21:15.470, the data from 7:14:38.478 to 7:21:14.970 is averaged at a sampling rate of 1 Hz, stored in the data stream and pushed to the L2 cache;

[0113] (8) The high-performance data processing unit performs time calibration on the timestamp information of the data stream according to the local standard time;

[0114] (9) The high-performance data processing unit separates the high-frequency response data and the low-frequency no-response data according to the recorded response time, the high-frequency response data is formed into an independent file in units of single response, and the low-frequency no-response data is formed into an independent file in units of hours;

[0115] (10) Extract the basic information of a single response: start time 7:14:35.754, end time 7:14:38.478, number of data sampling points 5448, data dimension 48, and data accuracy 0.001;

[0116] (11) Calculate the characteristic information of a single response: data baseline 4.52, overall mean of response data 5.02, upper 5% data mean 6.38, and lower 5% data mean 4.27;

[0117] (12) Uniformly number the single response data, and encapsulate the entity data, basic information and feature information into an independent data resource package; HLJ01_2020-09-29_07_5_1548547296060.

[0118] (13) Extract the basic information of the non-response data: data file size 12.5KB, number of data sampling points 2628, data dimension 48, and data precision 0.001;

[0119] (14) Unify the numbering of non-responsive data and encapsulate the entity data and basic information into independent data resource packages; HLJ01_2020-09-29_07.

Claims

1. A method for acquiring dynamic mechanical response data of a road structure, comprising the following steps:

1. High-frequency real-time collection of road structure mechanics data The mechanical sensors are buried in each structural layer of the road, and their signal data lines are connected to the high-frequency data acquisition unit after passing through the acquisition channel protection module. The working parameters of the data acquisition unit are set, and the acquisition frequency of the data acquisition unit is set to 2000Hz or higher. The working mode of the acquisition unit is set to 24-hour uninterrupted acquisition; 2. Filter low value density data The data stream volatility is calculated through the dynamic load response detection algorithm. When the volatility is greater than the set threshold, the current moment is recorded as the start time of the mechanical response of the road structure, and the data stream status is updated to "dynamic response". When the data stream volatility is lower than the set threshold, the current moment is recorded as the end time of the response, and the data stream status is updated to "no dynamic response". The non-response data is compressed according to the low-frequency sampling frequency based on the recorded time.

3. Processing of Mechanical Response Data Processing the processed data stream, including: data timestamp calibration; separation of high-frequency and low-frequency data, forming independent data units for each response; extraction and processing of response data metadata information; and extraction and processing of non-response data metadata information; 4. Data Resource Packaging Each independent data unit and its metadata information are uniformly packaged and numbered according to standards to form scientific data resources.

2. According to the acquisition method of claim 1, the data flow volatility P is calculated by: Where P is the volatility of the current data stream, x i is the real-time data in the dynamic window, and B is the average baseline value of the data stream.

3. According to the acquisition method of claim 1, the high frequency is 2000HZ, the low frequency is 1HZ, and the method of compressing the non-response data according to the low frequency sampling frequency is: the data stream from 0.5s after the end of each response to 0.5s before the start of the next response is averaged at a sampling rate of 1Hz, and the formula for the 1Hz sampling rate average processing is: where x j is the sampling value after averaging for 1 second, F is the sampling rate of the high-frequency acquisition unit, x i The sensor response data.

4. According to the acquisition method described in claim 1, the response data metadata information extraction and processing includes the following data information: data start time, data end time, number of data sampling points, data dimension, data accuracy, data baseline, overall mean of response data, upper 5% data mean and lower 5% data mean; the non-response data metadata information extraction and processing includes the following data information: data file size, number of data sampling points, data dimension and data accuracy.

5. The acquisition method according to claim 4, The data dimension is the number of sensors plus the data timestamp, and the calculation formula is: M=m+1, where m is the number of sensors; The data baseline calculation formula is: Where B is the data baseline, N is 5% of the total single response data, and x i Sensor response data; The overall mean calculation formula of the response data is: in is the overall mean of the response data, M is the total amount of single response data, x i Sensor response data; The calculation formula for the mean of the upper 5% data is: in is the mean of the top 5% data, N is 5% of the total single response data, x′ i Arrange the top 5% of the sensor response data in descending order; The formula for calculating the mean of the lower 5% data is: in is the mean of the next 5% of data, N is 5% of the total single response data, x″ i The sensor response data is sorted in descending order and contains the last 5% of the data.

6. According to the acquisition method of claim 1, the number is the number of the response data resource and the non-response data resource, which includes the following information: collection observation station, collection date, hour time, response sequence number within this hour and starting data timestamp information, and the code is unique.

7. The acquisition method according to claim 1 further comprises establishing an archive database of mechanical sensors, wherein the database storage fields include: Number, sensor type, model, manufacturer, burial location, burial depth, measurement direction, correction model, model parameters, zero position reading, zero load reading.

8. A system for acquiring dynamic mechanical response data of a road structure designed according to any acquisition method of claims 1 to 7, comprising a data acquisition subsystem, a data processing and temporary storage subsystem, and a data processing and packaging subsystem; The data acquisition subsystem includes mechanical sensors located at each structural layer of the road, an acquisition channel protection module and a high-frequency data acquisition unit. The signal data line of the mechanical sensor is connected to the high-frequency data acquisition unit through the acquisition channel protection module; the high-frequency data acquisition unit acquires dynamic mechanical response data in the mechanical sensor; The data processing and temporary storage subsystem includes a high-performance front-end data processing unit A, a first-level cache unit and a second-level cache unit. The response data collected by the high-frequency data collection unit is stored in the first-level cache unit in the form of a data stream; The high-performance front-end data processing unit A reads the response data from the first-level cache unit, calculates the data flow volatility in real time, and updates the data flow state; when a vehicle passes by, if the calculated volatility is greater than a set threshold, the current moment is recorded as the start time of the mechanical response of the road structure, and the data flow state is updated to "dynamic response"; after the vehicle passes, the data flow volatility is less than the set threshold, the current moment is recorded as the response end time, and the data flow state is updated to "no dynamic response"; the data from the response end time to the next response start time are compressed as a whole at a low-frequency sampling rate, and the dynamic response and no dynamic response data streams are temporarily stored in the second-level cache unit; The data processing and archiving subsystem includes a high-performance front-end data processing unit B, which reads data from the secondary cache unit and calibrates the timestamp information of the data according to the local standard time; separates the response data and the non-response data according to the recorded response time, forms an independent file with a single response as the unit of high-frequency response data, and forms an independent file with an hour as the unit of low-frequency non-response data; Extract basic information from response data and non-response data: start time, end time, number of data sampling points, data dimension, and data accuracy; Calculate the data information in the response data: data baseline, overall mean of response data, upper 5% data mean and lower 5% data mean; uniformly number the response data and non-response data, and encapsulate the entity data, basic information and feature information into independent data resource packages.

9. According to the acquisition system of claim 8, the arrangement position of the mechanical sensor is based on the center line of the wheel track on the left side of the road, and the mechanical sensors are buried layer by layer to the bottom of each structural layer of the road, wherein the strain sensors are buried in the asphalt layer and the water-stable layer in a 2x2 matrix form, and are used to measure the lateral and longitudinal strains at the bottom of the structural layer when a vehicle passes by, and the stress sensors are buried in all structural layers at the center position of the strain sensor matrix, including the asphalt layer, the water-stable layer and each layer of the roadbed, and are used to measure the vertical stress at the bottom of the structural layer when a vehicle passes by, wherein the strain sensors include pairs of lateral strain sensors and vertical strain sensors, and the lateral strain sensors and vertical strain sensors are arranged in a T-shape in the matrix.

10. The acquisition system according to claim 8, wherein the acquisition channel protection module comprises a surge pulse protector, an electromagnetic interference suppression filter, a low-frequency filter and a signal isolator connected in sequence.

Images