A multi-channel non-contact mechanical loading and measuring method and system for a stirrup U-shaped clamping type connection pull-out test

Abstract

This invention provides a multi-channel non-contact mechanical loading and measurement method and system for pull-out tests of U-shaped clamp connections for stirrups, relating to the field of bridge engineering testing technology. The method constructs a unified time-series reference for multi-source data, simultaneously acquiring confining pressure, pull-out force, distributed fiber optic frequency shift, and acoustic emission signals. Through wavelet denoising, wavelength division multiplexing decoupling, and strain conversion, strain data of the reinforcing steel and concrete are obtained. A unique slip field spatiotemporal reconstruction algorithm generates a continuous slip field, which, combined with acoustic emission monitoring, enables three-dimensional localization and type identification of microcracks in concrete. A bond force component decomposition model is constructed to quantify the proportions of cementing force, frictional resistance, and mechanical interlocking force, and to invert the effective anchorage stiffness and critical anchorage length. This invention solves the problems of distortion in traditional equipment simulation and the difficulty of measuring internal slippage of dense reinforcing steel bars with extensometers. It offers high experimental accuracy and comprehensive data, providing technical support for the study of the anchorage mechanism of clamp connections.

Description

Technical Field

[0001] This invention belongs to the field of bridge engineering testing technology, and in particular relates to a multi-channel non-contact mechanical loading and measurement method and system for pull-out testing of stirrup U-shaped clip connections. Background Technology

[0002] Interlocking connections are an important structural form for rebar connection in large concrete bridge towers and prefabricated components, and their bond anchorage performance directly affects the overall structural safety. Current research on the mechanical properties of interlocking connections mainly relies on bidirectional pull-out tests. However, existing bidirectional pull-out equipment cannot accurately simulate the oblique compression and hook splitting effects on concrete present in interlocking connections, resulting in significant differences between the test conditions and the actual engineering stress state. Furthermore, traditional extensometers cannot accurately measure the internal slippage between densely packed rebars, making it difficult to obtain the continuous distribution characteristics of slippage in the lap joint, leading to insufficient accuracy in the test results and failing to meet the needs of refined research on the anchorage mechanism of interlocking connections. Summary of the Invention

[0003] The purpose of this invention is to provide a multi-channel non-contact mechanical loading and measurement method and system for pull-out testing of stirrup U-shaped clip connections, so as to solve the problems mentioned in the background art.

[0004] In view of this, the present invention provides a multi-channel non-contact mechanical loading and measurement method for pull-out tests of stirrup U-shaped clamp connections, the method comprising: S1. Establish a unified time series reference for multi-source heterogeneous data, and assign a globally unified timestamp to confining pressure data, axial pull-out force data, distributed optical fiber frequency shift data, and acoustic emission signal data. S2. Synchronously acquire confining pressure time-series data, axial pull-out force time-series data, distributed fiber frequency shift time-series data, and multi-channel acoustic emission raw waveform time-series data of the plug-in connected specimen during the controlled confining pressure loading and axial pull-out loading process. S3. Perform wavelet threshold denoising, wavelength division multiplexing decoupling and frequency shift-strain conversion on the distributed optical fiber frequency shift time series data, and separate and output the steel bar strain time series data and the concrete strain time series data. S4. Based on the time series data of steel bar strain and concrete strain, a two-dimensional slip field distribution data that is continuous in time and spans the entire domain along the lap length of the steel bar is generated by using a slip field spatiotemporal reconstruction algorithm. S5. Perform filtering, amplification, event extraction, three-dimensional time difference localization and feature classification on the original acoustic emission waveform time series data to obtain three-dimensional coordinate time series data of microcracks inside concrete, crack propagation path time series data and crack type distribution data. S6. By integrating confining pressure time series data, axial pull-out force time series data, slip field distribution data, and crack evolution correlation data, a bonding force component decomposition model is constructed to quantitatively output the proportion of cementing force, frictional resistance, and mechanical interlocking force. S7. Based on the slip mutation characteristics and crack evolution law, invert the effective anchorage stiffness and critical anchorage length and output the visualization results.

[0005] In a further embodiment of the present invention, step S1 establishes a unified time-series benchmark for multi-source heterogeneous data, including: S11. Generate a high-precision global synchronization clock signal with a clock resolution of not less than 1μs, and assign a unified timestamp sequence to the four data channels of confining pressure, pull-out force, fiber frequency shift and acoustic emission. S12. Monitor the local clock deviation of each data channel in real time. When the timing deviation between channels exceeds 1ms, trigger dynamic timing correction to correct the deviation to ≤0.5ms. S13. Construct a time-series index mapping table, using the global timestamp as the primary key, and establish a one-to-one correspondence between the confining pressure data index, pull-out force data index, fiber optic frequency shift data index, and acoustic emission data index. S14. Generate a time-series synchronization verification log, recording the time point, deviation value, and correction amount of each correction, for data traceability and consistency verification.

[0006] In a further embodiment of the present invention, the S2 synchronously acquires multi-source time-series data, including: S21. Receive the confining pressure loading control command, parse the segmented confining pressure parameter sequence, generate the corresponding confining pressure loading drive signal, and synchronously collect the real-time confining pressure value output by the confining pressure sensor to form confining pressure time series data. S22. Receive the axial pull-out loading control command, parse the loading rate and target load parameters, generate the pull-out loading drive signal, and synchronously collect the real-time pull-out force value output by the force sensor to form axial pull-out force time sequence data. S23. Configure the distributed optical fiber demodulator to scan at a frequency of 100Hz, and collect Brillouin frequency shift signals of the optical fiber embedded in the steel bar and the optical fiber pre-embedded in the concrete in parallel. Separate the two frequency shift data according to the optical fiber wavelength label to form distributed optical fiber frequency shift time series data. S24. Configure the acoustic emission acquisition system with a sampling rate of 1MHz, and synchronously acquire the original waveforms output by no less than 4 acoustic emission sensors, retaining the amplitude, duration, rise time, and energy characteristic parameters of the waveforms to form multi-channel acoustic emission original waveform timing data.

[0007] In a further embodiment of the present invention, the S3 frequency domain decomposition and feature extraction includes: S31. Perform wavelet multi-scale decomposition on the distributed optical fiber frequency shift time series data, select 3 to 5 levels of decomposition scale, use a soft threshold function to suppress high-frequency noise components, and retain effective frequency shift signal components. S32. Based on the wavelength division multiplexing decoupling algorithm, a frequency domain filter bank is constructed according to the difference in center wavelength between the fiber embedded in the steel bar and the fiber pre-embedded in the concrete to separate the frequency shift time-series component of the fiber embedded in the steel bar and the fiber pre-embedded in the concrete. S33. Based on the Brillouin frequency shift-strain conversion model: ε=k·Δν, where ε is strain, Δν is the Brillouin frequency shift, and k is the calibration coefficient, the two frequency shift components are converted into steel strain time series data and concrete strain time series data respectively. S34. Time-series alignment of steel bar strain time-series data and concrete strain time-series data is performed. Data interpolation is completed based on the time-series index mapping table, and outlier values ​​in the strain data are removed using the 3σ criterion.

[0008] In a further embodiment of the present invention, the S4 slip field spatiotemporal reconstruction algorithm includes: discretizing the lap length of the reinforcing bar into N equidistant measuring points, where the reinforcing bar strain at measuring point i (i=1,2,…,N) at time t is denoted as εs(i,t) and the concrete strain is denoted as εc(i,t). The slip amount s(i,t) of the measuring point is calculated according to the following formula: Where τ is the integration time variable, and α is the spatial smoothing coefficient, ranging from 0.01 to 0.1. It is a two-dimensional Laplace operator; S41. Substitute the values ​​of each measurement point and each time step into the formula to calculate the slip s(i,t), and generate an N×T slip time series matrix, where T is the total number of sampling times; S42. Perform Kriging space interpolation on the slip time series matrix to supplement the slip data between measurement points and generate two-dimensional slip field grid data with a resolution of not less than 1mm×10ms. S43. Traverse the slip field grid data, identify the moment when the slip abrupt change exceeds the preset threshold, and mark it as the slip abrupt change point at the moment of steel bar yielding; S44. Generate a dynamic evolution sequence based on two-dimensional slip field grid data, and output slip field thermogram, slip-time curve and slip mutation characteristic parameters.

[0009] In a further embodiment of the present invention, the S5 acoustic emission crack localization and evolution analysis includes: S51. Perform 10kHz to 1MHz bandpass filtering and 40dB gain amplification on the original waveform timing data of multi-channel acoustic emission to eliminate low-frequency environmental noise and high-frequency electromagnetic interference. S52. Set the acoustic emission event trigger threshold, extract valid acoustic emission events with amplitudes exceeding the threshold, and record the trigger time of the event, the arrival time difference of each channel, and the characteristic parameters. S53. Based on the time difference of arrival of multiple channels, a three-dimensional time difference positioning equation system is constructed, and the least squares method is used to solve iteratively to obtain the three-dimensional spatial coordinates (x, y, z) of the microcrack. S54. Correlate the three-dimensional coordinates of the microcrack with the geometric coordinates of the hook apex according to the time series, calculate the time series data of the distance between the crack and the hook apex, and generate the time series data of the crack propagation path. S55. Construct a crack type identification classifier, using the amplitude, duration, rise time, and energy of the acoustic emission signal as input features, to distinguish between tension cracks, shear cracks, and interface peeling cracks, and output crack type distribution data.

[0010] In a further embodiment of the present invention, the S6 adhesive component decomposition model includes: S61. Extract the slip field data of the low slip stage with slip ≤ 0.05 mm, construct the nonlinear mapping model of cementing force-slip: F1=f1(s), and fit it to obtain the cementing force ratio η1=F1 / F, where F is the total pull-out force; S62. Extract slip field data for the stable slip stage with slip of 0.05 to 0.5 mm, and construct a frictional resistance-confining pressure coupled mapping model: F2=f2(s,σ), where σ is the real-time confining pressure. Fit the model to obtain the frictional resistance ratio η2=F2 / F. S63. Extract high-risk stage data with crack propagation rate ≥ 0.1 mm / s in the hook area, and construct a mechanical interlocking force-crack rate mapping model: F3 = f3(vc), where vc is the crack propagation rate. Fit the model to obtain the mechanical interlocking force ratio η3 = F3 / F. S64. Normalize and correct η1, η2, and η3 to satisfy η1+η2+η3=1, and output the quantitative results of the adhesive force components and the time-series curve of their proportions.

[0011] In a further embodiment of the present invention, S11 further includes: S111. Analyze the segmented confining pressure parameters to obtain the confining pressure gradient sequence {σ1,σ2,…,σm} along the height direction of the specimen, where m is the number of segments; S112. Bind the confining pressure gradient sequence with a global timestamp to generate a non-uniform gradient confining pressure time series signal to simulate the distribution difference of confining pressure with height within a depth range of 5 to 10 meters in the bridge tower. S113. Establish a mapping between confining pressure gradient and multi-source data so that each segment of confining pressure parameter corresponds to matching slip field data and crack evolution data.

[0012] In a further embodiment of the present invention, S44 further includes: S441. Calculate the slip stiffness K(t)=dF / ds based on the dynamic evolution data of the slip field, and extract the slip amount when the slip stiffness decays to 50% of the initial value as the basis for determining the effective anchoring stiffness. S442. Using the lap length corresponding to the slip change point as a benchmark, and extrapolating the slip field distribution data, determine the minimum anchorage length when the steel stress reaches the yield strength, and output the critical anchorage length value. S443. Generate the effective anchorage stiffness-confining pressure relationship curve, the critical anchorage length-slippage relationship curve, and the corresponding visualization charts.

[0013] A multi-channel non-contact mechanical loading and measurement system for pull-out testing of U-shaped stirrup connections, used to realize a multi-channel non-contact mechanical loading and measurement method for pull-out testing of U-shaped stirrup connections, comprising: The timing reference construction module is used to generate a high-precision global synchronization clock, dynamically correct channel timing deviations, build a timing index mapping table, and generate timing synchronization verification logs. The multi-source data synchronous acquisition module is used to parse confining pressure and pull-out control commands, synchronously acquire confining pressure / pull-out force timing data, acquire distributed fiber optic frequency shift data at 100Hz, and acquire multi-channel acoustic emission raw waveform data at 1MHz. The frequency domain decomposition and feature extraction module is used for wavelet thresholding noise reduction, wavelength division multiplexing decoupling, Brillouin frequency shift-strain conversion, time alignment and outlier removal; The slip field spatiotemporal reconstruction module is used to execute the slip field spatiotemporal reconstruction algorithm, generate the slip amount time series matrix, perform kriging space interpolation, identify slip abrupt change points, and output slip field dynamic evolution data. The acoustic emission crack location analysis module is used for acoustic emission waveform filtering and amplification, effective event extraction, three-dimensional time difference location, crack propagation path association, and intelligent crack type identification. The adhesive force component decomposition module is used to construct a mapping model of adhesive force / frictional resistance / mechanical interlocking force, normalize the component proportions, and output the adhesive force quantification results. The anchorage performance inversion and visualization module is used to calculate the slip stiffness attenuation coefficient, invert the effective anchorage stiffness and critical anchorage length, and generate various relationship curves and visualization charts.

[0014] The beneficial effects of this invention are: Firstly, it effectively simulates the actual stress state of the plug-in connection, improving the realism of the test conditions. By establishing a unified time-series benchmark for multi-source heterogeneous data and simultaneously collecting confining pressure, axial pull-out force, distributed fiber frequency shift, and acoustic emission signals, it can realistically reproduce the oblique compression and hook splitting effect on concrete during the stress process of the plug-in connection. This overcomes the defects of the traditional bidirectional pull-out equipment in simulating the working conditions, making the test conditions more in line with engineering reality and significantly improving the reliability of the test results.

[0015] Secondly, it achieves full-domain, high-precision non-contact measurement of internal slippage between dense steel bars; by using distributed optical fibers to collect continuous strain data of steel bars and concrete respectively, a two-dimensional slip field distribution with full-domain and time-continuous along the lap length is generated through a slip field spatiotemporal reconstruction algorithm, replacing the traditional single-point measurement method at the end of the extensometer, accurately obtaining the entire evolution process of internal slippage between dense steel bars, solving the technical problem that traditional extensometers cannot accurately measure internal slippage between dense steel bars, and providing high-precision and high-coverage data support for the study of the bonding and anchoring mechanism of plug-in connection. Attached Figure Description

[0016] Figure 1 This is a flowchart of the steps of the method of the present invention.

[0017] Figure 2 This is a graph showing the effective anchoring stiffness-confining pressure relationship of the present invention. Figure 3 This is a graph showing the relationship between the critical anchorage length and the amount of slippage in this invention. Figure 4 This is a schematic diagram of the sliding field thermodynamics of the present invention; Figure 5 This is a time-series curve showing the proportion of the adhesive component in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0019] In the description of this application, it should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. For ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters in the following drawings denote similar items; therefore, once an item is defined in one drawing, it need not be further discussed in subsequent drawings.

[0020] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and are not limited in number; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0021] It should be noted that in the description of this application, the directional terms such as "front, back, up, down, left, right", "horizontal, vertical, horizontal" and "top, bottom" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description. Unless otherwise stated, these directional terms do not indicate or imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the scope of protection of this application. The directional terms "inner" and "outer" refer to the inner and outer contours relative to the outline of each component itself.

[0022] It should be noted that, in this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0023] Example 1: This embodiment provides a multi-channel non-contact mechanical loading and measurement method for pull-out tests of U-shaped stirrup connections. The method includes: S1. Establish a unified time series benchmark for multi-source heterogeneous data, and assign globally unified timestamps to confining pressure data, axial pull-out force data, distributed fiber optic frequency shift data, and acoustic emission signal data. This step aims to construct a unified time framework for collaborative processing of multi-source data, effectively simulate the actual stress state of plug-in connections, and improve the realism of test conditions. By establishing a unified time series benchmark for multi-source heterogeneous data, high-precision time alignment of confining pressure, axial pull-out force, distributed fiber optic frequency shift, and acoustic emission signals is achieved. This provides a time series guarantee for realistically reproducing the splitting effect of oblique compression and hooks on concrete during the stress process of plug-in connections. It overcomes the distortion of traditional bidirectional pull-out equipment, which can only apply axial force and cannot synchronously couple confining pressure and dynamic damage monitoring. This makes the test conditions highly consistent with the engineering reality of three-dimensional compression of 5-10 meter deep concrete in bridge towers, significantly improving the reliability and engineering reference value of test results.

[0024] S2. Synchronously acquire confining pressure time-series data, axial pull-out force time-series data, distributed fiber frequency shift time-series data, and multi-channel acoustic emission raw waveform time-series data of the plug-in connection specimen during controlled confining pressure loading and axial pull-out loading. This step relies on a unified timing reference to initiate mechanical loading control and multi-dimensional sensor data acquisition in parallel, realizing the synchronous linkage of confining pressure loading, axial pull-out, fiber strain monitoring, and acoustic emission damage monitoring. It completely captures the raw data of the plug-in connection from initial stress, slip development, crack initiation to final failure, avoiding the fragmentation of working conditions and information loss caused by asynchronous data acquisition. It provides a comprehensive, synchronous, and reliable raw data foundation for subsequent slip field reconstruction, crack location, and adhesive force component decomposition.

[0025] S3. Perform wavelet thresholding, wavelength division multiplexing decoupling, and frequency shift-strain conversion on the distributed fiber optic frequency shift time series data to separate and output the steel reinforcement strain time series data and the concrete strain time series data. This step addresses the problem that distributed fiber optic sensing signals are susceptible to environmental vibration, electromagnetic interference, and temperature drift, and that crosstalk can easily occur between the signals from the fiber optic cables embedded in the steel reinforcement and the fiber optic cables embedded in the concrete. Through a progressive signal processing flow, noise suppression, signal decoupling, and physical quantity conversion are completed sequentially, accurately converting the original frequency domain signals into steel reinforcement and concrete strain data that can be directly used for slip calculation. This ensures the accuracy, stability, and independence of strain measurement from the source, providing high-quality strain input data for subsequent high-precision slip field reconstruction.

[0026] S4. Based on the time-series data of steel bar strain and concrete strain, a two-dimensional slip field distribution data with full domain and continuous time along the lap length of the steel bar is generated through a slip field spatiotemporal reconstruction algorithm. Secondly, it realizes full-domain, high-precision non-contact measurement of internal slip between dense steel bars. Continuous strain data of steel bars and concrete along the lap length are collected using distributed optical fibers. Through the slip field spatiotemporal reconstruction algorithm, the strain difference of discrete measuring points is spatiotemporally integrated and spatially smoothed to generate two-dimensional slip field distribution data with full domain coverage and continuous time dimension along the lap length of the steel bar. This replaces the limitations of traditional extensometers, which can only measure single-point slip at the end of the steel bar, cannot be arranged between dense steel bars, and are difficult to obtain the internal continuous slip distribution. It accurately captures the entire process evolution characteristics of internal slip between dense steel bars from micro-slip, stable slip to yield change, with a spatial resolution of up to 1 mm and a temporal resolution of up to 10 ms. It provides high-precision, high-coverage, and non-contact slip data support for the refined study of the bonding and anchoring mechanism of plug-in connections.

[0027] S5. Filtering, amplification, event extraction, three-dimensional time difference localization, and feature classification are performed on the original acoustic emission waveform time-series data to obtain three-dimensional coordinate time-series data of microcracks inside concrete, crack propagation path time-series data, and crack type distribution data. This step addresses the characteristics of invisible microcracks inside concrete, the highly transient nature of their initiation and propagation processes, and the complexity of their damage mechanisms. Through refined processing of the entire acoustic emission signal process, it achieves integrated dynamic monitoring of microcracks inside concrete, from signal capture, noise filtering, and event recognition to three-dimensional localization, path tracking, and type identification. This accurately reveals the initiation location, propagation direction, evolution rate, and spatial correlation with the hook apex of microcracks inside concrete under the hook splitting effect. It fills the technical gap in traditional experiments that cannot monitor the damage evolution inside concrete in real time, providing direct, dynamic, and quantitative damage data support for explaining the bonding failure mechanism of plug-in connections from the perspective of damage mechanism.

[0028] S6. Integrating confining pressure time-series data, axial pull-out force time-series data, slip field distribution data, and crack evolution correlation data, a bonding force component decomposition model is constructed to quantify the proportion of cementing force, frictional resistance, and mechanical interlocking force. This step integrates synchronous data from three core dimensions: mechanical loading, slip deformation, and damage evolution. Based on the differences in the contribution mechanisms of different slip stages and damage characteristics to bonding force, a multi-parameter coupled bonding force component decomposition model is constructed. This model accurately decomposes the total bonding force of the plug-in connection into three core components: cementing force dominated by the low slip stage, frictional resistance dominated by the stable slip stage, and mechanical interlocking force dominated by the hook splitting stage. The proportion of each component in the total bonding force is quantified, clarifying the mechanical contribution law of each bonding force component under different working conditions and different stress stages. This solves the bottleneck of existing technologies that cannot distinguish and quantify bonding force components, providing a quantitative basis for the research on the bonding and anchoring mechanism of plug-in connections and the optimization design of connection structures.

[0029] S7. Based on slip mutation characteristics and crack evolution laws, the effective anchorage stiffness and critical anchorage length are inverted and visualized results are output. This step relies on the slip mutation characteristics of the steel bar at the moment of yielding identified by slip field data, combined with the crack evolution rate and distribution law monitored by acoustic emission, to automatically calculate the two core anchorage performance parameters of the plug-in connection, namely the effective anchorage stiffness and the critical anchorage length, through the anchorage performance inversion algorithm. It generates multi-dimensional visualization results such as slip field thermogram, slip-time curve, anchorage stiffness-confining pressure relationship curve, and crack propagation trajectory animation, which intuitively present the spatiotemporal evolution characteristics of the plug-in connection anchorage performance. It realizes the closed loop of the entire process from raw data acquisition, solution analysis to result visualization, which greatly improves the readability and practicality of the test results and provides intuitive, quantitative and reliable technical support for plug-in connection engineering design, construction quality control and safety assessment.

[0030] In this embodiment, S1 establishes a unified timing reference for multi-source heterogeneous data, including: S11, generating a high-precision global synchronization clock signal with a clock resolution of not less than 1μs, and allocating a unified timestamp sequence for the four data channels: confining pressure, pull-out force, fiber frequency shift, and acoustic emission. In specific implementation, an existing commercial high-precision synchronization clock generator is used. This clock generator is based on isothermal crystal oscillator technology, and the time resolution of the output clock signal is stable at 1μs, with a time drift error of no more than 0.1μs / h, meeting the high-precision timing synchronization requirements of multiple channels. For the confining pressure data acquisition channel, axial pull-out force data acquisition channel, distributed fiber frequency shift data acquisition channel, and acoustic emission signal data acquisition channel, independent timestamp allocation units are configured respectively. Each unit is connected to the high-precision synchronization clock generator through a hardware synchronization bus to receive the global synchronization clock signal. When any data frame is acquired by each data channel, the corresponding timestamp allocation unit immediately extracts the current global clock time, generates a unique, continuous, and globally unified timestamp, and binds it to the data frame, realizing accurate timing alignment of the initial acquired data of the four types of data channels, and avoiding the problem of multi-channel data timing misalignment from the hardware level.

[0031] S12. Real-time monitoring of the local clock deviation of each data channel. When the timing deviation between channels exceeds 1ms, dynamic timing correction is triggered to correct the deviation to ≤0.5ms. In specific implementation, the existing digital clock deviation real-time monitoring algorithm is adopted, and a local clock counter is integrated into each data channel to continuously record the running time of the local clock of each channel. The synchronization clock deviation monitoring module reads the local clock count of each channel and the global synchronization clock count every 100ms, calculates the relative deviation between the local clock of each channel and the global synchronization clock, and then derives the timing deviation between any two data channels. The preset timing deviation is used. The threshold is 1ms. When the timing deviation between any two data channels exceeds this threshold, the system immediately triggers a dynamic timing correction process. The correction process adopts existing adaptive digital timing correction technology. On the one hand, it reduces the cumulative deviation between the local clock and the global clock by fine-tuning the sampling clock frequency of the corresponding data channel. On the other hand, it uses a high-order linear interpolation algorithm to perform time axis interpolation compensation on the data collected during the deviation period. Finally, the timing deviation between all channels is corrected and stably controlled within the range of ≤0.5ms, ensuring the timing consistency of multi-source data and providing a reliable timing foundation for subsequent cross-channel data fusion analysis.

[0032] S13. Construct a time-series index mapping table, using the global timestamp as the primary key, to establish a one-to-one correspondence between confining pressure data index, pull-out force data index, fiber optic frequency shift data index, and acoustic emission data index. Specifically, existing relational database indexing technology is used to create the time-series index mapping table in the local high-speed storage unit of the test system. This mapping table uses a 64-bit globally unified timestamp as the unique primary key field, and simultaneously sets four index fields: confining pressure data storage address, axial pull-out force data storage address, distributed fiber optic frequency shift data storage address, and acoustic emission raw waveform data storage address. After each data acquisition and corresponding global timestamp generation, the system immediately records the physical storage path or data block number of the four types of data corresponding to the current timestamp in the storage unit, forming a complete index record and writing it to the mapping table. The time-series index mapping table adopts an incremental writing mode, archiving index records of the entire test process in real time, supporting accurate retrieval by timestamp and rapid matching of multi-source data at corresponding moments, significantly improving the efficiency of subsequent data retrieval, fusion, and analysis, while ensuring the accuracy of data association.

[0033] S14. Generate a timing synchronization verification log, recording the time point, deviation value, and correction amount of each correction for data traceability and consistency verification. In specific implementation, existing real-time log generation and archiving technology is adopted, and a log recording unit is integrated into the dynamic timing correction module. Whenever the system triggers a dynamic timing correction process, the log recording unit automatically extracts and records key correction information, including: the globally unified timestamp corresponding to the correction trigger time, the maximum timing deviation between relevant data channels before correction, the clock frequency fine-tuning amount of each channel during correction, the time interval and compensation amount of interpolation compensation data, and the minimum timing deviation between channels after correction. The log file adopts the standard JSON format and is automatically archived and stored in the order of correction occurrence. The log data of the entire test process is tamper-proof and can be permanently retained. After the test, the timing synchronization verification log can be directly retrieved to fully trace the triggering cause, execution process, and correction effect of each timing correction. At the same time, the timing consistency of multi-source data is cross-verified through log data to ensure the traceability, authenticity, and reliability of test data, and meet the requirements of test data quality control and result verification.

[0034] In this embodiment, S2 synchronously acquires multi-source time-series data, including: S21, receiving confining pressure loading control commands, parsing to obtain a segmented confining pressure parameter sequence, generating corresponding confining pressure loading drive signals, and synchronously acquiring real-time confining pressure values ​​output by confining pressure sensors to form confining pressure time-series data; in specific implementation, the upper control terminal of the test system issues segmented confining pressure loading control commands according to the preset bridge tower depth confining pressure simulation scheme. The commands include 3-5 independent loading segments divided along the height direction of the specimen. Each loading segment corresponds to preset target confining pressure value, loading rate, loading holding time, loading termination conditions, and other parameters. All loading segment parameters are arranged in height order to form a segmented confining pressure parameter sequence; after receiving the confining pressure loading control commands, the system's command parsing module uses existing structured command parsing algorithms to extract each item in the segmented confining pressure parameter sequence. The parameters are formatted and validated to ensure their validity. Subsequently, the loading drive module generates a continuous analog drive signal adapted to the hydraulic confining pressure loading unit based on the segmented confining pressure parameter sequence. The drive signal includes the loading curve instructions for each loading segment, realizing non-uniform gradient confining pressure loading along the height direction of the specimen. Simultaneously, 3–4 high-precision piezoelectric confining pressure sensors are evenly distributed at the contact position between the hydraulic confining pressure loading unit and the outer surface of the concrete of the specimen. The sensor range covers 0–10 MPa, and the measurement accuracy reaches ±0.5%FS. The confining pressure sensors collect the actual confining pressure values ​​of each area of ​​the specimen in real time. The sampling frequency is consistent with the global synchronization clock (1kHz). The collected real-time confining pressure values ​​are sorted according to the globally unified timestamp to form structured confining pressure time series data and written to the storage unit in real time, truly recording the pressure changes throughout the entire confining pressure loading process.

[0035] S22. Receive the axial pull-out loading control command, parse the loading rate and target load parameters, generate a pull-out loading drive signal, and simultaneously collect the real-time pull-out force value output by the force sensor to form axial pull-out force time-series data. In specific implementation, the upper control terminal issues the axial pull-out loading control command according to the design bearing capacity of the plug-in connected specimen. The command includes parameters such as the preset loading rate (0.1–1 mm / min), multi-level target load values ​​(graded according to 20%, 40%, 60%, 80%, and 100% of the estimated ultimate bearing capacity), the duration of each load level, and the loading termination condition (specimen fracture or sudden load drop). After receiving the command, the system's command parsing module uses the existing loading command parsing algorithm to extract the loading rate and target load parameters. The system performs a validity check on the data. Based on the parsed parameters, the loading drive module generates a closed-loop control drive signal adapted to the servo electro-hydraulic actuator. The drive signal adopts a displacement-force dual closed-loop control mode to ensure a smooth loading process, accurate speed, and controllable load. Simultaneously, a high-precision spoke-type force sensor is installed at the connection position between the servo actuator and the pull-out end of the specimen. The sensor range covers 0–500kN, and the measurement accuracy reaches ±0.1%FS. The force sensor collects the axial pull-out force value applied to the specimen in real time. The sampling frequency is consistent with the global synchronization clock (1kHz). The collected real-time pull-out force values ​​are sorted according to the globally unified timestamp to form axial pull-out force time series data and stored in real time, completely recording the load change law of the specimen from initial loading to failure.

[0036] S23. Configure a distributed fiber optic demodulator with a scanning frequency of 100Hz to acquire Brillouin frequency shift signals from the fiber optic cable embedded in the reinforcing bar and the fiber optic cable pre-embedded in the concrete in parallel. Separate the two frequency shift data according to the fiber wavelength identifier to form distributed fiber optic frequency shift time series data. In specific implementation, select an existing commercial high-precision Brillouin distributed fiber optic demodulator and configure the core parameters of the demodulator as follows: fixed scanning frequency of 100Hz, spatial sampling interval of 10mm, wavelength resolution of 0.1nm, and frequency shift measurement accuracy of ±0.1MHz. Before the test, complete the fiber optic cable layout: implant a single-mode fiber as the fiber optic cable embedded in the reinforcing bar in the axial center deep hole of the lapped reinforcing bar of the plug-in connector. The fiber is continuously laid along the entire length of the reinforcing bar and is tightly attached to the reinforcing bar. Along the lapped length of the reinforcing bar, a second fiber with a different center wave is spirally pre-embedded in the concrete protective layer. Long single-mode optical fibers are used as embedded optical fibers in the concrete. The center wavelengths of the two fibers are selected as 1530nm and 1550nm respectively to avoid signal crosstalk. After the demodulator is started, it transmits pulse probe light in parallel to the fiber embedded in the steel bar and the fiber embedded in the concrete. It receives the Brillouin scattered light signal reflected back from the two fibers in real time and extracts the Brillouin frequency shift information in the scattered light. The signal separation module adopts the existing wavelength division multiplexing wavelength identification algorithm. Based on the difference in the center wavelength of the two fibers, a narrow bandpass filter is constructed to accurately separate the Brillouin frequency shift signal corresponding to the fiber embedded in the steel bar and the Brillouin frequency shift signal corresponding to the fiber embedded in the concrete. The two frequency shift signals are synchronously sorted according to a globally unified timestamp, merged to form distributed optical fiber frequency shift time series data and written to the storage unit in real time to completely record the frequency shift change information of the steel bar and concrete along the overlap length.

[0037] S24. Configure the acoustic emission acquisition system with a sampling rate of 1MHz, simultaneously acquiring the raw waveforms output by no fewer than 4 acoustic emission sensors, retaining the amplitude, duration, rise time, and energy characteristic parameters of the waveforms to form multi-channel acoustic emission raw waveform time-series data; in specific implementation, select an existing commercial multi-channel acoustic emission acquisition system, fixing the system sampling rate to 1MHz, analog-to-digital conversion accuracy to 16bit, and bandwidth covering 10kHz–1MHz to ensure complete capture of elastic wave signals generated by concrete microcracks; arrange no fewer than 4 high-precision acoustic emission sensors on the specimen surface in a spatial tetrahedral configuration, using piezoelectric sensors with a resonant frequency of 150kHz, and connecting them to the surface using a high-vacuum coupling agent. The specimen surface is tightly coupled to ensure that the elastic waves generated by the microcracks can be transmitted to the sensor without loss. After the acoustic emission acquisition system is started, it synchronously receives the continuous voltage raw waveform signals output by each sensor and performs digital sampling of the raw waveform signals in real time. The signal feature extraction module extracts four core feature parameters from the digital waveform in real time: waveform peak amplitude, waveform duration, rise time (time from threshold to peak), and waveform energy (absolute integral value). The raw waveform data and corresponding feature parameters of all channels are synchronously sorted and bound according to a globally unified timestamp to form multi-channel acoustic emission raw waveform time sequence data and stored in real time, completely preserving the original acoustic information of the entire process of microcrack initiation and propagation inside the concrete.

[0038] In this embodiment, the S3 frequency domain decomposition and feature extraction includes: S31, performing wavelet multi-scale decomposition on the distributed fiber frequency shift time series data, selecting 3 to 5 decomposition scales, using a soft threshold function to suppress high-frequency noise components, and retaining effective frequency shift signal components; specifically, using existing wavelet denoising algorithms, selecting the db4 orthogonal wavelet basis function, which has good time-frequency localization characteristics and is suitable for the noise distribution characteristics of fiber frequency shift signals; when performing wavelet multi-scale decomposition on the distributed fiber frequency shift time series data, adaptively selecting 3 to 5 decomposition scales according to the signal-to-noise ratio and noise frequency distribution, decomposing the original frequency shift time series data layer by layer into 1 low-frequency approximation component and 3 to 5 high-frequency detail components; wherein, the low-frequency approximation component contains fiber frequency shift components. The effective signal body corresponds to the actual strain changes of the steel bars and concrete; the high-frequency detail components mainly contain noise signals such as environmental vibration, electromagnetic interference, and temperature fluctuations, which have no actual physical meaning; during noise reduction, the existing soft threshold function is used to perform threshold processing on each high-frequency detail component: first, the standard deviation of each high-frequency component is calculated, and an adaptive threshold is set based on the 3σ criterion; for high-frequency noise components with amplitudes exceeding the threshold, attenuation processing is performed according to the threshold to suppress noise interference; for high-frequency components with amplitudes below the threshold, they are determined to be effective signal fluctuations and are completely preserved; after noise reduction, the low-frequency approximate component and the processed high-frequency detail component are reconstructed by wavelet to obtain the effective frequency-shifted signal component with noise interference removed, which significantly improves the signal-to-noise ratio of the frequency-shifted signal and ensures the accuracy of subsequent strain conversion.

[0039] S32. Based on the wavelength division multiplexing (WDM) decoupling algorithm, a frequency domain filter bank is constructed according to the difference in center wavelength between the fiber embedded in the reinforcing steel and the fiber pre-embedded in the concrete, separating the frequency shift time-series components of the fiber embedded in the reinforcing steel and the fiber pre-embedded in the concrete. In specific implementation, since the fiber embedded in the reinforcing steel and the fiber pre-embedded in the concrete use different center wavelengths (1530nm / 1550nm), the Brillouin frequency shift signals of the two fibers exhibit non-overlapping characteristic frequency bands in the frequency domain, providing a basis for signal decoupling. Using the existing WDM decoupling algorithm, a narrowband frequency domain filter bank is constructed based on the center wavelength parameters of the two fibers and the frequency band characteristics of the frequency shift signal. The system includes two bandpass filters with center frequencies matched to the 1530nm and 1550nm fiber frequency shift signals, respectively. The passband width of the filters is set to ±5MHz, and the transition band attenuation is greater than 40dB. The effective frequency shift signal component after S31 noise reduction is input into the frequency domain filter bank. Through the frequency filtering effect of the two bandpass filters, the residual crosstalk of the frequency shift signal of the other fiber is filtered out, and the frequency shift time sequence component corresponding to the fiber embedded in the steel bar and the fiber embedded in the concrete are accurately separated. The two separated frequency shift components are independent of each other and have no cross interference, providing clean frequency shift data for subsequent calculation of steel bar strain and concrete strain.

[0040] S33. Based on the Brillouin frequency shift-strain conversion model: ε = k·Δν, where ε is the strain, Δν is the Brillouin frequency shift, and k is the calibration coefficient. This converts the two frequency shift components into steel strain time-series data and concrete strain time-series data, respectively. In specific implementation, the Brillouin frequency shift-strain conversion model is a mature existing physical sensing model. Its principle is that when an optical fiber is subjected to axial strain, the frequency shift of the internal Brillouin scattered light is linearly positively correlated with the axial strain of the optical fiber. In the model expression, ε is the axial strain of the optical fiber (unit: με), Δν is the Brillouin frequency shift (unit: MHz), and k is the strain-frequency shift calibration coefficient (unit: με / MHz). The value of k is related to the optical fiber type, the demodulator's operating wavelength, and the ambient temperature. Before the experiment, the calibration coefficient k is determined through a standard strain calibration test: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] The fiber is fixed in a standard tensile testing machine, and a known graded standard strain is applied. The corresponding Brillouin frequency shift is collected synchronously, and the k value is obtained through linear fitting and fixed in the system. In this test, the k value is calibrated to 5.0 με / MHz. During the conversion calculation, for the time series component of the fiber optic frequency shift of the steel bar separated by S32, the Brillouin frequency shift Δνs corresponding to each global timestamp is extracted and substituted into the conversion model εs=k·Δνs to calculate the strain εs of each measuring point of the steel bar at the corresponding time. The strain time series data of the steel bar is formed by sorting by global timestamp. Similarly, for the time series component of the fiber optic frequency shift of the concrete, the frequency shift Δνc at each time is extracted and substituted into the model εc=k·Δνc to calculate the strain εc of each measuring point of the concrete, forming the time series data of concrete strain, realizing the accurate conversion from frequency domain signal to physical strain.

[0041] S34. Time-series alignment of steel bar strain time-series data and concrete strain time-series data is performed. Data interpolation is completed based on the time-series index mapping table, and outlier values ​​in the strain data are removed using the 3σ criterion. In practice, due to transient electromagnetic interference and slight fiber optic jitter during fiber optic signal acquisition, temporary signal interruptions may occur, resulting in local data loss in the steel bar strain time-series data and concrete strain time-series data. Preprocessing is required. The first step is time-series alignment: Based on the time-series index mapping table constructed in S13, the steel bar strain time-series data and concrete strain time-series data are aligned on the time axis using a globally unified timestamp as the reference. This ensures that the same global timestamp corresponds to a set of steel bar strain data and concrete strain data, eliminating time-series misalignment between the two strain data. The second step is interpolation completion: For data loss points found after time-series alignment (single or a few consecutive timestamps without strain data), The first step involves data completion. Using existing linear interpolation algorithms, strain data from two adjacent valid timestamps before and after a missing point are extracted. The strain value of the missing point is calculated through linear fitting to complete the data and ensure the temporal continuity of the steel strain time series data and the concrete strain time series data. The second step is outlier removal. Using the existing 3σ criterion outlier removal algorithm, anomalies are detected in both the steel strain time series data and the concrete strain time series data. The overall mean μ and standard deviation σ of the individual strain data are calculated, and the normal data range is set as [μ-3σ, μ+3σ]. Strain data exceeding this range are identified as outliers and removed. After outlier removal, a linear interpolation algorithm is used again to complete the data at the removed locations. Finally, continuous, complete, and anomaly-free high-quality steel strain time series data and concrete strain time series data are obtained, providing reliable strain input for subsequent spatiotemporal reconstruction of the slip field.

[0042] In this embodiment, the S4 slip field spatiotemporal reconstruction algorithm includes: discretizing the lap length of the reinforcing bar into N equidistant measuring points, where the reinforcing bar strain at measuring point i (i=1,2,…,N) at time t is denoted as εs(i,t) and the concrete strain is denoted as εc(i,t). The slip amount s(i,t) of the measuring point is calculated according to the following formula: Where τ is the integral time variable, and α is the spatial smoothing coefficient, ranging from 0.01 to 0.1. It is a two-dimensional Laplace operator; In practical implementation, the slip calculation formula is a spatiotemporally coupled slip calculation model. The core principle is that the slip is essentially the relative displacement between the steel bar and the concrete along the lap length, formed by the accumulation of strain difference over time. At the same time, a spatial smoothing term is introduced to suppress the abrupt change in the slip field caused by local strain noise, taking into account both the physical reality and spatial continuity of the slip field. Parameter settings and physical meaning: The total lap length L (unit: mm) of the steel bar is discretized into N equidistant measuring points. The measuring point spacing ΔL = L / N, where N is a positive integer. It is set according to the measurement accuracy requirements. In this experiment, the lap length L = 500 mm and the measuring point spacing ΔL = 10 mm, so N = 50. Measurement point i (i=1,2,…,N) corresponds to the i-th spatial position along the lap length; time t corresponds to a globally unified timestamp, representing the time dimension of slip development; εs(i,t) is the steel strain (unit: με) at measurement point i and time t, taken from the steel strain time series data processed by S34; εc(i,t) is the concrete strain (unit: με) at measurement point i and time t, taken from the concrete strain time series data processed by S34; τ is the integration time variable, representing any intermediate time from the start of the experiment (τ=0) to the current time t; the integration term... dτ is the time integral of the strain difference, which represents the cumulative amount of relative deformation between the steel bars and concrete over time and is the main contributor to the slip amount; α is the spatial smoothing coefficient, which ranges from 0.01 to 0.1. In this experiment, it is set to 0.05. It is used to balance the preservation of details and the suppression of noise in the slip field. If the value of α is too small, the smoothing effect is weak and noise is easily preserved. If the value of α is too large, the smoothing is over-smoothed and the details of slip abrupt changes are lost. s(i,t) is a two-dimensional Laplace operator used to perform spatial second-order difference smoothing on the slip field, suppressing spatial abrupt changes in the slip field caused by local strain noise at measurement points, and ensuring the spatial continuity of the slip field.

[0043] S41. Calculate the slip s(i,t) point by point and time by time by substituting into the formula, generating an N×T slip time series matrix, where T is the total number of sampling times. In specific implementation, a traversal calculation method is used to solve the slip in the entire time and space: First, parameter initialization: Determine the total number of measuring points N=50, and the total number of sampling times T corresponding to the total test duration (determined by the test duration and sampling frequency of 100Hz; for example, if the test duration is 10 minutes, then T=60000). Initialize an N-row T-column empty matrix as the slip time series matrix. Second, time-by-time and point-by-point traversal calculation: Traverse each sampling time t in time order (1≤t≤T), and at each time t, traverse each measuring point i in spatial order (1≤i≤N). For each (i,t) combination, extract the corresponding steel strain εs(i,t) and concrete strain εc(i,t), and substitute them into the slip calculation formula. Third, numerical solution: The time integral term in the formula... dτ is discretized using the existing trapezoidal numerical integration algorithm, which discretizes the integration interval [0,t] into t small intervals according to the sampling interval, and accumulates the strain difference integral value between each small interval; the two-dimensional Laplace operator s(i,t) is discretized using the existing five-point finite difference method. The second-order spatial difference is calculated using the slip of measurement point i and its adjacent measurement points before and after it. The fourth step is matrix filling: the slip s(i,t) calculated by each combination of (i,t) is filled into the i-th row and t-th column of the slip time series matrix. After traversal, a complete N×T slip time series matrix is ​​generated. The matrix rows correspond to spatial measurement points, the columns correspond to sampling times, and the matrix elements are the slip (unit: mm) of the corresponding spatiotemporal position, which completely records the spatiotemporal distribution characteristics of the card connection slip.

[0044] S42. Perform Kriging spatial interpolation on the slip time series matrix to supplement the slip data between measurement points and generate a two-dimensional slip field grid with a resolution of not less than 1 mm × 10 ms. In specific implementation, since the spacing between discrete measurement points is 10 mm, in order to improve the spatial resolution of the slip field and obtain the continuous slip distribution between measurement points, the existing ordinary Kriging spatial interpolation algorithm is used to spatially refine the slip time series matrix. The first step is to construct the interpolation grid: based on the spatial range (0–500 mm) and time range (0–T × 10 ms) of the slip time series matrix, a two-dimensional regular grid is constructed, with the spatial grid spacing set to 1 mm and the time grid spacing set to 10 ms to ensure that the slip field resolution is not less than 1 mm × 10 ms. The second step is to set the Kriging interpolation parameters: select the Gaussian variogram model and set the search neighborhood range to 3 times. The measurement point spacing is 30mm, and the number of effective measurement points participating in interpolation in the neighborhood is no less than 8 to ensure interpolation accuracy. The third step is grid-by-grid interpolation calculation: traverse all grid points to be interpolated in the two-dimensional grid, and for each grid point, search for the slip data of discrete measurement points in its neighborhood, and solve for the slip value of the grid point by using the Kriging interpolation algorithm. During the interpolation process, the spatial correlation of the slip of discrete measurement points is automatically considered to achieve unbiased and optimal interpolation. The fourth step is to generate slip field grid data: after the interpolation is completed, a two-dimensional slip field grid data covering the entire overlap length and the entire test duration is obtained, with a spatial resolution of 1mm and a temporal resolution of 10ms. Each grid point in the data contains the corresponding spatial location, time, and slip value, realizing the transformation of the slip field from discrete measurement points to the continuous whole field, accurately presenting the spatial distribution and temporal evolution details of the slip.

[0045] S43. Traverse the slip field grid data and identify the moment when the slip abrupt change exceeds a preset threshold, marking it as the slip abrupt change point at the moment of rebar yielding. In specific implementation, the moment of rebar yielding is accompanied by a sharp increase in slip. By identifying the abrupt change characteristics in the slip field, the yielding moment can be accurately located. The first step is to calculate the slip rate: For the two-dimensional slip field grid data, calculate the time change rate of slip at each spatial location along the time dimension (slip rate, unit: mm / s) to generate the spatiotemporal distribution data of slip rate. The second step is to set the abrupt change threshold: Based on the pre-experiment calibration and theoretical analysis, the preset slip abrupt change threshold is 0.1mm / 10ms (i.e., 10mm / s). This threshold can effectively... The first step distinguishes between normal slip development and yield abrupt change. The second step involves identifying abrupt change points by traversing the spatiotemporal distribution data of slip rate at each spatial location and time step. When the slip rate at a certain spatial location at a certain time exceeds a preset abrupt change threshold, and the slip rate remains above the threshold for three consecutive time steps, that time step is determined to be the instantaneous slip abrupt change point of the steel bar yielding. The third step involves marking the abrupt change point information by recording the corresponding global timestamp, spatial location, slip amount at the time of abrupt change, slip rate at the time of abrupt change, and other key information for the identified slip abrupt change point. This step can accurately capture the critical moment when the steel bar enters the yielding stage from the elastic stage, providing key feature points for subsequent inversion of anchorage performance parameters.

[0046] S44. Generate a dynamic evolution sequence based on two-dimensional slip field grid data, and output slip field heatmap, slip-time curve, and slip abrupt change characteristic parameters. In specific implementation, to intuitively present the spatiotemporal evolution law of the slip field, the two-dimensional slip field grid data is visualized and its characteristic parameters are extracted. The first step is to generate a dynamic evolution sequence of the slip field: in chronological order, the two-dimensional slip field grid data at different times are exported sequentially to generate a dynamic evolution sequence of the slip field with a frame interval of 10ms, which fully presents the entire process of slip from local initiation to gradual expansion to the whole domain. The second step is to output a slip field heatmap: select key characteristic moments (initial slip, stable slip, yield abrupt change, ultimate slip), and output the two-dimensional slip field grid data at the corresponding moments. The data is mapped to a color heatmap, with different colors representing the magnitude of slip, intuitively displaying the spatial distribution characteristics of the slip field. The third step outputs the slip-time curve: key spatial locations such as the midpoint and ends of the overlap length are selected, and data on the change of slip over time at the corresponding locations are extracted to generate a slip-time relationship curve, clearly presenting the evolution trend of slip over time. The fourth step outputs slip abrupt change characteristic parameters: information on slip abrupt change points marked by S43 is summarized, and core characteristic parameters such as the time of abrupt change, the location of the abrupt change, the amount of slip abruptly changed, and the rate of slip abruptly changed are output, providing a quantitative basis for anchoring performance analysis. Through the above outputs, the slip field data is transformed from a numerical matrix to a visualized result, significantly improving the readability and analytical efficiency of the slip evolution law.

[0047] In this embodiment, a further implementation is that the S5 acoustic emission crack localization and evolution analysis includes: S51, performing 10kHz~1MHz bandpass filtering and 40dB gain amplification on the multi-channel acoustic emission raw waveform time-series data to remove low-frequency environmental noise and high-frequency electromagnetic interference; in specific implementation, the acoustic emission raw waveform signal is easily affected by low-frequency mechanical vibration, environmental noise and high-frequency electromagnetic interference, so filtering and amplification preprocessing is required to extract the effective crack signal; the first step, bandpass filtering: using an existing finite-length unit impulse response (FIR) bandpass filter, setting the passband frequency to 10kHz~1MHz, and the stopband attenuation to be greater than 50dB. The process involves two steps: First, filtering out low-frequency environmental noise below 10kHz (such as mechanical vibration and on-site noise) and high-frequency electromagnetic interference above 1MHz (such as instrument electromagnetic radiation and line interference), while retaining the effective elastic wave signal frequency band generated by concrete microcracks. Second, gain amplification: the filtered acoustic emission waveform signal is amplified by a fixed gain of 40dB to increase the amplitude of the effective signal, preventing the weak crack signal from being drowned out by noise, while ensuring that the amplitude of the amplified signal does not exceed the range of the acquisition system to prevent signal saturation distortion. After preprocessing, an effective acoustic emission waveform signal with a significantly improved signal-to-noise ratio is obtained, laying the foundation for subsequent event extraction and location analysis.

[0048] S52. Set an acoustic emission event trigger threshold, extract valid acoustic emission events with amplitudes exceeding the threshold, and record the trigger time, arrival time difference of each channel, and characteristic parameters of the event. In specific implementation, the elastic wave signal generated when microcracks initiate and propagate inside concrete is a discrete pulse signal. By setting a trigger threshold, valid acoustic emission events can be extracted from the continuous waveform. First, set the trigger threshold: based on the noise level of the preprocessed acoustic emission waveform signal, set the amplitude trigger threshold to 40dB. Only when the peak amplitude of the waveform signal exceeds this threshold is it determined to be a valid acoustic emission event, and noise pulse interference is eliminated. Second, extract valid... The process involves three steps: First, traversing each channel of the preprocessed acoustic emission waveform signal to identify all pulse signals with peak amplitudes exceeding the trigger threshold, marking them as valid acoustic emission events. Second, recording event information: For each valid acoustic emission event, recording core information synchronously: the event trigger time (globally unified timestamp), the absolute arrival time of the event signal received by each channel sensor, the signal arrival time difference between each channel, the event amplitude, duration, rise time, and energy characteristic parameters. All valid acoustic emission event information is sorted by trigger time to form a valid acoustic emission event sequence, accurately recording the discrete event characteristics generated by concrete microcracks.

[0049] S53. Based on the time difference of arrival of multiple channels, a three-dimensional time difference positioning equation system is constructed, and the least squares method is used for iterative solution to obtain the three-dimensional spatial coordinates (x, y, z) of the microcrack. In specific implementation, the arrival time difference of signals from no less than 4 acoustic emission sensors is used to invert the spatial location of the microcrack through a three-dimensional time difference positioning algorithm. The first step is to construct the positioning equation system: Let the three-dimensional coordinates of the j-th sensor in the acoustic emission sensor array be (xj, yj, zj) (known, accurately measured before the experiment), the three-dimensional coordinates of the microcrack generated by a certain effective acoustic emission event be (x, y, z) (to be determined), the propagation speed of elastic wave in concrete be v (calibrated before the experiment, taken as 3500 m / s in this experiment), and the absolute arrival time of the signal received by the j-th sensor be tj. With the first sensor as a reference, the arrival time difference equation system is constructed: (√[(x-xj)] 2 +(y-yj) 2 +(z-zj) 2 ]-√[(x-x1) 2 +(y-y1) 2 +(z-z1) 2 The second step is iterative solution: This system of equations is a nonlinear system of equations, and the existing Gauss-Newton least squares method is used for iterative solution: First, the initial coordinate estimate of the microcrack is set, and the coordinate value is continuously corrected through iteration to minimize the sum of squares of the residuals of the system of equations; The convergence condition of the iteration is set to the coordinate correction amount being less than 0.1mm to ensure the positioning accuracy; The third step is to output the three-dimensional coordinates: After the iteration converges, the three-dimensional spatial coordinates (x,y,z) of the microcrack are output, with a positioning error ≤2mm, to accurately determine the crack initiation position of the microcrack inside the concrete; The above positioning calculation is performed for each valid acoustic emission event to obtain the three-dimensional coordinate time series data corresponding to all microcrack events.

[0050] S54. Correlate the three-dimensional coordinates of the microcrack with the geometric coordinates of the hook apex according to the time series, calculate the time-series data of the distance between the crack and the hook apex, and generate time-series data of the crack propagation path; In specific implementation, the core of the splitting effect of the hook on concrete is reflected in the initiation and propagation of cracks around the hook apex. By correlating the crack coordinates with the hook apex coordinates, the splitting evolution law can be revealed; First step, determine the hook apex coordinates: Before the experiment, accurately measure the geometric coordinates of the hook apex using a three-dimensional scanning method, and determine the three-dimensional coordinates (x0, y0, z0) of the apex center point; Second step, calculate the distance time-series data: For the three-dimensional coordinates (x, y, z) of each microcrack event, calculate its spatial distance d = √[(x-x0)] from the center point of the hook apex. 2 +(y-y0) 2 +(z-z0) 2The first step is to sort the events by their trigger times to form time-series data on the distance between the crack and the hook apex. The second step is to generate crack propagation paths: in chronological order, the coordinates of continuously occurring and spatially adjacent microcrack events are fitted with trajectories, and existing cubic spline interpolation algorithms are used to generate smooth crack propagation path time-series data. This data can clearly show the spatial trajectory of microcracks inside concrete from their initial initiation location to the surrounding area, as well as the dynamic distance changes between them and the hook apex, accurately revealing the crack propagation law under the hook splitting effect.

[0051] S55. Construct a crack type identification classifier, using the amplitude, duration, rise time, and energy of the acoustic emission signal as input features to distinguish between tension cracks, shear cracks, and interface peeling cracks, and output crack type distribution data. In specific implementation, the acoustic emission signal characteristics generated by different types of concrete cracks differ significantly. Intelligent crack type identification can be achieved through a machine learning classifier. The first step is to construct a feature dataset: extract four core feature parameters for each valid acoustic emission event: amplitude, duration, rise time, and energy, forming a four-dimensional feature vector. The second step is to construct a classifier: use the existing Support Vector Machine (SVM) classification algorithm to construct a crack type identification classifier. The classifier is trained using a pre-labeled sample dataset (acoustic emission feature data corresponding to tension, shear, and interface peeling cracks). The classifier parameters are optimized through training to ensure a classification accuracy of ≥95%. The third step is crack type classification: the four-dimensional feature vector of each valid acoustic emission event is input into the trained classifier, which automatically outputs the corresponding crack type label (tension / shear / interface peeling). The fourth step is outputting distribution data: according to the experimental time and spatial location, the number, distribution area, and occurrence time of different types of cracks are statistically analyzed to generate crack type distribution data and a visualization cloud map, clarifying the evolution ratio and distribution pattern of different types of cracks during the hook splitting process.

[0052] In this embodiment, a further implementation is that the S6 adhesive force component decomposition model includes: S61, extracting slip field data for the low-slip stage with slip ≤0.05mm, constructing a nonlinear mapping model of adhesive force-slip: F1=f1(s), and fitting to obtain the adhesive force ratio η1=F1 / F, where F is the total pull-out force; in specific implementation, in the initial stage of the plug-in connection under stress, the slip is extremely small (≤0.05mm), and the chemical adhesive force at the interface between the steel bar and concrete dominates the bonding effect, while the frictional resistance and mechanical interlocking force can be ignored; the first step is to extract low-slip stage data: from the two-dimensional slip field grid data, filter the time and spatial location data corresponding to all slip s≤0.05mm, and simultaneously extract the total pull-out force F (taken from the axial pull-out force time series data) at the corresponding time. The second step is to construct a nonlinear mapping model: the bonding force exhibits a nonlinear decay characteristic as the slip increases. An existing exponential decay model is selected to construct the bonding force-slip mapping relationship: F1=a1·exp(-b1·s)+c1, where F1 is the bonding force, s is the slip, and a1, b1, and c1 are model fitting parameters. The third step is model fitting: the least squares method is used to fit the model parameters with the Fs data of the low slip stage to determine the bonding force mapping model F1=f1(s). The fourth step is to calculate the proportion of bonding force: the slip s at each moment of the low slip stage is substituted into the model to calculate the bonding force F1 at the corresponding moment. Combined with the total pull-out force F at the same moment, the proportion of bonding force η1=F1 / F is obtained. This step accurately quantifies the contribution of the bonding force in the low slip stage to the total adhesion force.

[0053] S62. Extract slip field data in the stable slip stage where the slip amount is 0.05 to 0.5 mm, and construct a friction resistance-confining pressure coupling mapping model: F2=f2(s,σ), where σ is the real-time confining pressure, and the proportion of friction resistance η2=F2 / F is obtained through fitting; in specific implementation, after the slip amount enters the range of 0.05 to 0.5 mm, the interfacial chemical cementation force is basically invalid, and the friction (friction resistance) at the interface between steel bar and concrete dominates the bonding effect, and the friction resistance has a coupling relationship with both the slip amount and the confining pressure; Step 1: Extract data of the stable slip stage: from the two-dimensional slip field grid data, screen the time data corresponding to 0.05 mm < s ≤ 0.5 mm, synchronously extract the total drawing force F and real-time confining pressure σ (obtained from the confining pressure time series data) at the corresponding time; Step 2: Construct a coupling mapping model: the friction resistance tends to be stable with the increase of slip amount and increases approximately linearly with the increase of confining pressure, so construct a binary nonlinear coupling model: F2=a2·s^b2·σ+c2, where F2 is the friction resistance, s is the slip amount, σ is the confining pressure, and a2, b2 and c2 are model fitting parameters; Step 3: Model fitting: adopt the multivariate nonlinear least square method to fit the model parameters with the F-s-σ data of the stable slip stage, and determine the friction resistance coupling mapping model F2=f2(s,σ); Step 4: Calculate the proportion of friction resistance: substitute s and σ at each moment in the stable slip stage into the model to calculate the friction resistance F2 at the corresponding moment, and obtain the proportion of friction resistance η2=F2 / F in combination with the total drawing force F; this step accurately quantifies the contribution proportion of friction resistance in the stable slip stage.

[0054] S63. Extract high-risk stage data where the crack propagation rate in the hook region is ≥0.1mm / s, and construct a mechanical interlocking force-crack rate mapping model: F3=f3(vc), where vc is the crack propagation rate. The mechanical interlocking force ratio η3=F3 / F is obtained by fitting the model. In specific implementation, after the slip exceeds 0.5mm, the mechanical interlocking effect of the hook on the concrete becomes the main source of bond force, accompanied by rapid propagation of concrete splitting cracks. The first step is to extract high-risk stage data: from the crack propagation path time series data, select the time data corresponding to the crack propagation rate vc≥0.1mm / s in the hook region (≤50mm from the hook apex), and simultaneously extract the total pull-out force F at the corresponding time. The second step is to construct a mapping model... Modeling: The mechanical interlocking force exhibits a linear growth characteristic with the crack propagation rate. A linear mapping model is constructed: F3 = a3·vc + b3, where F3 is the mechanical interlocking force, vc is the crack propagation rate, and a3 and b3 are model fitting parameters. Third step, model fitting: Using the univariate linear least squares method, the model parameters are fitted with F-vc data from the high-risk stage to determine the mechanical interlocking force mapping model F3 = f3(vc). Fourth step, calculating the proportion of mechanical interlocking force: Substituting vc at each moment of the high-risk stage into the model, the corresponding mechanical interlocking force F3 is calculated. Combined with the total pull-out force F, the proportion of mechanical interlocking force η3 = F3 / F is obtained. This step accurately quantifies the contribution proportion of mechanical interlocking force in the hook splitting stage.

[0055] S64. Normalize η1, η2, and η3 to satisfy η1 + η2 + η3 = 1, and output the quantitative results of the adhesive force components and their proportion time-series curves. In practice, due to local overlap in the calculation of adhesive force components at each stage, the sum of η1, η2, and η3 calculated directly may deviate from 1, so normalization correction is required. The first step is normalization calculation: For η1, η2, and η3 at each time point, normalize them according to the formulas η1' = η1 / (η1 + η2 + η3), η2' = η2 / (η1 + η2 + η3), and η3' = η3 / (η1 + η2 + η3). The first step is to correct and ensure that η1'+η2'+η3'=1; the second step is to output the quantitative results: output the corrected proportions of adhesive force η1', frictional resistance η2', and mechanical interlocking force η3' at each time point, forming a quantitative result table of adhesive force components; the third step is to generate the proportion time-series curves: plot the time-series curves of η1', η2', and η3' as a function of time according to the test time sequence, intuitively presenting the dynamic evolution law of the proportion of each adhesive force component at different stress stages; this step achieves accurate normalization and quantification of the proportion of adhesive force components, and clarifies the dynamic changes of the adhesive force contribution mechanism throughout the entire process of plug-in connection.

[0056] In this embodiment, a further implementation is that S7, based on the slip mutation characteristics and crack evolution law, inverts the effective anchorage stiffness and critical anchorage length and outputs visualization results, including: S71, calculating the slip stiffness K(t)=dF / ds based on the dynamic evolution data of the slip field, and extracting the slip amount when the slip stiffness decays to 50% of the initial value as the basis for determining the effective anchorage stiffness; in specific implementation, the slip stiffness characterizes the ability of the plug-in connection to resist slip deformation and is the core indicator for evaluating anchorage performance; the first step is to calculate the slip stiffness: from the axial pull-out force time series data and the slip field grid data, extract the pull-out force F and slip amount s time series data at the midpoint of the overlap length, and sort them in time order. The process involves four steps: First, calculate the slip stiffness K(t) = dF / ds (the first derivative of the pull-out force with respect to the slip amount) to generate time-series data of the slip stiffness. Second, determine the initial slip stiffness: take the average slip stiffness at the beginning of the test (slip amount ≤ 0.02 mm) as the initial slip stiffness K0. Third, extract the attenuation feature point: traverse the time-series data of the slip stiffness to find the time and slip amount s50 corresponding to the slip stiffness K(t) = 0.5K0. This point is the critical feature point of the anchoring stiffness attenuation. Fourth, determine the effective anchoring stiffness: define the slip stiffness value corresponding to s50 as the effective anchoring stiffness Ke. The larger the effective anchoring stiffness, the stronger the resistance to slippage of the plug-in connection and the better the anchoring performance.

[0057] S72. Using the lap length corresponding to the slip mutation point as a benchmark, and combining slip field distribution data for extrapolation, determine the minimum anchorage length when the steel stress reaches the yield strength, and output the critical anchorage length value. In specific implementation, the critical anchorage length is the minimum lap length that ensures no bond failure occurs before the steel yields, and is a key parameter in engineering design. The first step is to determine the location of the slip mutation point: extract the spatial location corresponding to the earliest slip mutation point from the slip mutation points marked in S43. This location is the starting position of steel yielding, corresponding to the lap length L0. The second step is to extrapolate the slip field: based on the slip distribution law near L0 in the slip field grid data, use the existing linear extrapolation algorithm to extrapolate the slip distribution along the lap length direction to determine the minimum lap length when the steel stress reaches the yield strength (corresponding to the slip mutation characteristic). The third step is to output the critical anchorage length: define the minimum lap length obtained by extrapolation as the critical anchorage length Lcr. The smaller Lcr is, the higher the anchorage efficiency of the plug-in connection and the more economical the structural design.

[0058] S73. Generate effective anchorage stiffness-confining pressure relationship curves, critical anchorage length-slip relationship curves, and corresponding visualization charts. In specific implementation, to intuitively present the correlation between anchorage performance and confining pressure and slip, multi-dimensional visualization charts are generated. The first step is to generate relationship curves: based on effective anchorage stiffness data under different confining pressure levels, draw effective anchorage stiffness-confining pressure relationship curves to reveal the influence of confining pressure on anchorage stiffness; based on critical anchorage length data under different slip, draw critical anchorage length-slip relationship curves to clarify the influence of slip deformation on critical anchorage length. The second step is to generate visualization charts: generate a summary table of anchorage performance parameters, a slip field thermal atlas, crack propagation trajectory animation, and a time series diagram of the proportion of bonding components, etc. The third step is to output an analysis report: integrate all quantitative data and visualization charts to generate a plug-in connection anchorage performance analysis report, providing comprehensive, intuitive, and quantitative technical support for engineering design, construction optimization, and safety assessment.

[0059] In this embodiment, a further implementation is that S11 further includes: S111, analyzing the segmented confining pressure parameters to obtain the confining pressure gradient sequence {σ1,σ2,…,σm} along the height direction of the specimen, where m is the number of segments; specifically, the segmented confining pressure loading control command includes m independent loading segments (m=3~5) divided along the height direction of the specimen, each loading segment corresponding to a preset target confining pressure value; after the system analyzes the command, it extracts the target confining pressure values ​​of each loading segment, arranges them in order from bottom to top according to the height of the specimen, forming the confining pressure gradient sequence {σ1,σ2,…,σm}, where σ1 is the confining pressure of the bottom loading segment and σm is the confining pressure of the top loading segment. The confining pressure values ​​in the sequence can be the same or different, used to simulate the confining pressure differences at different depths of the bridge tower.

[0060] S112. Bind the confining pressure gradient sequence to a global timestamp to generate a non-uniform gradient confining pressure time-series signal to simulate the distribution difference of confining pressure with height within a depth range of 5 to 10 meters in the bridge tower. In specific implementation, bind each confining pressure parameter in the confining pressure gradient sequence to a global unified timestamp, and assign independent timing control instructions for the confining pressure loading process of each loading segment. Based on the bound parameters, the loading drive module generates a non-uniform gradient confining pressure time-series signal, which contains the loading curve of confining pressure with time for each height segment, to realize non-uniform gradient loading with high confining pressure at the bottom and low confining pressure at the top of the specimen, and accurately simulate the actual distribution characteristics of concrete confining pressure decreasing with height within a depth range of 5 meters (bottom) to 10 meters (top) in the bridge tower. S113. Establish a confining pressure gradient-multi-source data association mapping to match the slip field data and crack evolution data for each segment of confining pressure parameters. In specific implementation, based on the time-series index mapping table, the time-series data of confining pressure parameters for each height segment are associated and bound with the slip field grid data and crack evolution coordinate data of the corresponding time and spatial region, establishing a three-dimensional association mapping relationship between confining pressure gradient, slip field, and crack evolution. Through this mapping relationship, the slip distribution characteristics and crack evolution law of the corresponding region under any confining pressure gradient can be accurately queried, clarifying the influence mechanism of non-uniform confining pressure on the slip deformation and splitting damage of the plug-in connection, and providing accurate data association support for the analysis of confining pressure effects.

[0061] Example 2: A multi-channel non-contact mechanical loading and measurement system for pull-out tests of U-shaped stirrup connections is provided. This system enables multi-channel non-contact mechanical loading and measurement of U-shaped stirrup connections. The system includes a timing reference construction module, used to generate a high-precision global synchronization clock, dynamically correct channel timing deviations, construct a timing index mapping table, and generate a timing synchronization verification log. Specifically, it executes all operations S1, S11, S12, S13, and S14, generating a global synchronization clock with a resolution ≥1μs, real-time monitoring and correction of multi-channel timing deviations to ≤0.5ms, constructing a global timestamp index mapping table, and recording timing correction logs. This provides a unified timing reference for multi-source data across the entire system, ensuring data timing consistency and traceability. The multi-source data synchronous acquisition module is used to parse confining pressure and pull-out control commands, synchronously acquire confining pressure / pull-out force timing data, acquire distributed fiber frequency shift data at 100Hz, and acquire multi-channel acoustic emission raw waveform data at 1MHz. Specifically, it is used to execute all operations of S2, S21, S22, S23, and S24, parse segmented confining pressure and axial pull-out loading commands, drive the confining pressure and pull-out loading execution unit, synchronously acquire confining pressure and pull-out force mechanical data, acquire fiber frequency shift data at 100Hz, and acquire acoustic emission raw waveform data at 1MHz, realizing synchronous linkage of mechanical loading, fiber sensing, and acoustic emission monitoring, and outputting complete multi-source raw timing data. The frequency domain decomposition and feature extraction module is used for wavelet threshold denoising, wavelength division multiplexing decoupling, Brillouin frequency shift-strain conversion, time series alignment, and outlier removal. Specifically, it executes all operations S3, S31, S32, S33, and S34, performs wavelet denoising and wavelength division multiplexing decoupling on fiber frequency shift data, separates the frequency shift components of steel bars and concrete, converts them into strain data through the Brillouin model, completes time series alignment, interpolation completion, and outlier removal, and outputs high-quality steel bar and concrete strain time series data, providing reliable input for slip field reconstruction. The slip field spatiotemporal reconstruction module is used to execute the slip field spatiotemporal reconstruction algorithm, generate the slip amount time series matrix, perform kriging spatial interpolation, identify slip abrupt change points, and output slip field dynamic evolution data. Specifically, it executes all operations of S4, S41, S42, S43, and S44, runs the slip field spatiotemporal reconstruction algorithm, generates an N×T slip amount matrix, generates 1mm×10ms resolution slip field grid data through kriging interpolation, identifies abrupt change points of steel bar yield slip, and outputs slip field thermograms, slip-time curves, and abrupt change characteristic parameters, realizing full-domain, high-precision, and continuous measurement of the slip field. The acoustic emission crack location analysis module is used for acoustic emission waveform filtering and amplification, effective event extraction, three-dimensional time difference localization, crack propagation path association, and intelligent crack type identification. Specifically, it executes all operations of S5, S51, S52, S53, S54, and S55, filters and amplifies the acoustic emission waveform, extracts effective acoustic emission events, solves the three-dimensional coordinates of the crack through a three-dimensional time difference localization algorithm, generates the crack propagation path by associating the coordinates of the hook arc apex, identifies the crack type through a machine learning classifier, and outputs crack coordinates, path, and type distribution data to achieve dynamic monitoring of internal damage in concrete. The adhesive force component decomposition module is used to construct a mapping model of adhesive force / frictional resistance / mechanical interlocking force, normalize the component proportions, and output the quantitative results of adhesive force. Specifically, it executes all operations S6, S61, S62, S63, and S64, constructs mapping models of adhesive force-slip, frictional resistance-confining pressure, and mechanical interlocking force-crack rate, fits and calculates the proportion of each component, and outputs the quantitative results of adhesive force components and the proportion time series curves after normalization and correction, thus achieving accurate decomposition and quantification of adhesive force components. The anchorage performance inversion and visualization module is used to calculate the slip stiffness attenuation coefficient, invert the effective anchorage stiffness and critical anchorage length, and generate various relationship curves and visualization charts. Specifically, it is used to execute all operations of S7, S71, S72, S73, S111, S112, and S113, calculate the slip stiffness time series data, invert the effective anchorage stiffness and critical anchorage length, generate anchorage stiffness-confining pressure and critical anchorage length-slip relationship curves and visualization charts, complete the confining pressure gradient correlation mapping, output the anchorage performance analysis report, and realize the quantitative inversion and visualization of test results.

[0062] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A multi-channel non-contact mechanical loading and measurement method for pull-out tests of U-shaped stirrup connections, characterized in that, The method includes: S1. Establish a unified time series reference for multi-source heterogeneous data, and assign a globally unified timestamp to confining pressure data, axial pull-out force data, distributed optical fiber frequency shift data, and acoustic emission signal data. S2. Synchronously acquire confining pressure time-series data, axial pull-out force time-series data, distributed fiber frequency shift time-series data, and multi-channel acoustic emission raw waveform time-series data of the plug-in connected specimen during the controlled confining pressure loading and axial pull-out loading process. S3. Perform wavelet threshold denoising, wavelength division multiplexing decoupling and frequency shift-strain conversion on the distributed optical fiber frequency shift time series data, and separate and output the steel bar strain time series data and the concrete strain time series data. S4. Based on the time series data of steel bar strain and concrete strain, a two-dimensional slip field distribution data that is continuous in time and spans the entire domain along the lap length of the steel bar is generated by using a slip field spatiotemporal reconstruction algorithm. S5. Perform filtering, amplification, event extraction, three-dimensional time difference localization and feature classification on the original acoustic emission waveform time series data to obtain three-dimensional coordinate time series data of microcracks inside concrete, crack propagation path time series data and crack type distribution data. S6. By integrating confining pressure time series data, axial pull-out force time series data, slip field distribution data, and crack evolution correlation data, a bonding force component decomposition model is constructed to quantitatively output the proportion of cementing force, frictional resistance, and mechanical interlocking force. S7. Based on the slip mutation characteristics and crack evolution law, invert the effective anchorage stiffness and critical anchorage length and output the visualization results.

2. The method according to claim 1, characterized in that, The S1 establishes a unified time series benchmark for multi-source heterogeneous data, including: S11. Generate a high-precision global synchronization clock signal with a clock resolution of not less than 1μs, and assign a unified timestamp sequence to the four data channels of confining pressure, pull-out force, fiber frequency shift and acoustic emission. S12. Monitor the local clock deviation of each data channel in real time. When the timing deviation between channels exceeds 1ms, trigger dynamic timing correction to correct the deviation to ≤0.5ms. S13. Construct a time-series index mapping table, using the global timestamp as the primary key, and establish a one-to-one correspondence between the confining pressure data index, pull-out force data index, fiber optic frequency shift data index, and acoustic emission data index. S14. Generate a time-series synchronization verification log, recording the time point, deviation value, and correction amount of each correction, for data traceability and consistency verification.

3. The method according to claim 1, characterized in that, The S2 synchronously acquires multi-source time-series data, including: S21. Receive the confining pressure loading control command, parse the segmented confining pressure parameter sequence, generate the corresponding confining pressure loading drive signal, and synchronously collect the real-time confining pressure value output by the confining pressure sensor to form confining pressure time series data. S22. Receive the axial pull-out loading control command, parse the loading rate and target load parameters, generate the pull-out loading drive signal, and synchronously collect the real-time pull-out force value output by the force sensor to form axial pull-out force time sequence data. S23. Configure the distributed optical fiber demodulator to scan at a frequency of 100Hz, and collect Brillouin frequency shift signals of the optical fiber embedded in the steel bar and the optical fiber pre-embedded in the concrete in parallel. Separate the two frequency shift data according to the optical fiber wavelength label to form distributed optical fiber frequency shift time series data. S24. Configure the acoustic emission acquisition system with a sampling rate of 1MHz, and synchronously acquire the original waveforms output by no less than 4 acoustic emission sensors, retaining the amplitude, duration, rise time, and energy characteristic parameters of the waveforms to form multi-channel acoustic emission original waveform timing data.

4. The method according to claim 1, characterized in that, The S3 frequency domain decomposition and feature extraction include: S31. Perform wavelet multi-scale decomposition on the distributed optical fiber frequency shift time series data, select 3 to 5 levels of decomposition scale, use a soft threshold function to suppress high-frequency noise components, and retain effective frequency shift signal components. S32. Based on the wavelength division multiplexing decoupling algorithm, a frequency domain filter bank is constructed according to the difference in center wavelength between the fiber embedded in the steel bar and the fiber pre-embedded in the concrete to separate the frequency shift time-series component of the fiber embedded in the steel bar and the fiber pre-embedded in the concrete. S33. Based on the Brillouin frequency shift-strain conversion model: ε=k·Δν, where ε is strain, Δν is the Brillouin frequency shift, and k is the calibration coefficient, the two frequency shift components are converted into steel strain time series data and concrete strain time series data respectively. S34. Time-series alignment of steel bar strain time-series data and concrete strain time-series data is performed. Data interpolation is completed based on the time-series index mapping table, and outlier values ​​in the strain data are removed using the 3σ criterion.

5. The method according to claim 1, characterized in that, The S4 slip field spatiotemporal reconstruction algorithm includes: discretizing the lap length of the reinforcing bar into N equidistant measuring points, where the reinforcing bar strain at measuring point i (i=1,2,…,N) at time t is denoted as εs(i,t) and the concrete strain is denoted as εc(i,t). The slip amount s(i,t) of the measuring point is calculated according to the following formula: Where τ is the integration time variable, and α is the spatial smoothing coefficient, ranging from 0.01 to 0.

1. It is a two-dimensional Laplace operator; S41. Substitute the values ​​of each measurement point and each time step into the formula to calculate the slip s(i,t), and generate an N×T slip time series matrix, where T is the total number of sampling times; S42. Perform Kriging space interpolation on the slip time series matrix to supplement the slip data between measurement points and generate two-dimensional slip field grid data with a resolution of not less than 1mm×10ms. S43. Traverse the slip field grid data, identify the moment when the slip abrupt change exceeds the preset threshold, and mark it as the slip abrupt change point at the moment of steel bar yielding; S44. Generate a dynamic evolution sequence based on two-dimensional slip field grid data, and output slip field thermogram, slip-time curve and slip mutation characteristic parameters.

6. The method according to claim 1, characterized in that, The S5 acoustic emission crack localization and evolution analysis includes: S51. Perform 10kHz to 1MHz bandpass filtering and 40dB gain amplification on the original waveform timing data of multi-channel acoustic emission to eliminate low-frequency environmental noise and high-frequency electromagnetic interference. S52. Set the acoustic emission event trigger threshold, extract valid acoustic emission events with amplitudes exceeding the threshold, and record the trigger time of the event, the arrival time difference of each channel, and the characteristic parameters. S53. Construct a three-dimensional time difference positioning equation set based on the multi-channel arrival time difference, and solve iteratively using the least squares method to obtain the three-dimensional spatial coordinates (x, y, z) of the microcrack. S54. Correlate the three-dimensional coordinates of the microcrack with the geometric coordinates of the hook apex according to the time series, calculate the time series data of the distance between the crack and the hook apex, and generate the time series data of the crack propagation path. S55. Construct a crack type identification classifier, using the amplitude, duration, rise time, and energy of the acoustic emission signal as input features, to distinguish between tension cracks, shear cracks, and interface peeling cracks, and output crack type distribution data.

7. The method according to claim 1, characterized in that, The S6 adhesive component decomposition model includes: S61. Extract the slip field data of the low slip stage with slip ≤ 0.05 mm, construct the nonlinear mapping model of cementing force-slip: F1=f1(s), and fit it to obtain the cementing force ratio η1=F1 / F, where F is the total pull-out force; S62. Extract slip field data for the stable slip stage with slip of 0.05 to 0.5 mm, and construct a frictional resistance-confining pressure coupled mapping model: F2=f2(s,σ), where σ is the real-time confining pressure. Fit the model to obtain the frictional resistance ratio η2=F2 / F. S63. Extract high-risk stage data with crack propagation rate ≥ 0.1 mm / s in the hook area, and construct a mechanical interlocking force-crack rate mapping model: F3 = f3(vc), where vc is the crack propagation rate. Fit the model to obtain the mechanical interlocking force ratio η3 = F3 / F. S64. Normalize and correct η1, η2, and η3 to satisfy η1+η2+η3=1, and output the quantitative results of the adhesive force components and the time-series curve of their proportions.

8. The method according to claim 2, characterized in that, S11 further includes: S111. Analyze the segmented confining pressure parameters to obtain the confining pressure gradient sequence {σ1,σ2,…,σm} along the height direction of the specimen, where m is the number of segments; S112. Bind the confining pressure gradient sequence with a global timestamp to generate a non-uniform gradient confining pressure time series signal to simulate the distribution difference of confining pressure with height within a depth range of 5 to 10 meters in the bridge tower. S113. Establish a mapping between confining pressure gradient and multi-source data so that each segment of confining pressure parameter corresponds to matching slip field data and crack evolution data.

9. The method according to claim 5, characterized in that, S44 further includes: S441. Calculate the slip stiffness K(t)=dF / ds based on the dynamic evolution data of the slip field, and extract the slip amount when the slip stiffness decays to 50% of the initial value as the basis for determining the effective anchoring stiffness. S442. Using the lap length corresponding to the slip change point as a benchmark, and extrapolating the slip field distribution data, determine the minimum anchorage length when the steel stress reaches the yield strength, and output the critical anchorage length value. S443. Generate the effective anchorage stiffness-confining pressure relationship curve, the critical anchorage length-slippage relationship curve, and the corresponding visualization charts.

10. A multi-channel non-contact mechanical loading and measurement system for pull-out testing of U-shaped stirrup connections, characterized in that, To implement the method according to any one of claims 1 to 9, comprising: The timing reference construction module is used to generate a high-precision global synchronization clock, dynamically correct channel timing deviations, build a timing index mapping table, and generate timing synchronization verification logs. The multi-source data synchronous acquisition module is used to parse confining pressure and pull-out control commands, synchronously acquire confining pressure / pull-out force timing data, acquire distributed fiber optic frequency shift data at 100Hz, and acquire multi-channel acoustic emission raw waveform data at 1MHz. The frequency domain decomposition and feature extraction module is used for wavelet thresholding noise reduction, wavelength division multiplexing decoupling, Brillouin frequency shift-strain conversion, time alignment and outlier removal; The slip field spatiotemporal reconstruction module is used to execute the slip field spatiotemporal reconstruction algorithm, generate the slip amount time series matrix, perform kriging space interpolation, identify slip abrupt change points, and output slip field dynamic evolution data. The acoustic emission crack location analysis module is used for acoustic emission waveform filtering and amplification, effective event extraction, three-dimensional time difference location, crack propagation path association, and intelligent crack type identification. The adhesive force component decomposition module is used to construct a mapping model of adhesive force / frictional resistance / mechanical interlocking force, normalize the component proportions, and output the adhesive force quantification results. The anchorage performance inversion and visualization module is used to calculate the slip stiffness attenuation coefficient, invert the effective anchorage stiffness and critical anchorage length, and generate various relationship curves and visualization charts.

Images