Service life prediction method and system for bridge connection part

Abstract

The invention relates to the technical field of bridge engineering structure health monitoring, in particular to a service life prediction method and system for a bridge connection part, and the method comprises the steps: collecting the strain of a grouting sleeve, a reinforcement-grouting interface ultrasonic reflection signal and environment temperature and humidity data through a distributed sensor array; a three-dimensional dynamic void rate model is generated through fusion of a spatial topological mapping algorithm, a correlation function of load circulation and void expansion is established, the dynamic void rate is input into a bond strength coupling degradation model, after strength attenuation is quantized in combination with steel bar corrosion data and bond failure early warning is triggered, stress redistribution is calculated through multi-physics coupling simulation, and the dynamic void rate is obtained. Early warning is taken as a starting node, the actual load spectrum and temperature and humidity data are combined, Monte Carlo sampling is adopted to simulate rigidity degradation and output the residual life probability, the system comprises a multi-source collaborative sensing unit, a micro-void evolution analysis unit and a life prediction decision unit, and the safety control precision of bridge connection parts is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bridge engineering structure health monitoring, in particular to a life prediction method and system for bridge connecting parts. BACKGROUND

[0002] The bridge engineering structure health monitoring technology is an important technology. Under the background of the continuous extension of the service life of bridges and the increasing complexity of traffic loads, this technology is the key support for ensuring the safety of bridge structures and avoiding the sudden failure of connecting parts. It can not only capture the early damage of micro voids and stress concentration at the interface between the grouting sleeve and the steel bar in real time, but also quantify the damage accumulation effect combined with load cycles and environmental factors, providing accurate life evaluation basis for bridge maintenance, reducing the risk of bridge collapse caused by the failure of connecting parts, avoiding resource waste caused by excessive maintenance, promoting the upgrade of bridge operation and maintenance from post-maintenance to pre-warning, and adapting to the health management needs of connecting parts of various bridges of highways and railways. The existing bridge connecting part life prediction technology faces the core problems of one-sided damage identification, inaccurate strength degradation evaluation, and low life prediction reliability in actual application. Traditional technology relies on a single type of sensor and does not build a monitoring network that integrates strain, ultrasonic, and environmental temperature and humidity multidimensional data. It cannot accurately locate the initiation position of micro voids, which are often accompanied by stress concentration. Single data cannot distinguish between normal deformation and damage deformation, and cannot quantify the joint weakening effect of the coupling of micro voids and steel corrosion on the gripping strength. This technical limitation will have a chain effect: Firstly, without establishing a three-dimensional dynamic void rate model through multi-source data fusion, it can only roughly determine whether voids exist, and cannot master the correlation between void expansion rate and load cycle, leading to early damage being missed or misjudged. Secondly, the gripping strength degradation model does not include the coupling mechanism of voids and corrosion. When both act together, the decay rate of gripping force will be underestimated, and the gripping failure warning will be delayed. Finally, life prediction is based on a single load condition and deterministic analysis, without considering the influence of actual load spectrum and environmental temperature and humidity on stiffness degradation, resulting in a large deviation between predicted results and actual service state. Either the maintenance opportunity is delayed due to overestimated life, or excessive maintenance is caused by underestimating life. In the long run, it is difficult to ensure the structural safety of bridge connecting parts, and it also increases the operation and maintenance cost, which cannot meet the needs of precise control of the whole life cycle of bridges. In order to solve this technical problem, we provide a life prediction method and system for bridge connecting parts. SUMMARY

[0003] The purpose of the present application is to provide a life prediction method and system for bridge connecting parts to solve the problems raised in the background.

[0004] 1. Due to the one-sided nature of damage identification and the lack of integration of multi-dimensional data, it is difficult to locate micro-voids and quantify their expansion. Therefore, this case study uses a distributed sensor array to collect multiple types of data and employs a spatial topology mapping algorithm to integrate the data and build a three-dimensional dynamic void rate model, which can accurately identify damage and grasp the expansion pattern.

[0005] 2. Due to the low reliability of life prediction, the lack of coupling of multiple damage factors and the absence of consideration for actual working conditions, this case study uses a grip strength coupled degradation model to calculate strength attenuation and employs multiphysics simulation and Monte Carlo sampling to simulate stiffness degradation, which can improve the reliability of life prediction.

[0006] To achieve the above objectives, one objective of this invention is to provide a method for predicting the lifespan of bridge connection parts, comprising the following steps: S1. Collect strain data of the outer wall of the sleeve, ultrasonic reflection signal of the steel reinforcement grouting interface and ambient temperature and humidity data by a distributed sensor array installed at the stress part of the grouting sleeve, in order to build a multi-dimensional monitoring network. S2. Based on the ultrasonic reflection signal of the steel bar grouting interface, the interface acoustic energy loss coefficient is extracted to quantify the degree of bonding defects in the micro-void area. Combined with the strain data of the outer wall of the sleeve, the local stress concentration threshold area is identified to locate the micro-void initiation location. Finally, the discrete monitoring point data of the multi-dimensional monitoring network is fused through the spatial topology mapping algorithm to generate a three-dimensional dynamic void rate model of the steel bar-sleeve interface, output the dynamic void rate, and establish a correlation function between the number of load cycles and the void expansion rate to characterize the damage accumulation effect under cyclic loading. S3. Input the dynamic void rate into the preset bond strength coupling degradation model and set the critical threshold of the void rate. When the dynamic void rate exceeds the critical threshold, trigger the bond force failure warning. Integrate the concrete crack width and steel corrosion data, and calculate the stress redistribution of the connection part caused by bond failure through multi-physics field coupling simulation to quantify the structural internal force transfer caused by force transmission failure. S4. Taking the gripping force failure warning as the starting point for life prediction, the remaining stiffness of the connection part is inferred based on stress redistribution, and a time-varying reliability function is constructed by combining the actual load spectrum and environmental temperature and humidity data to calculate the remaining life probability of the bridge connection part.

[0007] The second objective of this invention is to provide a system for implementing the life prediction method for bridge connection parts, including any one of the above-mentioned methods, comprising: The multi-source collaborative sensing unit deploys four sets of composite sensing units in a circumferential manner at the junction of the curved and straight sections of the grouting sleeve. Each set of units integrates an adaptive focusing ultrasonic probe, a fiber optic strain gauge, and a temperature and humidity compensator. The sleeve strain, ultrasonic reflection signal at the rebar-grouting interface, and environmental temperature and humidity data are transmitted to the micro-void evolution analysis unit in real time via an industrial bus. The micro-loosening evolution analysis unit is internally provided with a sound energy-strain double threshold diagnosis sub-module, a three-dimensional dynamic loosening rate modeling sub-module and a gripping failure early warning sub-module, and the loosening quantitative result and early warning instruction are synchronously output to the life prediction decision unit; The life prediction decision unit receives the gripping force failure early warning signal and loosening evolution data of the micro-loosening evolution analysis unit, calculates the remaining stiffness of the connecting part through a multi-physical field coupling simulator, and drives a time-varying reliability analyzer to finally generate a bridge connecting part life evaluation report.

[0008] Compared with the prior art, the present application has the following beneficial effects: 1. The sleeve strain, interface ultrasonic reflection signal and environmental temperature and humidity data are collected by the distributed sensor array, the discrete monitoring point data are fused by combining the spatial topological mapping algorithm, the steel sleeve interface three-dimensional dynamic loosening rate model is generated, the correlation function of load cycle and loosening expansion is established, the technical effect of accurately identifying the micro-loosening initiation position and quantifying the expansion law is achieved, the problems of one-sided damage identification and difficulty in capturing early damage in the traditional technology are solved, and the advantages of early detection of connecting part hidden dangers and reliable data support for subsequent evaluation are provided.

[0009] 2. The gripping strength coupling degradation model is constructed, the dynamic loosening rate and steel corrosion data are integrated, the ribbing occlusion failure mechanism and corrosion expansion pressure model are embedded, the stress redistribution after gripping failure is calculated by combining the multi-physical field coupling simulation, the technical effect of accurately quantifying the gripping strength attenuation is achieved, the problems of not considering multi-damage coupling and inaccurate strength evaluation in the traditional model are solved, and the advantages of timely triggering the gripping failure early warning and avoiding early warning lag are provided.

[0010] 3. The gripping failure early warning is taken as the starting node, the actual load spectrum and environmental temperature and humidity data are combined, the stiffness degradation path under ten thousand times of load is simulated by using the Monte Carlo sampling technology, the residual life probability density curve is output, the technical effect of improving the reliability of life prediction is achieved, the problems of ignoring the working condition difference and large result deviation in the traditional prediction are solved, and the advantages of providing scientific basis for maintenance decision and balancing safety and operation cost are provided. DETAILED DESCRIPTION

[0011] Figure 1 The overall workflow diagram of the present application is shown in the figure; Figure 2 The overall structure schematic diagram of the present application is shown in the figure; The meanings of various labels in the figure are as follows: 1. Multi-source collaborative perception unit; 2. Micro-loosening evolution analysis unit; 3. Life prediction decision unit. DETAILED DESCRIPTION

[0012] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work belong to the protection scope of the present application.

[0013] Please refer to Figure 1 The present embodiment aims to provide a life prediction method for a bridge connection part, which comprises the following steps: S1, collecting sleeve outer wall strain data, steel grouting interface ultrasonic reflection signals and environmental temperature and humidity data by a distributed sensor array installed at the stress part of the grouting sleeve, for constructing a multidimensional monitoring network; S2, extracting interface acoustic energy loss coefficient based on the steel grouting interface ultrasonic reflection signals, for quantifying the bonding defect degree of the micro void area, and identifying the local stress concentration threshold area in combination with the sleeve outer wall strain data, for positioning the micro void initiation position, finally fusing the discrete monitoring point data of the multidimensional monitoring network through a spatial topology mapping algorithm to generate a three-dimensional dynamic void rate model of the steel-sleeve interface, outputting the dynamic void rate, and establishing a load cycle number-void expansion rate correlation function, for representing the damage accumulation effect under cyclic loading; S3, inputting the dynamic void rate into a preset bond strength coupling degradation model, and setting a void rate critical threshold, when the dynamic void rate exceeds the critical threshold, triggering a bond force failure warning, and fusing concrete crack width and steel corrosion data, through multi-physical field coupling simulation to calculate the stress redistribution of the connection part caused by bond failure, for quantifying the internal force transfer caused by force transmission failure; S4, taking the bond force failure warning as the life prediction starting node, backstepping the connection part residual stiffness based on the stress redistribution, and constructing a time-varying reliability function in combination with the actual load spectrum and environmental temperature and humidity data, for calculating the residual life probability of the bridge connection part.

[0014] The construction method of the multidimensional monitoring network comprises the following specific steps: Four groups of strain-ultrasonic composite sensing units are arranged circumferentially at the junction of the curved section and the straight section of the grouting sleeve, each group of units comprising a contact fiber Bragg grating strain gauge for capturing the multi-axial deformation of the sleeve outer wall and a self-adaptive focusing ultrasonic probe, wherein the ultrasonic probe penetrates the sleeve metal wall at an oblique incidence angle, the transmission frequency and the focusing depth are adjusted to match the steel rib pitch, and a temperature and humidity compensation module is integrated in each sensing unit to eliminate the interference of environmental temperature and humidity fluctuations on the ultrasonic signal propagation speed, finally forming a multidimensional monitoring network covering the full circumference of the sleeve.

[0015] The extraction method of the interface acoustic energy loss coefficient, and the specific steps are as follows: The ultrasonic reflection signals of the steel bar grouting interface are analyzed in time-frequency domain, the interface reflection wave and the multiple reflection noise of the sleeve wall in the signals are separated through wavelet packet decomposition technology, the energy attenuation rate of the characteristic frequency band is calculated, the acoustic energy loss coefficient is defined as the ratio of the defect reflection wave energy to the theoretical full-bond reflection wave energy, and finally the region with the acoustic energy loss coefficient greater than the preset reference value is determined as the micro void region.

[0016] The identification method of the local stress concentration threshold region, and the specific steps are as follows: The time-space distribution characteristics of the sleeve outer wall strain data are synchronously analyzed, when the reverse strain mutation of the adjacent strain gauges is detected and the mutation amplitude continuously increases, it is determined as a local stress concentration region, the acoustic energy-strain double threshold diagnosis mechanism is established combining with the spatial distribution of the acoustic energy loss coefficient, that is, in the acoustic energy loss coefficient over-limit region, if the strain mutation amplitude is detected to be greater than the set proportion of the material yield threshold at the same time, the region is positioned as the micro void initiation core region, and is marked as the key update node of the three-dimensional void ratio model.

[0017] The execution method of the spatial topological mapping algorithm, and the specific steps are as follows: The monitoring points of the multi-dimensional monitoring network are taken as the space control nodes, the initial topological grid of the sleeve-steel bar interface is constructed based on the Delaunay triangulation principle, the acoustic energy loss coefficient and the strain mutation data of the discrete nodes are converted into continuous space fields through the introduction of the radial basis function interpolation model, and the dynamic grid refinement technology is adopted to automatically encrypt the grid nodes in the micro void initiation core region, so as to realize the reconstruction of the three-dimensional data field.

[0018] It needs to be further explained that in the life prediction of the bridge connecting part, the construction of the accurate multi-dimensional monitoring network is the basis for obtaining the core data, the grouting sleeve is the key component of the steel bar connection, and the bending section and the straight section are the most prone to micro voids and stress concentration damage due to complex stress, superposition of bending stress and axial stress, so the sensing unit needs to be preferentially deployed at the intersection, and the specific implementation is as follows: The construction method of the multi-dimensional monitoring network involves the following steps: First, four sets of strain-ultrasonic composite sensing units are circumferentially arranged at the junction of the curved and straight sections of the grouting sleeve. The junction refers to the area in the grouting sleeve where the straight extension section transitions to the curved section. Due to the abrupt change in geometry, this area is prone to stress concentration under load, making it a high-risk location for micro-voids, i.e., tiny gaps at the interface between the reinforcing steel and the grouting material. Circumferential arrangement means that the sensing units are evenly distributed along the circumference of the sleeve. The four sets of units are located at 0°, 90°, 180°, and 270° positions on the sleeve, respectively, ensuring coverage. The sleeve is designed to cover the entire circumference of the casing to avoid damage due to blind spots in monitoring. Each unit contains two core components: a contact fiber optic strain gauge, which is tightly attached to the outer wall of the sleeve with a high-strength adhesive and can capture the multi-axial deformation of the sleeve under load, including axial tension / compression and circumferential expansion / contraction, reflecting the stress distribution on the outer wall of the sleeve in real time; and an adaptive focusing ultrasonic probe, which is installed at a 45° angle to the outer wall of the sleeve. The tilted incident angle ensures that the ultrasonic waves penetrate the metal wall and are perpendicularly incident on the rebar-grout interface, improving detection accuracy. Its core advantage is that it can adjust the transmission frequency (…). The frequency range is 2-5MHz, selected based on the sleeve thickness (2MHz for thick-walled sleeves and 5MHz for thin-walled sleeves). The focusing depth ensures the focal point accurately falls on the rebar surface, matching the rebar rib spacing to ensure clear ultrasonic detection of the interface bonding state at the ribs, avoiding blind spots caused by focusing deviations. Simultaneously, a temperature and humidity compensation module is integrated into each sensing unit to eliminate interference from environmental temperature and humidity fluctuations on the ultrasonic signal propagation speed. Temperature and humidity changes in the bridge's service environment cause variations in the propagation speed of ultrasonic waves in the metal sleeve and air medium; without compensation, this directly affects the interface detection accuracy. The temperature and humidity compensation module incorporates a high-precision temperature and humidity sensor to collect environmental data in real time. It then corrects the speed of the original signal collected by the ultrasonic probe using a preset temperature and humidity-sound velocity correction formula. Finally, the four composite sensing units transmit the collected sleeve strain data, rebar-grout interface ultrasonic reflection signals, and environmental temperature and humidity data to the backend analysis unit in real time via an industrial Ethernet bus. The monitoring ranges of the four units overlap, forming a multi-dimensional monitoring network that coordinates strain, ultrasound, and environmental parameters, covering the entire circumference of the sleeve, providing comprehensive data support for subsequent damage identification. After acquiring ultrasonic reflection signals through a multi-dimensional monitoring network, the interface acoustic energy loss coefficient needs to be extracted. This coefficient is a core indicator for quantifying the degree of micro-void at the interface between the steel reinforcement and the grout. Micro-void leads to changes in ultrasonic reflection energy. By analyzing these changes, the defect area can be accurately located. The specific implementation method is as follows: The specific steps of the extraction method of the interface acoustic energy loss coefficient are as follows: first, the ultrasonic reflection signal of the steel grouting interface is analyzed in time-frequency domain, the ultrasonic reflection signal is the electric signal reflected back after the ultrasonic wave emitted by the ultrasonic probe penetrates the sleeve wall and reaches the steel grouting interface, which is expressed as a voltage fluctuation curve changing with time, the time-frequency domain analysis is to analyze the signal from the time dimension (the change law of the signal with time) and the frequency dimension (the frequency component contained in the signal) at the same time, which avoids the limitation of single dimension analysis, and the specific process is as follows: The original ultrasonic reflection signal is divided into several segments with 10 μs as the time interval, each segment contains 1024 data points, for each segment, on the one hand, the peak amplitude (reflecting the strength of the reflected wave) and the arrival time (reflecting the depth of the reflection interface) of the signal are calculated in the time domain, on the other hand, the time domain signal is converted into the frequency domain signal by fast Fourier transform (FFT), and the power spectral density of the signal is obtained, which reflects the energy distribution of different frequency components, for example, the reflected wave of the steel grouting interface is usually concentrated in the frequency band of 2-3 MHz, while the multiple reflection noise of the sleeve wall is usually distributed below 1 MHz, through the time-frequency domain analysis, the characteristic difference between the effective signal and the noise can be preliminarily distinguished, then the interface reflection wave and the sleeve wall multiple reflection noise in the signal are separated by wavelet packet decomposition technology, the wavelet packet decomposition technology is a signal processing method which can decompose the signal into different frequency subbands, compared with the traditional Fourier transform, it can more accurately capture the local characteristics of the signal, and the specific separation process is as follows: The original ultrasonic reflection signal was decomposed into 4 wavelet packets using the db4 wavelet basis, resulting in 16 frequency sub-bands. Based on the characteristic frequency band and noise frequency band of the interface reflection wave determined by joint time-frequency domain analysis, the signal components within the characteristic frequency band were retained, while the signal amplitude within the noise frequency band was set to zero. Then, the retained characteristic frequency band signals were recombined using a wavelet packet reconstruction algorithm to obtain a denoised signal containing only the reflection wave from the steel bar-grout interface, completely eliminating the noise influence caused by multiple reflections from the sleeve wall. Subsequently, the energy attenuation rate of the characteristic frequency band was calculated. The energy attenuation rate is an indicator that measures the energy loss of ultrasonic waves during propagation and reflection. During the calculation, the denoised interface reflection wave signal was first processed in the characteristic frequency band (2-3... The total energy is calculated within MHz by integrating the square of the signal amplitude. Then, the theoretical fully bonded reflected wave energy (i.e., the total reflected wave energy under the same testing conditions when the steel bar and grout are completely bonded and without any defects) is retrieved from the laboratory as a benchmark. Finally, the energy attenuation rate is calculated as: (Theoretical fully bonded energy - Actual reflected energy) / Theoretical fully bonded energy × 100%. For example, if the theoretical energy is 100 V²·s and the actual energy is 70 V²·s, the energy attenuation rate is 30%. A higher attenuation rate indicates a more severe micro-void at the interface. The acoustic energy loss coefficient is defined as the ratio of the defect reflected wave energy to the theoretical fully bonded reflected wave energy. Wave energy refers to the reflected wave energy in the denoised signal caused by micro-voids. The actual reflected energy exceeds that of the fully bonded state. For example, the reflected energy is 70 V²·s in full bond, but 90 V²·s with micro-voids. The defect reflected wave energy is 20 V²·s. Therefore, the acoustic energy loss coefficient = defect reflected wave energy / theoretical fully bonded reflected wave energy, with a value range of 0-1. The closer the coefficient is to 1, the more severe the micro-voids at the interface. Micro-voids enhance ultrasonic reflection and increase defect energy. Regions with an acoustic energy loss coefficient greater than a preset benchmark value are defined as micro-void regions. The preset benchmark value is calibrated through laboratory tests for multiple groups with different micro-void degrees (void rate 0). Ultrasonic testing was conducted on steel bar-sleeve specimens with a void rate of 5%-20%. The correlation between the acoustic energy loss coefficient and the actual void rate was statistically analyzed. An acoustic energy loss coefficient of 0.2, corresponding to a void rate of 5%, was set as the baseline. Using the detection data from each ultrasonic probe in a multi-dimensional monitoring network, areas with an acoustic energy loss coefficient > 0.2 were marked as micro-void regions, providing a basis for subsequent location of micro-void initiation. After identifying the micro-void regions, the local stress concentration threshold area needs to be determined by combining the strain data of the sleeve's outer wall. The initiation and expansion of micro-voids are often accompanied by stress concentration; the acoustic energy loss coefficient alone cannot accurately locate the core damage area. A dual-threshold diagnostic mechanism needs to be established by combining strain data. The specific implementation method is as follows: The specific steps of the method for identifying the local stress concentration threshold area are as follows: first, synchronously analyze the time-space distribution characteristics of the strain data of the outer wall of the sleeve. The time-space distribution characteristics refer to the comprehensive performance of the strain data in the time dimension (trend changing with the load) and the space dimension (distribution difference along the circumference of the sleeve). Synchronous analysis needs to align the strain data collected by the fiber Bragg grating strain gauge with the acoustic energy loss coefficient data detected by the ultrasonic detection according to the time stamp, so as to ensure that both are for the same load working condition in the same period. The specific refinement process is as follows: In the time dimension, the fluctuation law of the strain data with the load cycle is analyzed, and the strain peak value corresponding to each load peak value is counted. In the space dimension, the strain distribution curves in the circumferential direction of the sleeve are drawn by comparing the strain data of the four groups of strain gauges, and the uneven area of strain distribution is identified. When the reverse strain mutation of adjacent strain gauges is detected and the mutation amplitude continuously increases, it is determined that it is a local stress concentration area. The reverse strain mutation refers to the fact that under the same load working condition, one group of adjacent strain gauges shows a tensile strain (positive value) mutation, and the other group of strain gauges shows a compressive strain (negative value) mutation. This phenomenon indicates that the area between the two groups of strain gauges is locally distorted due to uneven stress, which is a typical feature of stress concentration. The continuous increase of the mutation amplitude refers to the fact that in the last 3 load cycles, the difference value of the strain mutation shows an increasing trend, indicating that the degree of stress concentration is continuously increasing. When both the reverse strain mutation and the continuous increase of the amplitude meet the two conditions, the area between the adjacent strain gauges is determined to be a local stress concentration area. Then, combined with the spatial distribution of the acoustic energy loss coefficient, an acoustic energy-strain double threshold diagnosis mechanism is established. This mechanism is screened by the threshold conditions in two dimensions to avoid misjudgment of a single indicator. First, two thresholds are determined: one is the acoustic energy loss coefficient threshold, which is the preset reference value 0.2 in the foregoing, and the coefficient exceeding 0.2 is determined to be a potential micro void area. The other is the strain mutation amplitude threshold, which is the yield threshold of the sleeve material. The establishment process of the double threshold diagnosis mechanism is as follows: From the spatial distribution of the acoustic energy loss coefficient, extract all potential micro void areas with a coefficient greater than 0.2. In these areas, superimpose the strain data and select the areas with a strain mutation amplitude greater than 510με. Only the areas that meet both threshold conditions are retained, and the interference areas with acoustic energy exceeding the limit but normal strain or normal acoustic energy but exceeding strain are excluded to ensure the accuracy of the diagnosis result. That is, in the acoustic energy loss coefficient exceeding area, if the strain mutation amplitude exceeding the set proportion of the material yield threshold is detected at the same time, the area is located as the micro void initiation core area. The set proportion is 30% (510με) in the foregoing. The location is realized by the following steps: According to the ultrasonic detection data of the over-limit region of the acoustic energy loss coefficient, the range of the region on the circumference of the sleeve is determined, and the position with the maximum strain mutation amplitude is found within the range. The position is taken as the center, and each side is expanded by 5° to form a micro-disengagement initiation core region. Finally, it is marked as a key update node of the three-dimensional disengagement rate model. When constructing the three-dimensional dynamic disengagement rate model, the disengagement evolution data of the region needs to be focused on, and the acoustic energy loss coefficient and strain data of the core region are updated every 10 minutes to ensure that the model can reflect the damage development trend in real time. After determining the micro-disengagement initiation core region, a spatial topology mapping algorithm is used to convert the multi-dimensional monitoring network discrete data (acoustic energy loss coefficient, strain mutation data of each sensing unit) into a continuous three-dimensional data field. Discrete monitoring points can only reflect the local state, and continuous data field can fully present the disengagement distribution of the steel sleeve interface. The specific implementation is as follows: The execution method of the spatial topology mapping algorithm is as follows: first, the monitoring points of the multi-dimensional monitoring network are taken as spatial control nodes, which include the detection positions corresponding to the four groups of composite sensing units and the encrypted detection points of the micro-disengagement initiation core region. Each monitoring point contains three groups of data: three-dimensional coordinates (sleeve axial coordinate z, circumferential angle θ, radial depth r, r is fixed as the surface position of the steel bar), acoustic energy loss coefficient, and strain mutation amplitude. These monitoring points are taken as spatial control nodes to provide anchoring basis for subsequent grid construction. Then, based on the Delaunay triangulation principle, the initial topological grid of the sleeve-steel interface is constructed. The Delaunay triangulation principle is an algorithm for constructing a triangular grid from a set of discrete points on a plane or curved surface. Its core feature is to maximize the minimum internal angle, that is, the minimum internal angle of the generated triangular grid is as large as possible to avoid the generation of narrow triangles and prevent the subsequent interpolation calculation error from being too large. The specific construction process is as follows: The sleeve-steel bar interface is unfolded into a plane (with the axial coordinate z as the x-axis and the circumferential angle θ as the y-axis), the z-θ coordinates of all the spatial control nodes are projected onto the plane, the Delaunay triangulation algorithm is called, adjacent control nodes are connected, a triangular mesh covering the entire interface plane is generated, the three vertices of each triangle are control nodes, and then the plane mesh is mapped back to the cylindrical surface of the sleeve, the z-θ plane coordinates are converted into three-dimensional cylindrical coordinates z-θ-r through coordinate transformation, and the initial topological mesh of the sleeve-steel bar interface is formed. The mesh can preliminarily reflect the spatial structure of the interface, but due to the limited number of control nodes, the mesh is relatively sparse, and each triangle covers a circumferential range of 5°-10°. Then, the sound energy loss coefficient and strain jump data of the discrete nodes are converted into a continuous spatial field by introducing a radial basis function interpolation model. The radial basis function interpolation model is an interpolation method for fitting a continuous function from discrete point data. The core is to use radial basis functions, which are only related to the distance from the point to the origin, to construct the influence weight of each control node on any spatial point, so as to calculate the parameter value at the position of the non-control node. The specific conversion process is as follows: A Gaussian radial basis function (function form φ(r)=exp(-(εr)²), ε is the shape parameter, and 0.1 is taken by trial calculation) is selected, the sound energy loss coefficient of each spatial control node is taken as the known value, and an interpolation function is constructed. For the internal points of each triangle in the initial topological mesh, the non-control nodes, the radial basis function weight is calculated according to the distance between the three vertex control nodes and the non-control node, and the sound energy loss coefficient of the point is obtained by weighted summation. Repeat the above process to interpolate the strain jump amplitude. Finally, each spatial point on the sleeve-steel bar interface is assigned a sound energy loss coefficient and a strain jump amplitude to form a continuous spatial field, i.e. a field where the parameter value changes continuously with the spatial coordinates, solving the problem that discrete monitoring points cannot cover the entire interface. Then, the dynamic mesh refinement technique is used to automatically refine the mesh nodes in the micro-looseness initiation core area to realize the reconstruction of the three-dimensional data field. The dynamic mesh refinement technique is a technique that automatically increases the grid density according to the parameter change gradient and the rate of change of the parameter value in the continuous spatial field. The micro-looseness initiation core area needs to be refined to improve the calculation accuracy due to the dramatic changes in the sound energy loss coefficient and the strain jump amplitude. The specific refinement process is as follows: The gradient value of the sound energy loss coefficient in the continuous space field is calculated, the coefficient variation in unit distance is calculated, a gradient threshold is set, the area where the gradient value exceeds the threshold, i.e. the micro void initiation core area 40°-50°, is marked as a to-be-refined area, new grid nodes are added to the midpoint of each side of the original triangular grid in the to-be-refined area, the sound energy loss coefficient and the strain mutation amplitude of these new nodes are calculated, the new nodes and the original nodes are reconnected through radial basis function interpolation, a denser triangular grid is generated, each triangle covers a circumferential range of 1°-2° after refinement, and finally, the refined grid and the original grid are integrated to form a three-dimensional data field covering the entire sleeve-steel interface and the encrypted core area, and the data field reconstruction is completed to provide continuous and high-precision basic data for subsequent generation of a three-dimensional dynamic void ratio model.

[0019] The generation method of the three-dimensional dynamic void ratio model is as follows: The three-dimensional data field is mapped to a parameterized space on the surface of the steel bar, the microelement control volume is divided based on the geometric parameters of the steel bar rib, the dynamic void ratio in each microelement is defined as the weighted product of the sound energy loss coefficient and the strain mutation amplitude, the overall void ratio is calculated by adding the void volumes of all microelements, and a void boundary tracking algorithm is established to record the expansion trajectory of the void area combined with the load history, which is used to construct a load cycle number-void expansion rate correlation function.

[0020] The establishment method of the load cycle number-void expansion rate correlation function is as follows: A damage accumulation factor is defined in the microelement control volume, which is driven to grow by the ratio of the current load amplitude to the maximum historical load amplitude, a void volume growth rate equation is established with the damage accumulation factor as the independent variable by fitting the laboratory accelerated aging test data through nonlinear regression analysis, and the amplitude-frequency distribution in the load spectrum is input as a boundary condition to realize dynamic prediction of void evolution under variable amplitude load.

[0021] The construction method of the bond strength coupling degradation model is as follows: Based on the interface direct shear test database of grouting material and steel bar, a friction coefficient decay curve is established with the dynamic void ratio as the variable, a rib engagement effect failure mechanism is embedded in the model, when the void ratio exceeds the critical threshold, the lateral restraint force of the rib is forced to zero, an expansion pressure model of corrosion products is introduced to convert the steel corrosion data into interface contact pressure loss, and finally the quantitative relationship between the bond strength and the coupling degradation of the void ratio and the corrosion degree is output to form the bond strength coupling degradation model.

[0022] It needs to be further explained that after the three-dimensional data field reconstruction is completed, it needs to be converted into a three-dimensional dynamic void rate model that can intuitively reflect the void state of the steel sleeve interface. This model not only quantifies the current void degree, but also tracks the void expansion track to provide dynamic damage data for subsequent life prediction. The specific implementation is as follows: The method comprises the following specific steps: firstly, mapping the three-dimensional data field to the parameterized space of the steel surface; the parameterized space of the steel surface is a parameterized curved surface obtained by mathematical modeling of the outer surface of the steel; the cylindrical coordinate system is constructed with the steel axis as the z-axis (axial direction), the steel circumferential direction as the theta-axis (hoop direction), and the steel radial direction as the r-axis (radial direction); in the mapping process, the sound energy loss coefficient, strain mutation amplitude and corresponding three-dimensional coordinates (z, theta, r) of all space points in the three-dimensional data field are extracted; then, the points are accurately projected onto the parameterized curved surface of the steel surface through coordinate transformation; for example, a space point is (z=50 cm, theta=45°, r=3 cm) in the sleeve coordinate system, and the corresponding coordinate of the parameterized space of the steel surface is (z=50 cm, theta=45°, r=2 cm, the steel radius is fixed at 2 cm); each parameter value of the three-dimensional data field can be corresponded to a specific position on the steel surface, laying a spatial foundation for subsequent micro-element division; then, the micro-element control volume is divided based on the geometric parameters of the steel rib; the geometric parameters of the steel rib include the rib height (1-1.5 mm), the rib spacing (8-12 mm), and the rib width (3-5 mm), which are determined by the steel production standard; these parameters directly affect the biting action of the steel and the grouting material, and are the core area where micro voids are prone to occur; in the division, the units are divided along the steel axial direction (z-axis) according to the rib spacing (for example, 10 mm), along the hoop direction (theta-axis) according to 1° as a unit, and along the radial direction (r-axis) according to the rib height (for example, 1.2 mm); finally, a small cubic micro-element control volume with a volume of about 10 mm x (pi x steel diameter / 360) x 1.2 mm is formed; each micro-element accurately corresponds to a small area on the steel surface, especially covering the rib top and both sides, the high-incidence area of micro voids, to ensure the refinement of void quantification; subsequently, the dynamic void rate in each micro-element is defined as the weighted product of the sound energy loss coefficient and the strain mutation amplitude; the dynamic void rate is an index for quantifying the void degree in a single micro-element, with a value of 0-1, 0 representing no void, and 1 representing complete void; the weight distribution of the weighted product is based on laboratory test calibration; the sound energy loss coefficient directly reflects the interface bonding defects (weight 0.6), and the strain mutation amplitude reflects the driving effect of stress concentration on void (weight 0.4); the sum of the weights of the two is 1, ensuring the reasonableness of the index; for example, the sound energy loss coefficient of a micro-element is 0.35 (corresponding to slight void), and the strain mutation amplitude is 500με (converted to a coefficient of 0.3, because the material yield threshold corresponds to a strain of 1700με, 500με accounts for about 30%); thus, the dynamic void rate of the micro-element is 0.35 x 0.6 + 0.3 x 0.4 = 0.21 + 0.12 = 0.33, which directly reflects the void degree of the micro-element; the void volumes of all micro-elements are added to calculate the overall void rate; the void volume of each micro-element control volume = micro-element total volume x dynamic void rate (for example, the micro-element total volume is 0.1cm³, dynamic void ratio 0.33, void volume = 0.033cm³), the overall void ratio = (the sum of all micro-element void volumes) ÷ (the total contact volume of the steel bar-sleeve interface) × 100%, wherein the total contact volume is the actual contact volume of the steel bar and the grouting material (excluding the area of the steel bar that is not wrapped), for example, the sum of all micro-element void volumes is 10cm³, the total contact volume is 500cm³, the overall void ratio = 2%, which reflects the macro-void state of the entire steel bar-sleeve interface, and provides a core parameter for subsequent bond strength analysis, while a void boundary tracking algorithm is established to record the expansion trajectory of the void area in combination with the load history, the void boundary tracking algorithm is an algorithm for real-time capturing of the position change of the edge of the void area, the core logic of which is to identify micro-elements with a dynamic void ratio of 0.1, a critical value of the void boundary defined in engineering, below 0.1 is considered to be no void, above 0.1 is considered to be a void area, connect these micro-elements to form a closed boundary as the outline of the void area, and the algorithm is established as follows. The dynamic void ratio data of all micro-elements is extracted in time sequence (every 1 hour), for each time point, micro-elements with a dynamic void ratio of 0.1 are selected, their coordinates in the parameterized space are recorded, and a polygon fitting algorithm is used to connect these micro-element coordinates to generate the boundary outline of the void area, the overlapping area and the offset distance of the boundary outline of adjacent time points are calculated to determine the expansion direction and expansion speed of the void area, forming the void expansion trajectory, the load history refers to the actual load record borne by the bridge during service, including load amplitude, action time, and cycle number, which is collected by the bridge health monitoring system, and the void expansion trajectory is associated with the load history at the corresponding time, which can provide data support for constructing the load cycle number-void expansion rate correlation function, and finally used to construct the load cycle number-void expansion rate correlation function, the core of the function is to quantify the relationship between load cycle number and void expansion rate, and the specific establishment process is as follows: The void area corresponding to different load cycle numbers is extracted from the void expansion trajectory, the void expansion rate is calculated, and the area growth value per unit time or per unit cycle is calculated, and the corresponding data of cycle number and expansion rate is converted into a function relationship through linear fitting, which can intuitively reflect the cumulative effect of void under cyclic load, and provide damage evolution law for subsequent life prediction, after obtaining the load cycle number-void expansion rate correlation function, it is necessary to further realize the dynamic prediction of void evolution under variable amplitude load through the function, the actual load borne by the bridge is not constant amplitude, and only constant load analysis cannot accurately reflect the development of void, therefore, a damage accumulation factor needs to be introduced to construct a prediction model, and the specific implementation is as follows: The method for establishing the load cycle number-void expansion rate correlation function comprises the following specific steps: first, a damage accumulation factor is defined in the microelement control volume, which is driven to grow by the ratio of the current load amplitude to the historical maximum load amplitude, the damage accumulation factor is an index for quantifying the damage accumulation degree of the microelement under variable amplitude load, and the value is 0-1, 0 represents no damage, and 1 represents damage saturation, the core is the engineering application of the Miner linear cumulative damage theory, that is, a certain damage is caused by each load action, and the damage is cumulative, and the specific definition is: damage accumulation factor = Σ (damage increment of the i th load action), wherein the damage increment = (current load amplitude / historical maximum load amplitude) * 0.001 (basic damage coefficient of each load action, calibrated by accelerated aging test), for example, the historical maximum load amplitude of a microelement is 100 kN, and the load amplitude of a certain time is 80 kN, the damage increment of this time is (80 / 100) * 0.001 = 0.0008, if the microelement has accumulated damage of 0.2, the updated damage accumulation factor is 0.2 + 0.0008 = 0.2008, when the load amplitude exceeds the historical maximum load, the historical maximum load amplitude is updated first, and then the damage increment is calculated according to the new ratio (such as the load amplitude of 120 kN, the damage increment is (120 / 120) * 0.001 = 0.001), so as to ensure that the factor can reflect the additional damage caused by overloading, then the laboratory accelerated aging test data is fitted by nonlinear regression analysis, the laboratory accelerated aging test is to apply different amplitude cyclic loads to multiple groups of steel sleeve test pieces in the laboratory environment, and to monitor the void volume growth data of each group of test pieces under different cycle numbers, the void volume is measured every 100 cycles, and finally a three-dimensional test database containing load amplitude-cycle number-void volume is formed, and the specific process of nonlinear regression analysis is as follows: A suitable regression model is selected, an exponential function model y = a * e ^ (b * x) + c is selected according to the test data trend, wherein y is the void volume, x is the damage accumulation factor, a, b and c are to-be-solved parameters, the damage accumulation factor-void volume data pair in the test database is substituted into the model, the least square method (the residual sum of squares of the model predicted value and the test value is calculated, the parameters a, b and c are adjusted to minimize the residual) is used to fit the parameters, for example, a = 1.5, b = 5 and c = 0.5 are obtained by calculation, at this time, the error between the model predicted value and the test value is ≤5%, which meets the engineering precision requirement, the model significance is verified by F test, the significance level is set to 0.05, if the test result P < 0.05, it is indicated that the model can effectively fit the test data, then a void volume growth rate equation taking the damage accumulation factor as the independent variable is established, the void volume growth rate refers to the growth amount (cm³ / h) of the void volume per unit time, the damage accumulation factor needs to be converted into a time dimension index in combination with the load cycle frequency and the relationship between the actual bridge load cycle number and time, and the specific establishment process is as follows: According to the void volume model y=1.5e^(5x)+0.5 obtained by nonlinear regression (x is the damage cumulative factor), the derivative of x is obtained dy / dx=7.5e^(5x), the time variation rate of damage cumulative factor dx / dt=10x0.001=0.01 / h is calculated by the load cycle frequency (such as 10 times / h) and the damage increment of each cycle (such as 0.001), and according to the chain rule, the void volume growth rate v=dy / dt=dy / dxxdx / dt=7.5e^(5x)0.01=0.075e^(5x), the equation takes the damage cumulative factor x as the independent variable, and can calculate the void expansion rate in different damage states in real time, finally the amplitude-frequency distribution in the load spectrum is input as the boundary condition to realize the dynamic prediction of the void evolution under variable amplitude load, the load spectrum is the actual load statistical data recorded in the bridge design and health monitoring, including different load amplitudes (such as 50kN, 80kN, 100kN) and their corresponding action frequencies (such as 50 times, 30 times, 20 times per day); when the boundary condition is input, the amplitude-frequency data of the load spectrum is converted into the damage cumulative factor variation rate per unit time (such as 50kN load 50 times a day, each damage increment is 0.0005, and the daily damage increment is 50x0.0005=0.025), the variation rate is substituted into the void volume growth rate equation to calculate the void volume growth value in different time periods (such as 0.204x24=4.896cm³ in the first day), and the growth value is accumulated in time sequence to obtain the void volume evolution curve in a period of time in the future (such as the void volume reaches 60cm³ after 30 days), thereby realizing the dynamic prediction of the void evolution under variable amplitude load, and providing damage data in time dimension for subsequent bond strength degradation analysis. After the void evolution law is clear, a bond strength coupling degradation model needs to be constructed. The bond strength is the core index of force transmission at the interface between steel bars and grouting material, and the void and steel bar corrosion will jointly cause its degradation. Considering only a single factor will underestimate the risk of the structure, therefore, a coupling model needs to be established to quantify the joint effect. The specific implementation is as follows: The construction method of the bond strength coupling degradation model, the specific steps are as follows: first, a friction coefficient decay curve taking the dynamic void rate as the variable is established based on the interface direct shear test database of grouting material and steel bar. The interface direct shear test database of grouting material and steel bar is the basic data accumulated by laboratory direct shear test. During the test, multiple groups of steel bar-grouting material specimens with different dynamic void rates (0%, 5%, 10%, 15%, 20%) are prepared, an axial tension is applied on a universal testing machine, the shear force when the interface slips is measured, and the corresponding relationship of dynamic void rate-shear force-friction coefficient (friction coefficient= shear force ÷ interface normal pressure, the normal pressure is provided by the self weight of the specimen and the loading device, and is fixed at 100kN) is recorded. Finally, a database containing 500+ groups of data is formed. The specific process of establishing the friction coefficient decay curve is as follows: The friction coefficients corresponding to different dynamic void rates are extracted from the database (such as friction coefficient 0.6 when void rate is 0%, 0.5 when 5%, 0.35 when 10%, 0.2 when 15%, and 0.1 when 20%), a quadratic polynomial is used to fit these data, the fitting model is f=-0.002d²-0.02d+0.6, where f is the friction coefficient and d is the dynamic void rate, the fitting accuracy is verified, the average error between the fitting value and the test value is ≤3%, and finally the attenuation curve of dynamic void rate-friction coefficient is formed, which directly reflects the weakening effect of void on the interface friction transmission, then the ribbing occlusion effect failure mechanism is embedded in the model, when the void rate exceeds the critical threshold, the ribbing lateral restraint force is forced to zero, the ribbing occlusion effect is the mechanical occlusion effect of the ribbing on the surface of the steel bar and the grouting material, which is an important part of the holding strength (accounting for 40% of the total holding strength), when the dynamic void rate is less than 10%, the ribbing can still occlude the grouting material to transfer force, the critical threshold is determined to be 10% through tests, when the dynamic void rate exceeds 10%, the void area covers most of the ribbing, the occlusion effect fails, at this time the ribbing lateral restraint force needs to be forced to zero in the model, the specific embedding process is: An occlusion effect switch is added to the model, when the dynamic void rate d≤10%, the switch is opened, the ribbing lateral restraint force=0.4×holding strength reference value, the holding strength reference value is the holding strength without void and rust, when d>10%, the switch is closed, the ribbing lateral restraint force=0, to ensure that the model can truly reflect the failure process of the ribbing occlusion effect, and the rust product expansion pressure model is introduced to convert the steel bar rust data into interface contact pressure loss, the rust product expansion pressure model is used to calculate the pressure and contact pressure loss generated by the volume expansion of the rust product on the interface after the steel bar is rusted, the core principle is that the volume of the rust product is about 2-4 times the volume of the steel bar, the expansion will cause the separation of the steel bar and the grouting material interface, and reduce the contact pressure, the model is calibrated by laboratory tests, the input parameter is the steel bar rust depth collected by the bridge monitoring system, and the output parameter is the interface contact pressure loss, the specific conversion process is: The steel bar corrosion depth data (such as 0.2mm) is obtained, the corrosion product expansion amount (such as corrosion depth 0.2mm corresponding to expansion amount 0.5mm) is calculated according to the expansion pressure model, and the contact pressure loss amount = 50x0.5 = 25kN / m is calculated through the expansion amount-contact pressure loss relationship curve (calibrated as contact pressure loss amount = 50x expansion amount, unit kN / m·mm), if the initial interface contact pressure is 100kN / m, then the current contact pressure = 100-25 = 75kN / m, which reflects the weakening effect of corrosion on interface force transmission, and finally the quantitative relationship between the coupling degradation of the gripping strength and the coupling degradation of the corrosion degree is output, forming the coupling degradation model of the gripping strength, and the total degradation amount of the gripping strength is composed of the friction transmission degradation ribbing occlusion transmission degradation contact pressure degradation, and the specific quantitative process is: The friction transmission component = friction coefficient x current contact pressure (such as friction coefficient 0.35, current contact pressure 75kN / m, component = 26.25kN / m) is calculated, the ribbing occlusion transmission component (according to the void ratio, such as void ratio 12%>10%, component = 0) is calculated, and the total gripping strength = friction transmission component + ribbing occlusion transmission component (such as 26.25+0 = 26.25kN / m) is calculated. A three-dimensional quantitative relationship table of dynamic void ratio-corrosion depth-gripping strength (such as void ratio 12%, corrosion depth 0.2mm corresponding to gripping strength 26.25kN / m; void ratio 8%, corrosion depth 0.1mm corresponding to gripping strength 45kN / m) is established, and it is converted into a model formula that can be calculated in real time, and finally a coupling degradation model of gripping strength is formed. The model can accurately output the gripping strength value under different combinations of void ratio and corrosion degree, providing core mechanical parameters for subsequent stress redistribution analysis and life prediction.

[0023] The execution method of the multi-physical field coupling simulation is as follows: The sound energy loss coefficient, the gripping strength degradation data, the expansion amount of the concrete crack width and the steel bar corrosion depth are taken as joint inputs, the gripping strength degradation data is obtained from the gripping strength coupling degradation model, the interface friction contact nonlinear equation is solved by an implicit-explicit iterative algorithm, the concrete crack area is simulated by a dispersion crack model to reduce the stiffness, the steel bar corrosion area is subjected to an expansion strain load, and finally the principal stress redistribution cloud diagram and the internal force transfer path evolution animation of the connection part after the gripping failure are output to obtain the stress redistribution. Taking the gripping force failure warning as the time zero point, the stress redistribution result is input into the section internal force balance equation to back-propagate the remaining bending stiffness of the connection part, a random dynamic load time history is generated based on the vehicle flow axle weight distribution in the actual load spectrum, the concrete carbonation depth prediction model is driven combined with the environmental temperature and humidity data, and finally the Monte Carlo sampling technology is used to simulate the stiffness degradation path under the action of 10,000 times of load to output the probability density function curve of the remaining life.

[0024] It needs to be further explained that after the construction of the coupling degradation model of the gripping strength and the triggering of the gripping force failure warning, the stress redistribution caused by the gripping failure needs to be quantified through multi-physical field coupling simulation. The stress of the bridge connection part is the result of the synergistic action of the steel-grouting material-concrete multi-medium. The failure of a single component will lead to the change of the overall force transmission path. Only through multi-physical field simulation can this complex change be accurately captured. The specific implementation is as follows: The execution method of the multi-physical field coupling simulation is as follows: First, the sound energy loss coefficient, the gripping strength degradation data, the expansion amount of the concrete crack width, and the steel corrosion depth are taken as joint inputs. These four types of parameters correspond to the interface bonding state, the force transmission ability, the concrete damage, and the steel deterioration, respectively, and together constitute the core boundary conditions of the simulation. Among them, the gripping strength degradation data is obtained from the gripping strength coupling degradation model, and the specific obtaining process is as follows: First, the current dynamic void rate of the steel-sleeve interface is extracted from the three-dimensional dynamic void rate model (such as 15%), and the steel corrosion depth is obtained from the bridge corrosion monitoring system (such as 0.3mm). These two parameters are input into the gripping strength coupling degradation model. The model first queries the friction coefficient decay curve through the dynamic void rate (15% void rate corresponds to friction coefficient 0.2), and then reduces the rib side constraint force to zero according to the critical threshold of void rate exceeding 10%. At the same time, the interface contact pressure loss amount is calculated through the corrosion product swelling pressure model (0.3mm corrosion depth corresponds to loss amount 30kN / m, initial contact pressure 100kN / m, current contact pressure 70kN / m). Finally, the gripping strength degradation data is output (friction force transmission component = 0.2x70 = 14kN / m, total gripping strength = 14+0 = 14kN / m), which directly reflects the degree of force transmission ability decay after gripping failure. Then the interface friction contact nonlinear equation is solved through the implicit-explicit iterative algorithm. The interface friction contact nonlinear equation is a mathematical equation that describes the nonlinear behavior of the steel-grouting material interface under load friction sliding and contact separation. The core difficulty lies in the dynamic change of the interface state with the change of the load (such as from contact to separation), which needs to be gradually approached to the true solution through the iterative algorithm. The specific calculation process is as follows: The implicit algorithm is used to solve the overall equilibrium equation of the structure. The reinforcement, grouting material and concrete at the bridge connection part are established as a finite element model according to the actual size. The mesh is divided (fine mesh is used at the interface between the reinforcement and the grouting material, with an element size of 0.5 mm to ensure the capture of the interface behavior). The initial load (such as the vertical load of 50 kN generated by vehicle passing) is applied. The node displacement and stress are solved by the Newton-Raphson iteration method to obtain the preliminary interface contact state (such as the contact pressure of 10 kN / m in some areas and the contact pressure of 0 in some areas due to voiding). The explicit algorithm is used to process the nonlinear contact in the interface area. For the interface elements that are difficult to converge in the implicit algorithm, the explicit time integration method is switched. The load history is divided into steps of 0.001 seconds. In each step, the interface element contact state (contact or separation) is judged. The friction stress of the contact element is calculated (friction stress = friction coefficient x contact pressure). The friction stress of the separated element is set to 0. The implicit and explicit iterations are alternately executed. The implicit algorithm ensures the calculation accuracy of the overall structure, and the explicit algorithm solves the nonlinear problem of the interface. After each iteration, the calculation results are compared with the convergence criterion (such as the stress difference between the adjacent two iterations < 5%). If the convergence criterion is not met, the interface contact parameters are adjusted and the calculation is restarted until the convergence criterion is met. Finally, the nonlinear solution of the interface friction contact is obtained. The dispersion crack model is used to simulate the stiffness reduction in the concrete crack area. The dispersion crack model is a numerical simulation method that disperses the concrete cracks in the element interior rather than concentrating them on the element boundary. It is suitable for simulating the distributed micro-crack propagation of concrete under load. The core of the model is to reflect the stiffness reduction by reducing the elastic modulus of the crack area element. In practice, the crack width expansion (such as 0.2 mm) of the concrete crack is first obtained from the bridge crack monitoring system. According to the crack width-stiffness reduction coefficient relationship curve (0.2 mm crack corresponds to a reduction coefficient of 0.7 through experimental calibration), the elastic modulus of the concrete element with cracks in the finite element model is adjusted from 30 GPa to 30 x 0.7 = 21 GPa to simulate the stiffness reduction caused by cracks. At the same time, the model automatically calculates the stress transfer path change in the crack area (such as stress transfer from the crack area to the surrounding intact concrete), ensuring that the simulation results match the actual damage state. The expansion strain load is applied to the steel corrosion area. The volume of the corrosion product (such as iron hydroxide and iron oxide) generated after the steel corrosion will expand (about 2-4 times the volume of the original steel). This expansion will generate extrusion stress on the surrounding grouting material and concrete, further aggravating the interface voiding and concrete cracking. According to the steel corrosion depth (0.3 mm), the expansion amount of the corrosion product (0.3 mm corrosion depth corresponds to a radial expansion amount of 0.6 mm) is calculated through the material mechanics formula. In the finite element model, the radial expansion strain (strain value = expansion amount / steel radius, steel radius 10 mm, strain value = 0.06), simulating the expansion effect of rust products, the expansion strain will be converted into extrusion stress (such as 2 MPa) of grouting material, the stress will be superimposed with external load, affecting the overall stress distribution, and finally outputting the main stress redistribution cloud of the connection part after the grip failure and the evolution animation of the internal force transfer path, the main stress redistribution cloud is generated by the post-processing module of the finite element software, and the main stress of each area of the connection part is intuitively displayed by different colors (such as red representing high stress area and blue representing low stress area), before the grip failure, the stress is mainly concentrated on the steel bar-grouting material interface (such as 30 MPa), after the failure, the interface stress is reduced to 5 MPa, and the stress is transferred to the sleeve outer wall and the surrounding concrete (the stress of the sleeve outer wall is increased to 25 MPa, and the stress of the concrete is increased to 15 MPa), the evolution animation of the internal force transfer path shows the transfer process of the internal force from the original transfer path (steel bar-grouting material-concrete) to the new path (steel bar-sleeve-concrete) in time sequence (such as 0-1 second), clearly presenting the evolution law of grip failure-interface force interruption-sleeve and concrete bearing main load, by analyzing the cloud and the animation, the stress change data (such as sleeve stress growth rate, concrete stress concentration area position) of each component can be extracted, and the stress redistribution can be quantitatively obtained, which provides a basis for subsequent residual stiffness calculation, after obtaining the stress redistribution data, the residual life probability of the connection part needs to be calculated based on this, the life prediction needs to be combined with the structure stiffness degradation law, the actual load and the environmental influence, the uncertainty is quantified through probability analysis, and the prediction deviation caused by deterministic analysis is avoided, and the specific implementation is as follows. The implementation process of taking the grip force failure warning as the starting node of life prediction, and the specific steps are as follows: first, taking the grip force failure warning as the time zero point, inputting the stress redistribution result into the section internal force balance equation to reverse the remaining bending stiffness of the connection part, when the grip force failure warning is triggered (the dynamic void rate exceeds the critical threshold of 20%, and the grip strength is reduced to less than 30% of the design value), the system marks this moment as time zero (t=0), which marks that the connection part enters the life decay stage, and the remaining bending stiffness reverse process is as follows: The stress distribution data (such as sleeve stress 25 MPa, concrete stress 15 MPa, and steel stress 20 MPa) of the key section (such as the middle section of the sleeve) of the connection part is extracted from the stress redistribution result, the section internal force balance equation is established, the total bending moment of the section = the bending moment borne by the sleeve + the bending moment borne by the concrete + the bending moment borne by the steel, wherein the bending moment borne by each component is calculated by multiplying the stress by the corresponding section modulus (such as the sleeve section modulus 100 cm 4 , the bending moment borne = 25 MPa x 100 cm 4=2500 N-m), the curvature of the section is calculated according to the formula of material mechanics (curvature = bending moment / (elastic modulus x sectional moment of inertia), considering the elastic modulus after stiffness reduction), the overall residual bending stiffness of the connecting part is obtained by multiplying the sectional bending stiffness by the length of the component (e.g. the original stiffness is 10 5 N-m 2 , the residual stiffness is 4 x 10 4 N-m 2 ), which reflects the ability of the structure to resist deformation after the grip failure. Then, a random dynamic load time history is generated based on the axle weight distribution of the traffic flow in the actual load spectrum. The actual load spectrum is collected by the bridge weighing system and traffic monitoring camera for a long period of time, which contains different axle weights (e.g. 20 kN, 50 kN, 100 kN, 150 kN) and their corresponding frequencies (e.g. 20 times, 50 times, 30 times, 10 times per day). The specific process of generating the random dynamic load time history is as follows: Statistical analysis of the axle weight data in the load spectrum is performed to determine the probability distribution type of the axle weight (e.g. normal distribution, mean 60 kN, standard deviation 20 kN). A random number generation algorithm (e.g. Monte Carlo random sampling) is used to randomly generate an axle weight sequence for a future period of time (e.g. 1 year) according to the probability distribution of the axle weight (e.g. 55 kN in the first hour, 70 kN in the second hour). Combined with the time distribution law of the bridge traffic flow (e.g. dense traffic flow during the morning peak period from 7 to 9 am, 1 load every 10 minutes; 1 load every 30 minutes during the flat peak period), the axle weight sequence is arranged in chronological order to form a random dynamic load time history corresponding to the time-axle weight. This time history can truly reflect the variable amplitude load characteristics of the bridge, avoiding the prediction bias of stiffness degradation caused by the use of constant load. A concrete carbonation depth prediction model is driven by environmental temperature and humidity data. Concrete carbonation refers to the chemical reaction between carbon dioxide in the air and calcium hydroxide in concrete to form calcium carbonate, which reduces the alkalinity of concrete and further accelerates the corrosion of steel reinforcement. The rate is significantly affected by temperature and humidity (carbonation accelerates when temperature rises and humidity is moderate). The concrete carbonation depth prediction model is an empirical model based on Fick's second law. The input parameters are environmental temperature and humidity, exposure time, and strength level of concrete. The output parameter is carbonation depth. The specific driving process is as follows: Real-time temperature and humidity data are acquired from a bridge environment monitoring system, and exposure time from time zero is counted, the temperature and humidity data are converted into carbonation rate coefficients, the rate coefficients and exposure time are substituted into a prediction model, the carbonation depth = rate coefficient x exposure time, the carbonation depth is calculated, the protection layer thickness effective coefficient of the concrete is adjusted according to the carbonation depth, the coefficient will affect the protection ability of the concrete to the steel bar, and then is reflected in the stiffness degradation model, the carbonation deepening leads to the risk of cracking of the protection layer increasing, and the stiffness reduction coefficient increases, finally, the Monte Carlo sampling technology is used to simulate the stiffness degradation path under the action of 10,000 times of load, and the probability density function curve of the remaining life is output, the Monte Carlo sampling technology is a probability analysis method for simulating the influence of uncertain factors on the results through a large number of random sampling, which is used here to quantify the influence of factors such as load amplitude and material parameter fluctuation on the stiffness degradation, and the specific process is as follows: The sampling parameters and distribution are determined, the random dynamic load amplitude, the concrete carbonation rate coefficient and the steel bar corrosion rate are selected as random variables, 10,000 times of sampling are performed, a group of parameter values is extracted from the distribution of each random variable in each sampling, the stiffness degradation path is simulated, each group of sampling parameters is substituted into the stiffness degradation model, the load effect is calculated by the random dynamic load, the material degradation degree is calculated by the carbonation and corrosion data, and then the remaining stiffness after each load cycle is obtained, the stiffness-time curve under the group of parameters is obtained, the life termination condition is defined, when the remaining stiffness is reduced to 20% (2x10 4 N·m 2 ) of the original stiffness, it is determined that the connecting part fails, the life value of this sampling is recorded, the life values of 10,000 times of sampling are counted, and the probability density function curve is drawn, the horizontal axis of the curve is the life value, and the vertical axis is the occurrence probability corresponding to the life, the life value corresponding to the peak value of the curve is the most possible remaining life, and the dispersion degree of the curve reflects the uncertainty of the life prediction, the smaller the dispersion degree, the more reliable the prediction, through the curve, the probability distribution of the remaining life of the connecting part can be directly obtained, and a scientific basis is provided for bridge maintenance decision.

[0025] The second object of the present application is to provide a system for realizing the life prediction method for the bridge connecting part according to any one of the above. The multi-source collaborative perception unit 1 is arranged with four groups of composite sensing units at the junction of the curved section and the straight section of the grouting sleeve, each group of units is integrated with an adaptive focusing ultrasonic probe, a fiber bragg grating strain gauge and a temperature and humidity compensator, and the sleeve strain, the steel bar-grouting interface ultrasonic reflection signal and the environmental temperature and humidity data are transmitted in real time to the micro void evolution analysis unit 2 through an industrial bus; The micro void evolution analysis unit 2 is internally provided with an acoustic energy-strain double threshold diagnosis sub-module, a three-dimensional dynamic void rate modeling sub-module and a gripping failure early warning sub-module, and the void quantization result and the early warning instruction are synchronously output to the life prediction decision unit 3; The life prediction decision unit 3 receives the grip force failure early warning signal and the void evolution data of the micro-void evolution analysis unit 2, calculates the residual stiffness of the connecting position through a multi-physical field coupling simulator, drives a time-varying reliability analyzer, and finally generates a bridge connecting position life evaluation report.

[0026] The present application collects grouting sleeve strain, steel-grouting interface ultrasonic reflection signal and environmental temperature and humidity data through a distributed sensor array, generates a three-dimensional dynamic void rate model through a spatial topology mapping algorithm, establishes a correlation function between load cycles and void expansion, inputs the dynamic void rate into a coupling degradation model of the grip strength, quantifies the strength attenuation combined with the steel corrosion data, triggers the grip failure early warning, calculates the stress redistribution through a multi-physical field coupling simulation, uses the Monte Carlo sampling simulation to simulate the stiffness degradation combined with the actual load spectrum and temperature and humidity data, and outputs the residual life probability, so that the system contains a multi-source collaborative perception unit 1, a micro-void evolution analysis unit 2 and a life prediction decision unit 3, and the safety control precision of the bridge connecting position is improved.

[0027] The basic principles, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited by the above examples, and the above examples and descriptions in the specification are only preferred examples of the present application and are not intended to limit the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for predicting the lifespan of bridge connection parts, characterized in that: Includes the following steps: S1. Collect strain data of the outer wall of the sleeve, ultrasonic reflection signal of the steel reinforcement grouting interface and ambient temperature and humidity data by a distributed sensor array installed at the stress part of the grouting sleeve, in order to build a multi-dimensional monitoring network. S2. Based on the ultrasonic reflection signal of the steel bar grouting interface, the interface acoustic energy loss coefficient is extracted to quantify the degree of bonding defects in the micro-void area. Combined with the strain data of the outer wall of the sleeve, the local stress concentration threshold area is identified to locate the micro-void initiation location. Finally, the discrete monitoring point data of the multi-dimensional monitoring network is fused through the spatial topology mapping algorithm to generate a three-dimensional dynamic void rate model of the steel bar-sleeve interface, output the dynamic void rate, and establish a correlation function between the number of load cycles and the void expansion rate to characterize the damage accumulation effect under cyclic loading. S3. Input the dynamic void rate into the preset bond strength coupling degradation model and set the critical threshold of the void rate. When the dynamic void rate exceeds the critical threshold, trigger the bond force failure warning. Integrate the concrete crack width and steel corrosion data, and calculate the stress redistribution of the connection part caused by bond failure through multi-physics field coupling simulation to quantify the structural internal force transfer caused by force transmission failure. S4. Taking the gripping force failure warning as the starting point for life prediction, the remaining stiffness of the connection part is inferred based on stress redistribution, and a time-varying reliability function is constructed by combining the actual load spectrum and environmental temperature and humidity data to calculate the remaining life probability of the bridge connection part.

2. The life prediction method for bridge connection parts according to claim 1, characterized in that: The specific steps for constructing the multi-dimensional monitoring network are as follows: Four sets of strain-ultrasonic composite sensing units are arranged circumferentially at the junction of the curved and straight sections of the grouting sleeve. Each unit contains a contact fiber optic strain gauge to capture multi-axial deformation of the outer wall of the sleeve and an adaptive focusing ultrasonic probe. The ultrasonic probe penetrates the metal wall of the sleeve at an inclined incident angle. The emission frequency and focusing depth are adjusted to match the spacing of the steel ribs. At the same time, a temperature and humidity compensation module is integrated in each sensing unit to eliminate the interference of environmental temperature and humidity fluctuations on the propagation speed of ultrasonic signals, thus forming a multi-dimensional monitoring network covering the entire circumference of the sleeve.

3. The life prediction method for bridge connection parts according to claim 1, characterized in that: The method for extracting the interface acoustic energy loss coefficient comprises the following steps: A joint time-frequency domain analysis was performed on the ultrasonic reflection signal at the grouting interface of the steel reinforcement. The interface reflection wave and the multiple reflection noise of the sleeve wall were separated by wavelet packet decomposition. The energy attenuation rate of the characteristic frequency band was calculated, and the acoustic energy loss coefficient was defined as the ratio of the energy of the defect reflection wave to the energy of the theoretical fully bonded reflection wave. Finally, the region with an acoustic energy loss coefficient greater than the preset benchmark value was identified as the micro-void region.

4. The life prediction method for bridge connection parts according to claim 3, characterized in that: The method for identifying the local stress concentration threshold region comprises the following steps: The spatiotemporal distribution characteristics of strain data on the outer wall of the sleeve are analyzed synchronously. When a reverse strain mutation is detected in an adjacent strain gauge and the mutation amplitude continues to increase, it is determined to be a local stress concentration area. Combined with the spatial distribution of the acoustic energy loss coefficient, an acoustic-strain dual threshold diagnostic mechanism is established. That is, in the area where the acoustic energy loss coefficient exceeds the limit, if the strain mutation amplitude is detected to exceed the set proportion of the material yield threshold, the area is located as the core area of ​​micro-void initiation and marked as the key update node of the three-dimensional void rate model.

5. The life prediction method for bridge connection parts according to claim 1, characterized in that: The specific steps for executing the spatial topology mapping algorithm are as follows: The monitoring points of the multi-dimensional monitoring network are used as spatial control nodes. An initial topological mesh of the sleeve-reinforcement interface is constructed based on the Delaunay triangulation principle. By introducing a radial basis function interpolation model, the acoustic energy loss coefficient and strain mutation data of discrete nodes are transformed into a continuous spatial field. Then, dynamic mesh refinement technology is used to automatically densify the mesh nodes in the micro-void initiation core area to realize the reconstruction of the three-dimensional data field.

6. The life prediction method for bridge connection parts according to claim 5, characterized in that: The specific steps for generating the three-dimensional dynamic void rate model are as follows: The three-dimensional data field is mapped to the parameterized space of the steel bar surface. The micro-element control volume is divided based on the geometric parameters of the steel bar ribs. The dynamic voiding rate in each micro-element is defined as the weighted product of the acoustic energy loss coefficient and the strain mutation amplitude. The voiding volumes of all micro-elements are then summed to calculate the overall voiding rate. At the same time, a voiding boundary tracking algorithm is established. Combined with the load history record, the expansion trajectory of the voiding area is used to construct the correlation function of load cycle number and voiding expansion rate.

7. The life prediction method for bridge connection parts according to claim 1, characterized in that: The method for establishing the correlation function between the number of load cycles and the rate of vacancy expansion is as follows: A damage accumulation factor is defined within the micro-element control volume. This factor is driven by the ratio of the current load amplitude to the historical maximum load amplitude. By fitting laboratory accelerated aging test data through nonlinear regression analysis, an equation for the growth rate of the void volume with the damage accumulation factor as the independent variable is established. The amplitude-frequency distribution in the load spectrum is used as the boundary condition input to achieve dynamic prediction of void evolution under variable amplitude load.

8. The life prediction method for bridge connection parts according to claim 1, characterized in that: The specific steps for constructing the grip strength coupling degradation model are as follows: Based on the direct shear test database of the interface between grout and steel reinforcement, a friction coefficient decay curve with dynamic void rate as the variable is established. The rib interlocking effect failure mechanism is embedded in the model. When the void rate exceeds the critical threshold, the lateral constraint force of the rib is forcibly reduced to zero. At the same time, the expansion pressure model of corrosion products is introduced to convert the steel reinforcement corrosion data into the interface contact pressure loss. Finally, the quantitative relationship of the bond strength with the coupling degradation of void rate and corrosion degree is output, forming a bond strength coupling degradation model.

9. The life prediction method for bridge connection parts according to claim 1, characterized in that: The specific steps for performing the multiphysics coupling simulation are as follows: The acoustic energy loss coefficient, bond strength degradation data, concrete crack width expansion, and steel corrosion depth are used as joint inputs. The bond strength degradation data is obtained by the bond strength coupled degradation model. The interface friction contact nonlinear equation is solved by an implicit-explicit iterative algorithm. The stiffness reduction in the concrete crack area is simulated by a diffuse crack model, and the expansion strain load is applied to the steel corrosion area. Finally, the principal stress redistribution cloud map and internal force transfer path evolution animation of the connection part after bond failure are output to obtain stress redistribution. Using the gripping force failure warning as the zero point, the stress redistribution results are input into the internal force balance equation of the section to back-calculate the remaining bending stiffness of the connection. Based on the redistribution of the vehicle flow axis in the actual load spectrum, a random dynamic load time history is generated. Combined with environmental temperature and humidity data, a concrete carbonation depth prediction model is driven. Finally, Monte Carlo sampling technology is used to simulate the stiffness degradation path under 10,000 loads and output the probability density function curve of the remaining life.

10. A system for implementing the life prediction method for bridge connection parts according to any one of claims 1-9, characterized in that, include: The multi-source collaborative sensing unit (1) deploys four sets of composite sensing units in a circumferential manner at the junction of the curved section and the straight section of the grouting sleeve. Each set of units integrates an adaptive focusing ultrasonic probe, a fiber optic strain gauge and a temperature and humidity compensator. The sleeve strain, the ultrasonic reflection signal of the steel bar-grouting interface and the ambient temperature and humidity data are transmitted to the micro-void evolution analysis unit (2) in real time through the industrial bus. The micro-void evolution analysis unit (2) has a built-in acoustic energy-strain dual threshold diagnosis submodule, three-dimensional dynamic void rate modeling submodule and grip failure early warning submodule, which outputs the void quantification results and early warning instructions to the lifetime prediction decision unit (3) in a synchronized manner. The life prediction decision unit (3) receives the gripping force failure warning signal and the delamination evolution data from the micro-delamination evolution analysis unit (2), calculates the remaining stiffness of the connection part through the multi-physics field coupling simulator, and drives the time-varying reliability analyzer to finally generate the life assessment report of the bridge connection part.

Images