Method for intelligently predicting performance degradation and evaluating durability limit state of reinforced concrete structure

Abstract

The invention provides a reinforced concrete structure performance degradation intelligent prediction and durability limit state evaluation method. The method comprises the following steps: establishing a random variable probability model of environment and structure parameters; a convolutional neural network is fused to construct a chloride ion diffusion intelligent prediction model, and Monte Carlo sampling is adopted to analyze probability distribution of the initial corrosion time of the steel bar, so that prediction of the steel bar de-blunt time is realized; establishing a steel bar time-varying corrosion model to obtain the change condition of the steel bar corrosion loss rate along with time; further, an incremental static analysis method is adopted, and a multi-scale finite element model of the reinforced concrete structure under different corrosion rate conditions is established through random sampling; obtaining a critical load value in a limit state, and obtaining a failure probability curve under different corrosion degrees; finally, the bearing capacity failure time of the reinforced concrete structure is obtained by defining a bearing capacity reduction coefficient, and evaluation of the durability limit state of the reinforced concrete structure in the corrosion state is achieved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of concrete structures, and particularly relates to a reinforced concrete structure performance degradation intelligent prediction and durability limit state evaluation method. BACKGROUND

[0002] Reinforced concrete structures are widely used in various engineering fields due to their good mechanical properties and durability. However, during service, reinforced concrete structures are exposed to complex environments for a long time, especially in coastal, industrial or saline areas, and are easily eroded by chloride ions, leading to problems such as steel corrosion, concrete cracking and structural performance degradation. Steel corrosion not only reduces its cross-sectional area and mechanical properties, but also may cause the cracking and spalling of the protective layer, thereby accelerating the durability degradation and seriously threatening the safety and service life of the structure.

[0003] In the prior art, reinforced concrete durability analysis mainly relies on chloride ion diffusion models based on Fick's diffusion law and corrosion evolution models based on empirical formulas, but traditional methods often ignore the complex effects of environmental factors, material heterogeneity and changes in service conditions, resulting in insufficient prediction accuracy. In addition, existing researches mainly focus on single deterministic analysis, and do not fully consider the randomness, uncertainty in the material performance degradation process and the multiscale coupling effect between corrosion and overall structural performance degradation.

[0004] In recent years, artificial intelligence technology, especially convolutional neural networks, has shown good application potential in the field of material performance prediction, providing a new approach for modeling complex diffusion processes. At the same time, methods based on random sampling and probability statistics, such as the Monte Carlo method, can effectively capture the uncertainty in the structural performance degradation process and improve the reliability of durability evaluation. Therefore, it is urgent to develop a reinforced concrete structure performance degradation intelligent prediction and durability limit state evaluation method that can comprehensively consider environmental changes, material degradation mechanisms, structural multiscale responses and random uncertainties, in order to realize scientific prediction of the service life and failure risk of the structure and provide a theoretical basis for engineering design, maintenance and reinforcement decisions. SUMMARY

[0005] The application aims to provide a reinforced concrete structure performance degradation intelligent prediction and durability limit state evaluation method to realize the evaluation of the durability limit state of reinforced concrete structures under corrosion conditions.

[0006] The technical solution for realizing the application is as follows:

[0007] A reinforced concrete structure performance degradation intelligent prediction and durability limit state evaluation method comprises the following steps:

[0008] Step one, establish the probability model of environmental parameters and reinforced concrete structure parameters random variables;

[0009] Step two, based on the effective chloride ion diffusion coefficient, build the intelligent prediction model of chloride ion diffusion;

[0010] Step three, with the chloride ion concentration on the surface of the steel bar in the concrete reaching the set critical chloride ion concentration threshold as the limit state criterion of the beginning of steel bar corrosion, the initial corrosion time of the steel bar is calculated according to the effective chloride ion diffusion coefficient;

[0011] Step four, the Monte Carlo sampling algorithm is used to sample the environmental parameters and reinforced concrete structure parameters, and the probability density distribution of the initial corrosion time of the steel bar is obtained by combining the intelligent prediction model of chloride ion diffusion, and the initial corrosion time of the steel bar corresponding to different test samples is calculated;

[0012] Step five, establish the time-varying corrosion model of the steel bar to obtain the change of the corrosion loss rate of the steel bar with time;

[0013] Step six, using the incremental static force analysis method, n times of random sampling are carried out from the established probability model of random variables, and a multi-scale finite element analysis model of the reinforced concrete structure is established for different steel bar corrosion rates;

[0014] Step seven, using the step-by-step loading calculation method, multiple finite element nonlinear analysis is carried out to obtain the maximum interlayer displacement angle of the structure under different loads, and the maximum interlayer displacement angle is taken as the structure damage index to obtain the interlayer displacement angle-load relationship curve under different corrosion loss rates;

[0015] Step eight, the interlayer displacement angle limit value of the normal use limit state and the bearing capacity limit state is defined respectively, the critical load value corresponding to the limit value is extracted in each interlayer displacement angle-load curve, and then the probability density function of the critical load is obtained, and the failure probability curve under two limit states under different corrosion degrees is obtained;

[0016] Step nine, the bearing capacity reduction coefficient K is used to represent the ratio of the bearing capacity of the structure after corrosion to the original bearing capacity, when the coefficient K decreases to the specified limit value with the change of service time, it is regarded as durability failure, and when K decreases to the specified limit value, the time corresponding to the corrosion loss rate is the bearing capacity failure time of the reinforced concrete structure, that is, the durability limit state of the reinforced concrete structure under the corrosion state is evaluated.

[0017] Compared with the prior art, the present application has the following advantages:

[0018] The application introduces a convolutional neural network to construct an intelligent prediction model of chloride ion diffusion, which can comprehensively consider the coupling effect of multiple factors such as material performance and environmental changes, and has higher prediction accuracy and adaptability compared with traditional diffusion models based on empirical formula or simplified assumptions. By establishing a probability distribution model of environment and structure parameters and combining with the Monte Carlo sampling method, the uncertainty characteristics of each stage in the material degradation and structure performance degradation process can be systematically captured, and the scientificity and reliability of durability evaluation are significantly improved.

[0019] In the finite element modeling process, the application adopts a multi-scale modeling strategy, uses solid element refinement in the key area, and uses simplified element modeling in the non-key area, which balances the fine representation of local corrosion effect and the calculation efficiency of overall structure response analysis, and is suitable for durability simulation of large-scale structures in engineering scale. In addition, the application considers the non-uniform corrosion (local pitting) morphology in the steel bar corrosion model, and simulates the cracking and spalling process of the concrete cover through the birth and death element technology, which can more truly reflect the mechanical performance degradation law of reinforced concrete structures during service, and improves the physical consistency of the structure degradation process.

[0020] The application can quantitatively predict the failure risk of normal use limit state and bearing capacity limit state by introducing the maximum interlayer displacement angle as a damage index and deriving the failure probability curve based on the critical load probability density function, which has good engineering application value and promotion prospect. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 The method flowchart of the embodiment of the application is shown in the figure;

[0022] Figure 2 The chloride ion diffusion coefficient prediction diagram is shown in the figure;

[0023] Figure 3 The initial rust time probability distribution diagram is shown in the figure;

[0024] Figure 4 The steel bar corrosion rate model fitting curve is shown in the figure;

[0025] Figure 5 The steel bar pitting model diagram is shown in the figure;

[0026] Figure 6 The structure multi-scale modeling method diagram is shown in the figure;

[0027] Figure 7 The interlayer displacement angle-load curve diagram is shown in the figure;

[0028] Figure 8 The critical load probability distribution diagram is shown in the figure. DETAILED DESCRIPTION

[0029] The application will be further described below in conjunction with the drawings and specific embodiments.

[0030] The method comprises the following steps:

[0031] Step one: establish a probability model of environmental and structural parameter random variables, which provides a database for subsequent convolutional neural network and Monte Carlo sampling.

[0032] Step two: establish an intelligent prediction model of chloride ion diffusion based on the basic framework of Fick's second law, and the specific expression is as follows:

[0033]

[0034] wherein C(x, t) is the chloride ion concentration at a distance x from the erosion surface at an erosion time t, C s is the chloride ion concentration on the surface of the concrete, k t is the temperature correction coefficient, k h is the humidity correction coefficient, k s is the concrete strength correction coefficient, D is the effective chloride ion diffusion coefficient, and erf(·) is the Gaussian error function.

[0035] The important parameter of chloride ion diffusion coefficient D in the intelligent prediction model of chloride ion diffusion is obtained through the prediction of the convolutional neural network. The input data of the convolutional neural network are temperature and humidity, concrete porosity, water-cement ratio, curing age, and protection layer thickness, and the output data is the effective chloride ion diffusion coefficient.

[0036] Step three: set a critical chloride ion concentration threshold, and take the chloride ion concentration on the surface of the steel bar in the concrete reaching the set critical chloride ion concentration threshold as the limit state criterion for the initial corrosion of the steel bar.

[0037]

[0038] wherein t i is the initial corrosion time of the steel bar, x is the protection layer thickness, and C cr is the critical chloride ion concentration.

[0039] Step four: The above calculation of the initial corrosion time of steel bars belongs to a deterministic method. The deterministic method usually assumes that factors such as temperature and humidity, protective layer thickness, and chloride ion concentration are constant, and does not fully consider the random nature of these factors, so uncertainty analysis is needed. Specifically, the Monte Carlo sampling algorithm is used to sample large-scale environmental and structural parameter random variables, and then the probability density distribution of the initial corrosion time of steel bars is obtained by combining the intelligent prediction model of chloride ion diffusion, and the initial corrosion time of steel bars corresponding to different test samples is calculated. Through statistical analysis, the probability density histogram and probability density curve are obtained, and the intelligent prediction of the initial corrosion time of steel bars is realized.

[0040] Step five: Establish a time-varying corrosion model of steel bars to obtain the change of the corrosion loss rate of steel bars with time. The time-varying corrosion model of steel bars is as follows:

[0041] First, calculate the depth p(t) of the corrosion pit of steel bars at time t:

[0042] p(t) = 0.0116i corr (t)R

[0043] i corr (t) = jln(t + p) + k j > 0

[0044] where i corr is the corrosion current density at time t; R is the ratio coefficient between the maximum corrosion rate and the uniform corrosion rate, which can be taken as 4-8; j, p, k are constants related to the properties of reinforced concrete, which are obtained by test fitting.

[0045] The width of the corrosion pit of steel bars after serving t years is a(t), and the cross-sectional area of the corrosion loss of steel bars is S(t), which is calculated as follows:

[0046]

[0047] where S1 is the area of the sector MON S 扇形MON minus the area of the triangle MON S 三角形MON ; S2 is the area of the sector MPN S 扇形MPN minus the area of the triangle MPN S 三角形MPN ; θ1 is the angle MON; θ2 is the angle MPN; where O is the center of the steel bar section, P is the upper vertex of the steel bar section, and M and N are the intersection points of the corrosion loss section and the steel bar section, as shown in Figure 5 The calculation formula is as follows:

[0048]

[0049] where D0 is the initial diameter of the uncorroded steel bar.

[0050] After obtaining the residual cross-sectional area of the steel bar, the steel bar corrosion loss rate is calculated according to the following formula:

[0051]

[0052] Wherein, S0 is the original cross-sectional area of the uncorroded steel bar.

[0053] Step six: using the incremental static analysis method, n times of random sampling is carried out from the established probability model of random variables, and the influence of different corrosion degrees is considered, and the steel reinforced concrete frame structure with different steel bar corrosion loss rates is designed, as shown in Figure 5 The multi-scale finite element analysis model of the reinforced concrete structure is established. The multi-scale finite element analysis model is that the beam-column joint area is modeled in detail by using solid elements, and the remaining part is modeled by using beam elements. The influence of steel bar corrosion on the performance of the structure is embodied by adjusting the existing steel bar constitutive relation, and the cracking and spalling effect caused by corrosion is simulated in time by using the birth-death element technology in the concrete cover.

[0054] Step seven: using the step-by-step loading calculation method, multiple finite element nonlinear analyses are carried out to obtain the maximum interlayer displacement angle of the structure under different loads. Taking the maximum interlayer displacement angle as the structural damage index, the interlayer displacement angle-load relationship curves under different corrosion loss rates are obtained, wherein one corrosion loss rate can correspond to n statistical sample curves.

[0055] Step eight: the interlayer displacement angle limit value of the normal use limit state and the bearing capacity limit state is defined respectively, the critical load value corresponding to the limit value is extracted in each interlayer displacement angle-load curve, based on the n critical loads, the distribution goodness fitting test is carried out, the distribution type and distribution parameters of the critical load can be obtained, and then the distribution probability density function of the critical load under different corrosion degrees is obtained, and the failure probability curve under two limit states can be obtained by integration.

[0056] Step nine: the mode load value of the critical load distribution probability density function obtained under different corrosion states is defined as the bearing capacity of the structure after corrosion, and the bearing capacity reduction coefficient K is used to represent the ratio of the bearing capacity of the structure after corrosion to the bearing capacity of the uncorroded structure. When the coefficient K decreases to the specified limit value with the service time, it is considered as the durability failure. In this study, when the bearing capacity reduction coefficient K decreases to 0.92, it is regarded as the durability limit state, and when K decreases to 0.92, the corrosion loss rate corresponding to the time is the bearing capacity failure time of the reinforced concrete structure, that is, the durability limit state of the reinforced concrete structure under the corrosion state is evaluated.

[0057] Example 1

[0058] The first step of durability assessment of reinforced concrete structures is to establish the probability model of random variables such as environmental parameters and structural parameters. Environmental parameters include chloride ion concentration, temperature and humidity, steel corrosion current density, etc. Structural parameters include concrete strength and size, steel size and performance, etc. All data are collected through field sampling, laboratory testing and literature, and the corresponding probability model is established. Some probability models of parameters can be referred to Table 1.

[0059] Table 1 Probability model of environmental and structural parameters

[0060]

[0061] After establishing the probability model of random variables, an intelligent prediction model of chloride ion diffusion is constructed based on Fick's law to simulate the diffusion process of chloride ions in concrete under the environment. The chloride ion diffusion equation adopts the basic framework of Fick's second law, and the specific expression is:

[0062]

[0063] where C(x, t) is the chloride ion concentration at a distance of x from the erosion surface at an erosion time of t, C s is the chloride ion concentration on the surface of concrete, k t is the temperature correction coefficient, k h is the humidity correction coefficient, k s is the concrete strength correction coefficient, D is the effective chloride ion diffusion coefficient, and erf(·) is the Gaussian error function.

[0064] To improve the accuracy, efficiency and generalization ability of the prediction of chloride ion diffusion coefficient, the important parameter of chloride ion diffusion coefficient D in the intelligent prediction model of chloride ion diffusion is obtained by the prediction of convolutional neural network. The input data of convolutional neural network are temperature and humidity, concrete porosity, water-cement ratio, curing age, protective layer thickness and other factors that may affect the behavior of chloride ion diffusion, and the output data are effective chloride ion diffusion coefficient. The self-learning ability of convolutional neural network is used to train the prediction model. Through training on a large number of training data sets, the neural network can learn the complex nonlinear relationship between these input features and chloride ion diffusion coefficient, so as to realize the prediction of effective chloride ion diffusion coefficient.

[0065] The convolutional neural network model structure is composed of an input layer, multiple convolutional layers, pooling layers, batch normalization layers, fully connected layers, and an output layer to fully extract the key features of chloride ion diffusion. First, the input layer normalizes the original data, making all features converge to the same numerical range to improve training stability. Then, two layers of 3x3 convolutional kernels are used to extract local features of the diffusion data, and ReLU is used as the activation function for each convolutional layer to ensure that the gradient remains stable during backpropagation. The pooling layer uses a 2x2 filter for dimensionality reduction, thereby reducing computational complexity and improving generalization ability. Then, the data is standardized by the batch normalization layer to avoid problems such as gradient vanishing or gradient explosion. After the high-dimensional features are fused by the fully connected layer, the diffusion coefficient D is finally output.

[0066] To measure the prediction error of the convolutional neural network model, the mean squared error (MSE) is used as the loss function. During optimization, the Adam adaptive optimization algorithm is used for gradient update, and the learning rate, number of convolutional layers, filter size, and other key parameters are adjusted through hyperparameter optimization to improve model prediction accuracy. The data set is divided into training set, validation set, and test set in the ratio of 7:2:1, and 5-fold cross-validation is performed to reduce the risk of overfitting. The final CNN prediction results obtained by training are shown in Figure 2 .

[0067] Using the established chloride ion diffusion model, the time-varying curve of chloride ion concentration at each depth of the concrete is obtained. A critical chloride ion concentration threshold is set. When the chloride ion concentration at a certain depth first exceeds this threshold, it is determined that the steel bar has reached the initial rust condition, and the time at which this occurs is recorded as the initial rust time.

[0068] The above method for determining the initial time of steel bar corrosion belongs to a deterministic method. However, deterministic models usually assume that factors such as concrete quality, protective layer thickness, and environmental conditions remain constant, and do not fully consider the randomness of these factors. In fact, these parameters have certain variability in actual engineering. To more reasonably assess the time-varying behavior of steel bar corrosion, key influencing factors can be probabilistically modeled to determine their probability distribution types, and uncertainty analysis can be performed based on the Monte Carlo method. Specifically, using MATLAB software in combination with the above deterministic formula, large-scale random sampling is performed in samples that meet the probability distribution of each initial parameter, and the initial corrosion time of the steel bar corresponding to different test samples is calculated. Through statistical analysis, the probability density histogram and probability density curve can be plotted, as shown in Figure 3 , thereby achieving the evaluation of the limit state of the initial corrosion of the steel bar.

[0069] After the steel bar starts to corrode, a time-varying corrosion model of the steel bar is established to obtain the change of the steel bar corrosion amount over time. First, the depth p(t) of the steel bar corrosion pit at time t is calculated:

[0070] p(t) = 0.0116i corr (t)R

[0071] i corr (t) = jln(t + p) + k j > 0

[0072] where i corr is the corrosion current density at time t; R is the ratio coefficient between the maximum corrosion rate and the uniform corrosion rate, which can be taken as 4-8; j, p, k are constants related to the characteristics of reinforced concrete, which are obtained by fitting experiments, and the fitting curve is shown in Figure 4 .

[0073] The width of the corrosion pit of the steel bar after serving t years is a(t), and the cross-sectional area of the steel bar after corrosion is S(t), and the calculation formula is as follows:

[0074]

[0075] where S1 is the area S 扇形MON of the sector MON minus the area S 三角形MON of the triangle MON; S2 is the area S 扇形MPN of the sector MPN minus the area S 三角形MPN of the triangle MPN; θ1 is the angle MON; θ2 is the angle MPN; where O is the center of the steel bar cross section, P is the upper vertex of the steel bar cross section, and M and N are the two intersection points of the corrosion loss cross section and the steel bar cross section, as shown in Figure 5 , and the calculation formula is as follows:

[0076]

[0077] where D0 is the initial diameter of the uncorroded steel bar.

[0078] After obtaining the remaining cross-sectional area of the steel bar, the corrosion loss rate of the steel bar is calculated according to the following formula:

[0079]

[0080] where S0 is the original cross-sectional area of the uncorroded steel bar.

[0081] The main analysis process of the main incremental static analysis method for the durability limit state of reinforced concrete structures is shown in the following steps. First, random sampling is performed from the probability model of each random variable described above to establish a multi-scale finite element model of the reinforced concrete structure with different corrosion levels. In order to improve the efficiency of the calculation and analysis of the time-varying seismic performance of the structure while reflecting the characteristics of the structure as much as possible, a multi-scale modeling method is used for calculation. That is, for the beam-column joint area which is special in construction, complex in stress and more sensitive to corrosion, a solid element is used for fine modeling, and for the beams and columns in the non-sensitive area, a beam element is used for modeling (in the.inp file, the *rebar command is used to insert steel bars according to the actual reinforcement). The solid element and the beam element are coupled through the "Kinematic coupling" in ABAUQS, and the modeling method and the multi-scale model are shown in Figure 6

[0082] To consider the influence of different corrosion levels, a frame structure with different steel corrosion rates can be designed. The corrosion of steel will cause the characteristic parameter values of the steel constitutive relationship to degrade, while the overall stress-strain curve shape remains almost unchanged. Therefore, a degradation model of the characteristic parameter values of the corroded steel under cyclic loading can be used to modify the constitutive relationship of the corroded steel.

[0083] The influence of corrosion on concrete is mainly in the protective layer, while the influence on the internal core concrete is smaller. With the increase of the loading displacement, the protective layer continuously falls off, and with the increase of the corrosion degree, the protective layer falling off situation continuously worsens. This simulation will timely set up the live and dead elements of the concrete protective layer according to the observed concrete protective layer falling off situation in the test and reality to simulate the influence of corrosion on concrete. The concrete constitutive model in the solid element modeling part adopts the plastic damage constitutive model provided by ABAQUS, and the concrete constitutive model Concrete02 is used for the concrete part of the beam element modeling to simulate the nonlinear behavior of the concrete in the beam element modeling part.

[0084] The maximum inter-story drift angle is used as the structural damage index to establish the inter-story drift angle-load relationship curve. By repeating the above analysis n times using n sets of random variables, n curves can be obtained as shown in Figure 7 For the normal service limit state and the bearing capacity limit state of the structure, the inter-story drift angle limit value is defined, and on each curve, the limit value corresponds to a load value, i.e. the critical load. Based on the n critical loads, distribution goodness fitting test is performed to obtain the distribution type and distribution parameters of the critical loads, and then the distribution probability density function of the critical loads is obtained as shown in Figure 8 The failure probability curves under the two limit states under different corrosion levels can be obtained by integration.

[0085] ​The ratio of the bearing capacity of the structure after corrosion to the original bearing capacity is expressed by a bearing capacity reduction coefficient K. When the coefficient K decreases to a specified limit with the service time, the durability failure is considered to occur. Optionally, the bearing capacity reduction coefficient K is decreased to 0.92 as the durability limit state. When the coefficient K decreases to 0.92, the time corresponding to the corrosion loss rate is the failure time of the reinforced concrete structure, that is, the durability limit state of the reinforced concrete structure in the corrosion state is evaluated.

[0086] The above embodiments are only preferred embodiments of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and equivalent replacements can be made, and these improved and equivalent replaced technical solutions of the claims of the present application all fall within the protection scope of the present application.

Claims

1. A method for intelligent prediction of performance degradation of reinforced concrete structures and durability limit state assessment, characterized in that: include: Step 1: Establish a probability model of random variables of environmental parameters and reinforced concrete structure parameters; Step 2: Based on the effective chloride ion diffusion coefficient, an intelligent prediction model for chloride ion diffusion is constructed; Step 3: Using the chloride ion concentration on the steel bar surface in the concrete reaching the set critical chloride ion concentration threshold as the limit state criterion for the onset of steel bar corrosion, and calculating the initial corrosion time of the steel bar based on the effective chloride ion diffusion coefficient; Step 4: Using the Monte Carlo sampling algorithm to sample environmental parameters and reinforced concrete structure parameters, combined with the chloride ion diffusion intelligent prediction model to obtain the probability density distribution of the initial corrosion time of steel bars, and calculate the initial corrosion time of steel bars corresponding to different test samples; Step 5: Establish a time-varying steel corrosion model to obtain the change of steel corrosion loss rate over time; Step 6: Using the incremental static analysis method, random sampling is performed n times from the established probability model of random variables, and a multi-scale finite element analysis model of the reinforced concrete structure is established for different steel corrosion rates; Step 7: Using a step-by-step loading calculation method, multiple finite element nonlinear analyses are performed to obtain the maximum inter-story displacement angle of the structure under different loads. The maximum inter-story displacement angle is used as a structural damage indicator to obtain the inter-story displacement angle-load relationship curve under different corrosion loss rates; Step 8: Define the story drift angle limits for the serviceability limit state and the ultimate bearing capacity limit state, extract the critical load value corresponding to the limit value from each story drift angle-load curve, and then obtain the probability density function of the critical load, and obtain the failure probability curves under the two limit states at different corrosion levels; Step 9. The bearing capacity reduction coefficient K is used to represent the ratio of the structural bearing capacity after corrosion to the original bearing capacity. When the coefficient K drops to the specified limit with the service time, it is regarded as a durability failure. When K drops to the specified limit, the time of the corresponding corrosion loss rate is the bearing capacity failure time of the reinforced concrete structure, thereby realizing the evaluation of the durability limit state of the reinforced concrete structure under the corrosion state.

2. The method for intelligent prediction of reinforced concrete structure performance degradation and durability limit state assessment according to claim 1 is characterized in that: The chloride ion diffusion intelligent prediction model chloride ion diffusion coefficient is predicted by convolutional neural network. The input data of convolutional neural network are temperature and humidity, concrete porosity, water-cement ratio, curing age, and protective layer thickness. The output data is the effective chloride ion diffusion coefficient.

3. The method for intelligent prediction of reinforced concrete structure performance degradation and durability limit state assessment according to claim 1 is characterized in that: The expression of the intelligent prediction model for chloride ion diffusion is: Where C(x,t) is the chloride ion concentration at a distance x from the erosion surface when the erosion time is t, C s is the chloride ion concentration on the concrete surface, k t is the temperature correction coefficient, k h is the humidity correction factor, k s is the concrete strength correction factor, D is the effective chloride ion diffusion coefficient, and erf(·) is the Gaussian error function.

4. The method for intelligent prediction of reinforced concrete structure performance degradation and durability limit state assessment according to claim 1 is characterized in that: The initial corrosion time of steel bars is calculated by the following formula: Among them, t i is the initial corrosion time of steel bar, x is the thickness of protective layer, C cr is the critical chloride ion concentration, C s is the chloride ion concentration on the concrete surface, k t is the temperature correction coefficient, k h is the humidity correction factor, k s is the concrete strength correction factor, erf(·) is the Gaussian error function, and D is the effective chloride ion diffusion coefficient.

5. The method for intelligent prediction of reinforced concrete structure performance degradation and durability limit state assessment according to claim 1 is characterized in that: The time-varying steel corrosion model obtains the change of steel corrosion loss rate over time: Where S(t) and a(t) are the cross-sectional area and width of the steel bar after t years of service, respectively; p(t) is the depth of the pit after t years of service; D0 is the initial diameter of the uncorroded steel bar; S0 is the original cross-sectional area of ​​the uncorroded steel bar; S1 is the area of ​​the sector MON 扇形MON Subtract the area S of triangle MON 三角形MON ; S2 is the area S of the sector MPN 扇形MPN Subtract the area S of triangle MPN 三角形MPN ; θ1 is the angle MON; θ2 is the angle MPN; where O is the center of the steel bar section, P is the upper vertex of the steel bar section, and M and N are the two intersection points of the corrosion loss section and the steel bar section.

6. The method for intelligent prediction of reinforced concrete structure performance degradation and durability limit state assessment according to claim 5 is characterized in that: The depth p(t) satisfies: p(t)=0.0116i corr (t)R i corr (t)=jln(t+p)+k j>0 Among them, i corr is the corrosion current density at time t; R is the ratio coefficient between the maximum corrosion rate and the uniform corrosion rate, which can be 4 to 8; j, p, k are constants related to the characteristics of reinforced concrete and obtained through experimental fitting.

Images