Evaluation method and system for chloride ion diffusion in concrete structure

Abstract

The invention discloses a method and system for evaluating chloride ion diffusion in a concrete structure, and the method comprises the steps: carrying out a test of corroding concrete by chloride ions under a set working condition environment, and obtaining chloride ion concentrations under different erosion depths; the chloride ion diffusion coefficient under the test is solved by using the chloride ion concentrations under different erosion depths; performing expression fitting on the chloride ion diffusion coefficient under the test to obtain a chloride ion diffusion coefficient prediction model, and setting the chloride ion diffusion coefficient prediction model as a boundary condition to the neural network model; performing iterative output on the neural network model, and if the change of the output value of the neural network model between the mth time and the (m + 1) th time is smaller than a threshold value, setting the mth output value to a chloride ion diffusion coefficient prediction model to obtain an analyzed prediction model; and calculating to obtain the chloride ion diffusion coefficient under the actual working condition by using the analyzed prediction model. The degree of dependence on data can be reduced, and the prediction precision of the chloride ion diffusion coefficient is improved.

Description

Technical Field

[0001] The invention relates to the field of material determination, and in particular to an evaluation method and system for chloride ion diffusion in a concrete structure. Background Art

[0002] At present, although the current chloride ion diffusion coefficient prediction model can reveal the diffusion behavior of chloride ions in concrete structures to a certain extent, the factors it considers are relatively single, and only considers the effects of environmental factors, material factors and load factors on the chloride ion transmission law, ignoring their coupling effect.

[0003] The existing diffusion coefficient prediction models mainly rely on linear regression, nonlinear regression and traditional machine learning models. When there are many input variables, it is difficult to reveal their complex mapping relationships using linear regression and nonlinear regression theories. The generalization ability of traditional machine learning models is more dependent on the scope of training data. When encountering new scenarios beyond the scope of training data, they perform poorly and cannot accurately evaluate the diffusion coefficient of chloride ions under complex service environment conditions.

[0004] Therefore, a new chloride ion diffusion evaluation method and system are needed to reduce the dependence on data and improve the prediction accuracy of chloride ion diffusion coefficient. Summary of the invention

[0005] In view of this, the purpose of the present invention is to overcome the defects in the prior art and provide an evaluation method and system for chloride ion diffusion in concrete structures, which can reduce the dependence on data and improve the prediction accuracy of chloride ion diffusion coefficient.

[0006] The method for evaluating chloride ion diffusion in a concrete structure of the present invention comprises:

[0007] Conduct chloride ion corrosion tests on concrete under set working conditions to obtain chloride ion concentrations at different corrosion depths;

[0008] Using the chloride ion concentration at different corrosion depths, the chloride ion diffusion coefficient under the test is solved;

[0009] An expression for the chloride ion diffusion coefficient under the test is formulated to obtain a chloride ion diffusion coefficient prediction model, and the chloride ion diffusion coefficient prediction model is set as a boundary condition in the neural network model;

[0010] The neural network model is made to iteratively output, and if the change of the output value of the neural network model between the mth time and the m+1th time is less than a threshold value, the mth output value is set to the chloride ion diffusion coefficient prediction model to obtain a predicted model after analysis;

[0011] The chloride ion diffusion coefficient under actual working conditions is calculated using the analytical prediction model.

[0012] Furthermore, a test of chloride ion corrosion of concrete was conducted under the set working conditions, specifically including:

[0013] Prepare concrete specimens with different water-cement ratios and aggregate volume fractions, and place the concrete specimens in an environmental test chamber;

[0014] In the environmental test chamber, different ambient temperatures, ambient humidity and stress levels are set for the concrete specimens, and the five sides of the environmental test chamber are sealed to allow chloride ions to corrode the concrete from only one side;

[0015] After the concrete specimens have been corroded for different periods of time, they are taken out in batches and the chloride ion concentration at different corrosion depths is tested using a chloride ion meter.

[0016] Furthermore, the chloride ion diffusion coefficient under the test is solved, specifically including: substituting the chloride ion concentration under different corrosion depths into the expression of Fick's second law to solve the chloride ion diffusion coefficient under the test.

[0017] Further, the chloride ion diffusion coefficient prediction model is determined according to the following formula:

[0018]

[0019] Among them, D i is the chloride ion diffusion coefficient under different working conditions, D 0 represents the diffusion coefficient of chloride ions after k days of corrosion, T represents the ambient temperature, T 0 represents the curing temperature, H represents the ambient humidity, σ represents the stress level, t represents the erosion time, t 0 The value is k days, W / C is the water-cement ratio, v represents the volume fraction of aggregate, β i To estimate the parameters, i=1,2,3,4,5,6.

[0020] Furthermore, the input values ​​of the neural network model include ambient temperature, ambient humidity, stress level, water-cement ratio, aggregate volume fraction and erosion time; and the output values ​​of the neural network model include estimated parameters.

[0021] Further, the parsed prediction model is determined according to the following formula:

[0022]

[0023] Among them, D i is the chloride ion diffusion coefficient under different working conditions, T represents the ambient temperature, T 0 represents the curing temperature, H represents the ambient humidity, σ represents the stress level, t represents the erosion time, t 0 The value is k days, W / C is the water-cement ratio, and v represents the volume fraction of aggregate.

[0024] An evaluation system for chloride ion diffusion in a concrete structure, comprising a test unit, a model unit and an evaluation unit;

[0025] The test unit is used to conduct a test on chloride ion corrosion of concrete under a set working condition environment to obtain chloride ion concentrations at different corrosion depths; and to solve the chloride ion diffusion coefficient under the test by using the chloride ion concentrations at different corrosion depths;

[0026] The model unit is used to formulate an expression for the chloride ion diffusion coefficient under the test to obtain a chloride ion diffusion coefficient prediction model, and set the chloride ion diffusion coefficient prediction model as a boundary condition to the neural network model; the neural network model is iteratively output, and if the change of the output value of the neural network model between the mth time and the m+1th time is less than a threshold value, the mth output value is set to the chloride ion diffusion coefficient prediction model to obtain a parsed prediction model;

[0027] The evaluation unit is used to calculate the chloride ion diffusion coefficient under actual working conditions using the analyzed prediction model.

[0028] Furthermore, a test of chloride ion corrosion of concrete was conducted under the set working conditions, specifically including:

[0029] Prepare concrete specimens with different water-cement ratios and aggregate volume fractions, and place the concrete specimens in an environmental test chamber;

[0030] In the environmental test chamber, different ambient temperatures, ambient humidity and stress levels are set for the concrete specimens, and the five sides of the environmental test chamber are sealed to allow chloride ions to corrode the concrete from only one side;

[0031] After the concrete specimens have been corroded for different periods of time, they are taken out in batches and the chloride ion concentration at different corrosion depths is tested using a chloride ion meter.

[0032] Further, the chloride ion diffusion coefficient prediction model is determined according to the following formula:

[0033]

[0034] Among them, D i is the chloride ion diffusion coefficient under different working conditions, D 0 represents the diffusion coefficient of chloride ions after k days of corrosion, T represents the ambient temperature, T 0 represents the curing temperature, H represents the ambient humidity, σ represents the stress level, t represents the erosion time, t 0 The value is k days, W / C is the water-cement ratio, v represents the volume fraction of aggregate, β i To estimate the parameters, i=1,2,3,4,5,6.

[0035] Further, the parsed prediction model is determined according to the following formula:

[0036]

[0037] Among them, D i is the chloride ion diffusion coefficient under different working conditions, T represents the ambient temperature, T 0 represents the curing temperature, H represents the ambient humidity, σ represents the stress level, t represents the erosion time, t 0 The value is k days, W / C is the water-cement ratio, and v represents the volume fraction of aggregate.

[0038] The beneficial effects of the present invention are as follows: a method and system for evaluating chloride ion diffusion in a concrete structure disclosed in the present invention comprehensively considers the influence of material factors, environmental factors and loads on the chloride ion diffusion coefficient, and embeds mathematical formulas into the machine learning model based on the physical machine learning method, which can significantly improve the prediction accuracy and generalization ability of the model, and combines the advantages of physical laws and data-driven to ensure that the prediction results not only conform to the existing data, but also follow the laws of physics, thereby breaking the dependence of traditional machine learning models on large-scale experimental data, and providing a technical reference for further analyzing the durability performance of reinforced concrete structures. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0040] Figure 1 It is a schematic diagram of the test scene of the specimen of the present invention;

[0041] Figure 2 It is a schematic diagram of the sampling process of the test piece of the present invention;

[0042] Figure 3 Schematic diagram of the weight relationship of each input variable of the present invention;

[0043] Figure 4 A schematic diagram of the neural network model updating process of the present invention;

[0044] Figure 5 It is a schematic diagram of the relative error comparison results of the present invention. DETAILED DESCRIPTION

[0045] The present invention is further described below in conjunction with the accompanying drawings, as shown in the drawings:

[0046] This embodiment discloses a method for evaluating chloride ion diffusion in a concrete structure, comprising the following steps:

[0047] Conduct chloride ion corrosion tests on concrete under set working conditions to obtain chloride ion concentrations at different corrosion depths;

[0048] Using the chloride ion concentration at different corrosion depths, the chloride ion diffusion coefficient under the test is solved;

[0049] An expression for the chloride ion diffusion coefficient under the test is formulated to obtain a chloride ion diffusion coefficient prediction model, and the chloride ion diffusion coefficient prediction model is set as a boundary condition in the neural network model;

[0050] The neural network model is made to iteratively output, and if the change of the output value of the neural network model between the mth time and the m+1th time is less than a threshold value, the mth output value is set to the chloride ion diffusion coefficient prediction model to obtain a predicted model after analysis;

[0051] The chloride ion diffusion coefficient under actual working conditions is calculated using the analytical prediction model.

[0052] In this embodiment, the present invention uses a multifunctional environmental test chamber to conduct tests on concrete specimens with different water-cement ratios and coarse aggregate volume fractions under different ambient temperatures, ambient humidity and stress levels. The test scenarios are as follows: Figure 1 By carrying out durability tests on concrete structures under the coupling of multiple factors, the diffusion law of chloride ions in concrete structures is determined, providing support for the establishment of a chloride ion diffusion coefficient prediction model under complex environments.

[0053] The test of chloride ion corrosion concrete is carried out under the set working conditions, including:

[0054] Prepare concrete specimens with different water-cement ratios and aggregate volume fractions, and place the concrete specimens in an environmental test chamber;

[0055] In the environmental test chamber, different ambient temperatures, ambient humidity and stress levels were set for the concrete specimens. During the test, in order to intuitively describe the corrosion law of chloride ions, epoxy resin was used to seal five of the concrete surfaces, allowing chloride ions to corrode the concrete from only one of the surfaces.

[0056] After the concrete specimens have been corroded for different periods of time, they are taken out in batches and the chloride ion concentration at different corrosion depths is tested using a chloride ion meter. Figure 2 Concrete powder was drilled at the hole diameter position shown. Then, the powder at different positions was collected using a reagent bottle and completely dissolved in pure water. Finally, the chloride ion concentration meter was used to test the chloride ion content in the concrete powder at different depths to prepare for the construction of a prediction model for the chloride ion transfer coefficient.

[0057] In this embodiment, solving the chloride ion diffusion coefficient under the test specifically includes: substituting the chloride ion concentration under different corrosion depths into the expression of Fick's second law to solve the chloride ion diffusion coefficient under the test.

[0058] Among them, Fick's second law expression is:

[0059]

[0060] Where C(x,t) represents the chloride ion concentration (%) at each depth, t represents the erosion time, x represents the erosion depth, D represents the chloride ion diffusion coefficient, and C s Indicates the surface chloride ion concentration.

[0061] erf() represents the error function, which is used to describe the probability cumulative function of the normal distribution. Its specific expression is shown in the following formula, which can be solved by the least squares method.

[0062]

[0063] Among them, β represents the integral variable, which means accumulation from 0 to a given value z; z is a free variable and can take any real value; according to different application scenarios, z can represent a variable, such as the standard deviation multiple in a normal distribution. When z is equal to 0, erf() is equal to 0; when z approaches positive infinity, erf() approaches 1; when z approaches negative infinity, erf() approaches -1.

[0064] In this embodiment, the grey correlation theory is used to judge the influence of each input variable (ambient temperature, ambient humidity, water-cement ratio, etc.) on the chloride ion diffusion coefficient, so as to determine the reasonable form of the diffusion coefficient regression equation. The results are as follows: Figure 3 shown.

[0065] according to Figure 3 Based on the rules shown in , combined with the influence of various factors on chloride ion diffusion under the coupling of multiple factors, the expression of chloride ion diffusion coefficient is proposed, and the prediction model of chloride ion diffusion coefficient is obtained:

[0066]

[0067] Among them, D i is the chloride ion diffusion coefficient under different working conditions, D 0 represents the diffusion coefficient of chloride ions after 28 days of corrosion, T represents the ambient temperature, T 0 represents the curing temperature, H represents the ambient humidity, σ represents the stress level, t represents the erosion time, t 0 The value is 28 days, W / C is the water-cement ratio, v represents the volume fraction of aggregate, β i To estimate the parameters, i = 1, 2, 3, 4, 5, 6. Among them, the curing temperature refers to the temperature maintained by the concrete during the curing period. The specific curing temperature will vary according to the actual working condition design. It is generally recommended to be around 20°C to ensure the smooth hydration reaction of the concrete.

[0068] In this embodiment, in order to ensure the calculation accuracy of the diffusion coefficient prediction model, the diffusion coefficient prediction model is directly embedded into the structure or loss function of the neural network model as a boundary condition. Through these constraints, the results output by the neural network model are more consistent with physical reality and unreasonable predictions are avoided. The specific analysis steps are as follows: Figure 4 The input values ​​of the neural network model include ambient temperature, ambient humidity, stress level, water-cement ratio, aggregate volume fraction and erosion time; the output values ​​of the neural network model include the estimated parameter β i , i=1,2,3,4,5,6.

[0069] The neural network model is iterated to output. If the change in the output value of the neural network model between the mth and m+1th times is less than the threshold, for example, for the estimated parameter β 1 , the 900th value is 1.02, the 901th value is 1.019, and the threshold is 0.001; |1.02-1.019| / 1.02 is less than 0.001, so the 900th output value 1.02 is taken as the estimated parameter β 1 The principles for determining the values ​​of other estimated parameters are the same and will not be elaborated here.

[0070] According to the above processing, the parsed prediction model is obtained:

[0071]

[0072] Among them, D i is the chloride ion diffusion coefficient under different working conditions, T represents the ambient temperature, T 0 represents the curing temperature, H represents the ambient humidity, σ represents the stress level, t represents the erosion time, t 0 The value is 28 days, W / C is the water-cement ratio, and v represents the volume fraction of aggregate.

[0073] The present invention also relates to an evaluation system for chloride ion diffusion in a concrete structure. The evaluation system corresponds to the evaluation method of the above embodiment and can be understood as a system for implementing the above evaluation method. The system includes a test unit, a model unit and an evaluation unit.

[0074] The test unit is used to conduct a test on chloride ion corrosion of concrete under a set working condition environment to obtain chloride ion concentrations at different corrosion depths; and to solve the chloride ion diffusion coefficient under the test by using the chloride ion concentrations at different corrosion depths;

[0075] The model unit is used to formulate an expression for the chloride ion diffusion coefficient under the test to obtain a chloride ion diffusion coefficient prediction model, and set the chloride ion diffusion coefficient prediction model as a boundary condition to the neural network model; the neural network model is iteratively output, and if the change of the output value of the neural network model is less than a threshold value, the output value at this time is set to the chloride ion diffusion coefficient prediction model to obtain a parsed prediction model;

[0076] The evaluation unit is used to calculate the chloride ion diffusion coefficient under actual working conditions using the analyzed prediction model.

[0077] The prediction model analyzed by the present invention can be used to solve the chloride ion diffusion coefficient in concrete under different conditions, and compare it with the test results. The comparison results are as follows: Figure 5 The results show that the relative error between the test results and the prediction results of the present invention is within 10%, indicating that the present invention can achieve accurate prediction of the chloride ion diffusion coefficient under complex environments, can better meet the needs of actual engineering, and thus provide technical support for analyzing the durability of reinforced concrete structures.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A method for evaluating chloride ion diffusion in a concrete structure, characterized in that: include: Conduct chloride ion corrosion tests on concrete under set working conditions to obtain chloride ion concentrations at different corrosion depths; Using the chloride ion concentration at different corrosion depths, the chloride ion diffusion coefficient under the test is solved; An expression for the chloride ion diffusion coefficient under the test is formulated to obtain a chloride ion diffusion coefficient prediction model, and the chloride ion diffusion coefficient prediction model is set as a boundary condition in the neural network model; The neural network model is made to iteratively output, and if the change of the output value of the neural network model between the mth time and the m+1th time is less than the threshold value, the mth output value is set to the chloride ion diffusion coefficient prediction model to obtain the analyzed prediction model; The chloride ion diffusion coefficient under actual working conditions is calculated using the analytical prediction model.

2. The method for evaluating chloride ion diffusion in a concrete structure according to claim 1, characterized in that: The test of chloride ion corrosion concrete is carried out under the set working conditions, including: Prepare concrete specimens with different water-cement ratios and aggregate volume fractions, and place the concrete specimens in an environmental test chamber; In the environmental test chamber, different ambient temperatures, ambient humidity and stress levels are set for the concrete specimens, and the five sides of the environmental test chamber are sealed to allow chloride ions to corrode the concrete from only one side; After the concrete specimens have been corroded for different periods of time, they are taken out in batches and the chloride ion concentration at different corrosion depths is tested using a chloride ion meter.

3. The method for evaluating chloride ion diffusion in a concrete structure according to claim 1, characterized in that: Solving the chloride ion diffusion coefficient under the test specifically includes: substituting the chloride ion concentration under different corrosion depths into the expression of Fick's second law to solve the chloride ion diffusion coefficient under the test.

4. The method for evaluating chloride ion diffusion in a concrete structure according to claim 1, characterized in that: The chloride ion diffusion coefficient prediction model is determined according to the following formula: Among them, D i is the chloride ion diffusion coefficient under different working conditions, D0 is the chloride ion diffusion coefficient after k days of corrosion, T is the ambient temperature, T0 is the curing temperature, H is the ambient humidity, σ is the stress level, t is the corrosion time, t0 is k days, W / C is the water-cement ratio, v is the volume fraction of aggregate, β i To estimate the parameters, i=1,2,3,4,5,6.

5. The method for evaluating chloride ion diffusion in a concrete structure according to claim 1, characterized in that: The input values ​​of the neural network model include ambient temperature, ambient humidity, stress level, water-cement ratio, aggregate volume fraction and erosion time; the output values ​​of the neural network model include estimated parameters.

6. The method for evaluating chloride ion diffusion in a concrete structure according to claim 1, characterized in that: The parsed prediction model is determined according to the following formula: Among them, D i is the chloride ion diffusion coefficient under different working conditions, T represents the ambient temperature, T0 represents the curing temperature, H represents the ambient humidity, σ represents the stress level, t represents the erosion time, t0 is k days, W / C represents the water-cement ratio, and v represents the volume fraction of aggregate.

7. A system for evaluating chloride ion diffusion in concrete structures, characterized in that: It includes test unit, model unit and assessment unit; The test unit is used to conduct a test on chloride ion corrosion of concrete under a set working condition environment to obtain chloride ion concentrations at different corrosion depths; and to solve the chloride ion diffusion coefficient under the test using the chloride ion concentrations at different corrosion depths; The model unit is used to formulate an expression for the chloride ion diffusion coefficient under the test to obtain a chloride ion diffusion coefficient prediction model, and set the chloride ion diffusion coefficient prediction model as a boundary condition to the neural network model; the neural network model is iteratively output, and if the change of the output value of the neural network model between the mth time and the m+1th time is less than a threshold value, the mth output value is set to the chloride ion diffusion coefficient prediction model to obtain a parsed prediction model; The evaluation unit is used to calculate the chloride ion diffusion coefficient under actual working conditions using the analyzed prediction model.

8. The system for evaluating chloride ion diffusion in concrete structure according to claim 7, characterized in that: The test of chloride ion corrosion concrete is carried out under the set working conditions, including: Prepare concrete specimens with different water-cement ratios and aggregate volume fractions, and place the concrete specimens in an environmental test chamber; In the environmental test chamber, different ambient temperatures, ambient humidity and stress levels are set for the concrete specimens, and the five sides of the environmental test chamber are sealed to allow chloride ions to corrode the concrete from only one side; After the concrete specimens have been corroded for different periods of time, they are taken out in batches and the chloride ion concentration at different corrosion depths is tested using a chloride ion meter.

9. The system for evaluating chloride ion diffusion in concrete structure according to claim 7, characterized in that: The chloride ion diffusion coefficient prediction model is determined according to the following formula: Among them, D i is the chloride ion diffusion coefficient under different working conditions, D0 is the chloride ion diffusion coefficient after k days of corrosion, T is the ambient temperature, T0 is the curing temperature, H is the ambient humidity, σ is the stress level, t is the corrosion time, t0 is k days, W / C is the water-cement ratio, v is the volume fraction of aggregate, β i To estimate the parameters, i=1,2,3,4,5,6.

10. The system for evaluating chloride ion diffusion in concrete structure according to claim 7, characterized in that: The parsed prediction model is determined according to the following formula: Among them, D i is the chloride ion diffusion coefficient under different working conditions, T represents the ambient temperature, T0 represents the curing temperature, H represents the ambient humidity, σ represents the stress level, t represents the erosion time, t0 is k days, W / C represents the water-cement ratio, and v represents the volume fraction of aggregate.

Images