Method and system for predicting corrosion fatigue life of inhaul cable steel wire under multi-factor coupling effect

Abstract

The invention discloses an inhaul cable steel wire corrosion fatigue life prediction method and system under a multi-factor coupling effect, and the method comprises the steps: carrying out the corrosion fatigue test of a to-be-tested steel wire under a set working condition environment, and obtaining the tested corrosion fatigue life; the environmental parameters in the working condition environment serve as first variables, the tested corrosion fatigue life serves as second variables, the nonlinear relation between the first variables and the second variables is analyzed, and an analysis result is obtained; and according to an analysis result, fitting the first variable and the second variable by utilizing multivariate nonlinear regression to obtain a corrosion fatigue life prediction model. According to the method, the influence rule of each factor on the fatigue life can be comprehensively considered, and the corrosion fatigue life of the steel wire under different corrosion degrees can be rapidly and accurately evaluated.

Description

Technical Field

[0001] The invention relates to the field of material corrosion, and in particular to a method and system for predicting the corrosion fatigue life of a cable steel wire under the coupling effect of multiple factors. Background Art

[0002] High-strength steel wire, as the main load-bearing material of cable-stayed bridges, is prone to corrosion and damage in complex service environments. Once the steel wire is corroded, its corrosion fatigue life is shortened under the action of load, which reduces the reliability of the cable and threatens the safety and reliability of the cable bridge.

[0003] The existing technology used to predict the corrosion fatigue life of cable wire takes relatively simple factors into consideration, resulting in poor prediction ability. It cannot truly reflect the actual evolution of the corrosion fatigue life of the wire in a complex service environment, resulting in poor prediction accuracy, unable to meet actual engineering needs, and cannot be widely promoted and used in actual engineering projects.

[0004] Therefore, a method and system for predicting the corrosion fatigue life of cable steel wire under the coupling of multiple factors is needed, which can comprehensively consider the influence of various factors on fatigue life and quickly and accurately evaluate the corrosion fatigue life of steel wire under different degrees of corrosion. 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 a method and system for predicting the corrosion fatigue life of cable steel wire under the coupling of multiple factors, which can comprehensively consider the influence of various factors on fatigue life and quickly and accurately evaluate the corrosion fatigue life of steel wire under different degrees of corrosion.

[0006] The method for predicting the corrosion fatigue life of a cable steel wire under the coupling effect of multiple factors of the present invention comprises:

[0007] Under the set working conditions, the steel wire to be tested is subjected to corrosion fatigue test to obtain the tested corrosion fatigue life;

[0008] Taking the environmental parameters under the working environment as the first variable and the tested corrosion fatigue life as the second variable, the nonlinear relationship between the first variable and the second variable is analyzed to obtain the analysis results;

[0009] According to the analysis results, the first variable and the second variable are fitted using multivariate nonlinear regression to obtain the corrosion fatigue life prediction model.

[0010] Furthermore, the working environment includes different ambient temperatures, ambient humidity, corrosion factor concentrations and stress levels.

[0011] Furthermore, the environmental parameters include environmental temperature, environmental humidity, corrosion factor concentration, stress level and corrosion time.

[0012] Furthermore, the nonlinear relationship between the first variable and the second variable is analyzed, specifically including:

[0013] A random forest regression model is used, with the first variable as the input variable and the second variable as the output variable;

[0014] Standardize or normalize the input variables, use the random forest algorithm to train the random forest regression model, and use the integration of multiple decision trees to learn the relationship between the input variables and the output variables;

[0015] After training, SHAP values ​​are used to quantify the degree of influence of each input variable and their interactions on the predictions of the random forest regression model.

[0016] Furthermore, the corrosion fatigue life prediction model is determined according to the following formula:

[0017] N=T 0.65 ×h 0.32 ×t 0.87 ×μ 1.17 ×Δσ 1.03 ×10 6 ;

[0018] Wherein, N represents the corrosion fatigue life of the cable wire, T represents the ambient temperature, h represents the ambient humidity, t represents the corrosion time, μ represents the corrosion factor concentration, and Δσ represents the stress level.

[0019] A cable wire corrosion fatigue life prediction system under multi-factor coupling, comprising a test unit, an analysis unit and a prediction unit;

[0020] The test unit is used to perform a corrosion fatigue test on the steel wire to be tested under a set working condition to obtain the tested corrosion fatigue life;

[0021] The analysis unit is used to analyze the nonlinear relationship between the first variable and the second variable by taking the environmental parameter under the working condition as the first variable and the tested corrosion fatigue life as the second variable to obtain the analysis result;

[0022] The prediction unit is used to fit the first variable and the second variable using multivariate nonlinear regression according to the analysis result to obtain a corrosion fatigue life prediction model.

[0023] Furthermore, the working environment includes different ambient temperatures, ambient humidity, corrosion factor concentrations and stress levels.

[0024] Furthermore, the environmental parameters include environmental temperature, environmental humidity, corrosion factor concentration, stress level and corrosion time.

[0025] Furthermore, the nonlinear relationship between the first variable and the second variable is analyzed, specifically including:

[0026] A random forest regression model is used, with the first variable as the input variable and the second variable as the output variable;

[0027] Standardize or normalize the input variables, use the random forest algorithm to train the random forest regression model, and use the integration of multiple decision trees to learn the relationship between the input variables and the output variables;

[0028] After training, SHAP values ​​are used to quantify the degree of influence of each input variable and their interactions on the predictions of the random forest regression model.

[0029] Furthermore, the corrosion fatigue life prediction model is determined according to the following formula:

[0030] N=T 0.65 ×h 0.32 ×t 0.87 ×μ 1.17 ×Δσ 1.03 ×10 6 ;

[0031] Wherein, N represents the corrosion fatigue life of the cable wire, T represents the ambient temperature, h represents the ambient humidity, t represents the corrosion time, μ represents the corrosion factor concentration, and Δσ represents the stress level.

[0032] The beneficial effects of the present invention are as follows: a method and system for predicting the corrosion fatigue life of cable steel wire under the action of multi-factor coupling is disclosed in the present invention. By comprehensively considering the influence of various factors on fatigue life, the complex nonlinear relationship between various factors and the corrosion fatigue life of steel wire is analyzed by using a machine learning model. Based on nonlinear regression analysis, a prediction model for the corrosion fatigue life of cable steel wire under the influence of multi-factor coupling is established, thereby obtaining a calculation method for the corrosion fatigue life of cable steel wire, comprehensively revealing the decay law of the corrosion fatigue life of rusted steel wire, and being able to quickly and accurately evaluate the corrosion fatigue life of steel wire under different degrees of corrosion. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0034] Figure 1 It is a schematic diagram of the process flow of the corrosion fatigue life prediction method of the present invention;

[0035] Figure 2 It is a schematic diagram of the corrosion fatigue test process of the present invention;

[0036] Figure 3 It is a schematic diagram of characteristic parameter analysis results of the present invention;

[0037] Figure 4 It is a schematic diagram of the error comparison result of the present invention. DETAILED DESCRIPTION

[0038] The present invention is further described below in conjunction with the accompanying drawings, as shown in the drawings:

[0039] This embodiment discloses a method for predicting the corrosion fatigue life of a cable steel wire under the coupling effect of multiple factors, comprising the following steps:

[0040] Under the set working conditions, the steel wire to be tested is subjected to corrosion fatigue test to obtain the tested corrosion fatigue life;

[0041] Taking the environmental parameters under the working environment as the first variable and the tested corrosion fatigue life as the second variable, the nonlinear relationship between the first variable and the second variable is analyzed to obtain the analysis results;

[0042] According to the analysis results, the first variable and the second variable are fitted using multivariate nonlinear regression to obtain the corrosion fatigue life prediction model.

[0043] In this embodiment, high-strength parallel steel wire is selected as the steel wire to be tested. According to the artificial accelerated corrosion test operating procedures, the high-strength parallel steel wire is subjected to corrosion fatigue under different environmental temperatures, environmental humidity, corrosion factor concentrations, and stress levels. After different periods of corrosion fatigue, the steel wire is taken out, and the rust products on the surface of the steel wire are cleared with a rust remover, and then the steel wire is left to stand and dried. After that, the fatigue life test is carried out. The entire corrosion fatigue process is as follows. Figure 2 By changing the temperature, humidity, corrosion factor concentration and stress level, the influence of each factor on the fatigue life of the steel wire can be better revealed.

[0044] In this embodiment, the environmental parameters include ambient temperature, ambient humidity, corrosion factor concentration, stress level and corrosion time. Since corrosion is dynamic and directly related to fatigue life, setting the corrosion time can accurately simulate the corrosion process, quantitatively evaluate the corrosion degree, and analyze the interaction effect with other environmental parameters.

[0045] In this embodiment, the nonlinear relationship between the first variable and the second variable is analyzed, specifically including:

[0046] Using a machine learning model, such as a random forest regression model, the first variable is used as an input variable and the second variable is used as an output variable; that is, ambient temperature, ambient humidity, corrosion factor concentration, stress level, and corrosion time are used as input variables, and the corrosion fatigue life of the steel wire is used as an output variable;

[0047] Standardize or normalize the input variables to improve the performance of the random forest regression model. Use the random forest algorithm to train the random forest regression model, using an ensemble of multiple decision trees to learn the relationship between the input variables and the output variables;

[0048] After training, the importance of features is analyzed, that is, the interaction effects between them are analyzed to understand which input variables have a greater impact on fatigue life. For example, SHAP values ​​are used to quantify the impact of each input variable and its interaction on the prediction of the random forest regression model. Among them, the SHAP values ​​of erosion time and ambient temperature in the interaction effect analysis are as follows: Figure 3 As shown in the figure, the SHAP value represents the additive feature value, which is used to judge the significant impact between parameters. The SHAP value not only shows the impact of a single feature on the prediction results, but can also be used to compare the significant impact between different features. By calculating the absolute value of the SHAP value of each feature, it can be concluded which features have a greater impact on the model prediction results.

[0049] Through the above analysis, the complex nonlinear relationship between the first variable and the second variable can be obtained, which provides technical support for the subsequent multivariate nonlinear regression analysis.

[0050] In this embodiment, on the basis of clarifying the interaction effect of each variable, based on the multivariate nonlinear regression theory, a prediction model for the corrosion fatigue life of the cable steel wire under the coupling effect of multiple factors can be obtained.

[0051] According to the collected data of the first variable and the second variable and the analysis result of the nonlinear relationship, a suitable nonlinear regression model is selected, such as polynomial regression, support vector regression (SVR) or neural network.

[0052] The collected data is used to fit the selected nonlinear regression model and solve the fitting parameters in the nonlinear regression model. Existing statistical software or programming tools can be used to solve the fitting parameters; in addition, during the fitting process, if overfitting or underfitting occurs, the fitting parameters are adjusted to obtain the optimal fitting parameters.

[0053] Using the optimal fitting parameters obtained, the corrosion fatigue life prediction model is determined according to the following formula:

[0054] N=T 0.65 ×h 0.32 ×t 0.87 ×μ 1.17 ×Δσ 1.03 ×10 6 ;

[0055] Wherein, N represents the corrosion fatigue life of the cable wire, T represents the ambient temperature, h represents the ambient humidity, t represents the corrosion time, μ represents the corrosion factor concentration, and Δσ represents the stress level.

[0056] The present invention also relates to a corrosion fatigue life prediction system for cable steel wires under the action of multi-factor coupling. The prediction system corresponds to the corrosion fatigue life prediction method of the above embodiment and can be understood as a system for implementing the above corrosion fatigue life prediction method. The system includes a test unit, an analysis unit and a prediction unit.

[0057] The test unit is used to perform a corrosion fatigue test on the steel wire to be tested under a set working condition to obtain the tested corrosion fatigue life;

[0058] The analysis unit is used to analyze the nonlinear relationship between the first variable and the second variable by taking the environmental parameter under the working condition as the first variable and the tested corrosion fatigue life as the second variable to obtain the analysis result;

[0059] The prediction unit is used to fit the first variable and the second variable using multivariate nonlinear regression according to the analysis result to obtain a corrosion fatigue life prediction model.

[0060] In order to verify the accuracy of the prediction model of the present invention, corresponding experimental research was carried out to test the corrosion fatigue life of steel wires with different corrosion degrees. 100 samples were selected and the test results were compared with the model prediction results. The relative error results of the two are shown in Figure 2. Figure 4 The results show that the calculation errors of the 100 samples are all within 10%, indicating that the prediction model proposed in the present invention can better predict the corrosion fatigue life of the corroded steel wire.

[0061] The present invention provides an objective and scientific method and system for predicting the corrosion fatigue life of cable steel wires, reveals the attenuation law of the fatigue life of cable steel wires under the coupling of multiple factors, and constructs a prediction model for the corrosion fatigue life of cable steel wires under the coupling of multiple factors. The calculation model can be used to calculate the corrosion fatigue life of steel wires at different degrees of corrosion, providing technical support for evaluating the service status of cable steel wires under complex service environments.

[0062] 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 predicting the corrosion fatigue life of cable steel wire under the coupling effect of multiple factors, characterized by: include: Under the set working conditions, the steel wire to be tested is subjected to corrosion fatigue test to obtain the tested corrosion fatigue life; Taking the environmental parameters under the working environment as the first variable and the tested corrosion fatigue life as the second variable, the nonlinear relationship between the first variable and the second variable is analyzed to obtain the analysis results; According to the analysis results, the first variable and the second variable are fitted using multivariate nonlinear regression to obtain the corrosion fatigue life prediction model.

2. The method for predicting the corrosion fatigue life of cable steel wire under the multi-factor coupling effect according to claim 1 is characterized by: The working environment includes different ambient temperatures, ambient humidity, corrosion factor concentrations and stress levels.

3. The method for predicting the corrosion fatigue life of cable steel wire under the multi-factor coupling effect according to claim 1 is characterized in that: The environmental parameters include environmental temperature, environmental humidity, corrosion factor concentration, stress level and corrosion time.

4. The method for predicting the corrosion fatigue life of cable steel wire under the multi-factor coupling effect according to claim 1 is characterized in that: The nonlinear relationship between the first variable and the second variable is analyzed, including: A random forest regression model is used, with the first variable as the input variable and the second variable as the output variable; Standardize or normalize the input variables, use the random forest algorithm to train the random forest regression model, and use the integration of multiple decision trees to learn the relationship between the input variables and the output variables; After training, SHAP values ​​are used to quantify the degree of influence of each input variable and their interactions on the predictions of the random forest regression model.

5. The method for predicting the corrosion fatigue life of cable steel wire under the multi-factor coupling effect according to claim 1 is characterized in that: The corrosion fatigue life prediction model is determined according to the following formula: N=T 0.65 ×h 0.32 ×t 0.87 ×μ 1.17 ×Δσ 1.03 ×10 6 ; Wherein, N represents the corrosion fatigue life of the cable wire, T represents the ambient temperature, h represents the ambient humidity, t represents the corrosion time, μ represents the corrosion factor concentration, and Δσ represents the stress level.

6. A system for predicting the corrosion fatigue life of cable steel wire under the action of multiple factors coupling, characterized by: It includes experimental unit, analysis unit and prediction unit; The test unit is used to perform a corrosion fatigue test on the steel wire to be tested under a set working condition to obtain the tested corrosion fatigue life; The analysis unit is used to analyze the nonlinear relationship between the first variable and the second variable by taking the environmental parameter under the working condition as the first variable and the tested corrosion fatigue life as the second variable to obtain the analysis result; The prediction unit is used to fit the first variable and the second variable using multivariate nonlinear regression according to the analysis result to obtain a corrosion fatigue life prediction model.

7. The system for predicting the corrosion fatigue life of cable steel wire under the multi-factor coupling effect according to claim 6 is characterized by: The working environment includes different ambient temperatures, ambient humidity, corrosion factor concentrations and stress levels.

8. The system for predicting the corrosion fatigue life of cable steel wire under the multi-factor coupling effect according to claim 6 is characterized by: The environmental parameters include environmental temperature, environmental humidity, corrosion factor concentration, stress level and corrosion time.

9. The system for predicting the corrosion fatigue life of cable steel wire under the multi-factor coupling effect according to claim 6 is characterized in that: The nonlinear relationship between the first variable and the second variable is analyzed, including: A random forest regression model is used, with the first variable as the input variable and the second variable as the output variable; Standardize or normalize the input variables, use the random forest algorithm to train the random forest regression model, and use the integration of multiple decision trees to learn the relationship between the input variables and the output variables; After training, SHAP values ​​are used to quantify the degree of influence of each input variable and their interactions on the predictions of the random forest regression model.

10. The system for predicting the corrosion fatigue life of cable steel wire under the multi-factor coupling effect according to claim 6 is characterized in that: The corrosion fatigue life prediction model is determined according to the following formula: N=T 0.65 ×h 0.32 ×t 0.87 ×μ 1.17 ×Δσ 1.03 ×10 6 ; Wherein, N represents the corrosion fatigue life of the cable wire, T represents the ambient temperature, h represents the ambient humidity, t represents the corrosion time, μ represents the corrosion factor concentration, and Δσ represents the stress level.

Images